Manual

Advanced Decision Support Software for the Millennium
Part I: Quick Start Guide

Part II: Tutorials
support@expertchoice.com
sales@expertchoice.com
info@expertchoice.com
www.expertchoice.com
Quick Start Guide and Tutorials Copyright 2000-2002 Expert Choice, Inc.,
Pittsburgh PA
Software Copyright 1982-2002, Patents: 4,613,946; 5,995,728; 6,067,719; other

patents pending. Expert Choice, Inc., Pittsburgh, PA
Credits
Based on the theory, the Analytic Hierarchy Process: Thomas L. Saaty Ph.D.
System Design: Ernest H. Forman, D.Sc.
Programming: Ernest H. Forman, D.Sc.
Anatoly Shvartsman
Michael R. Forman
Deployment: Michael Korpics
Technical Support: John Zottola
Documentation: Mary Ann Selly
We wish to thank James Devlin of our Professional Services Staff for reviewing this
document.
Due to our commitment to constantly improve our software, as well as the

documentation, this manual may not include the most current changes. Refer to
on-line help for the latest changes to our documentation.
Table of Contents
Part I: Expert Choice Quick Start Guide.................................................... 1
Introduction / Installing / Starting Expert Choice ........................................ 3

Installing Expert Choice........................................................................... 3
Starting Expert Choice ............................................................................ 4
Building An Expert Choice Model ............................................................. 5
ModelView ............................................................................................. 6
Build the Model by Entering Objectives in the TreeView ............................ 8
Enter the Alternatives in the Alternatives Pane.......................................... 8
Making Judgments / Pairwise Comparisons .............................................. 9
Making Verbal Judgments for the Alternatives .......................................... 9
Calculated Priorities ................................................................................ 11
Making Numerical & Graphical Judgments for the Alternatives ................... 12
Making Judgments for the Objectives ....................................................... 13
Viewing Your Results - Synthesizing ........................................................ 14
Ideal and Distributive Synthesis Modes .................................................... 15
Sensitivity Analysis ................................................................................. 16
The Ratings Approach ............................................................................ 18
Create a Ratings Model........................................................................... 19
Defining a Ratings Formula and Ratings Intensities................................... 20
Defining a Ratings Formula ..................................................................... 20
Defining Rating Intensities ....................................................................... 20
Prioritizing Ratings Intensities .................................................................. 22
Rating Alternatives with Ratings Intensities............................................... 23
Understanding the Ratings Intensities Just Entered................................... 25
To View Totals........................................................................................ 26
Global versus Active Alternatives ............................................................. 26
To Make an Alternative Active ................................................................. 26
To Make an Active Alternative Inactive ..................................................... 27
Examining Active Alternatives Priorities .................................................... 27
What To Do Next .................................................................................... 28
Part II: EXPERT CHOICE TUTORIALS .................................................... 31
About the Expert Choice Tutorials............................................................ 33

Introduction to Expert Choice................................................................... 35
Overview of Expert Choice Features ........................................................ 36
Group Features ...................................................................................... 37
i
Expert Choice Quick Start Guide and Tutorials Table of Contents
About the Internet and Intranet Features .................................................. 37

About the Keypads ................................................................................. 38
What Expert Choice is used for: ............................................................... 38
Installing Expert Choice........................................................................... 39
Starting Expert Choice ............................................................................ 40
Conventions Used in the Tutorials............................................................ 40
Mouse Conventions ................................................................................ 40
Using the Keyboard to Select a Menu Command ...................................... 41
Shortcut keys ......................................................................................... 42
Standard Editing Keys............................................................................. 42
Lesson 1 Building A Model ................................................................... 43

Overview................................................................................................ 43
Differentiating Between Objectives and Criteria ........................................ 43
Starting / Creating a New Model (File) ...................................................... 44
To Create a Model .................................................................................. 45
Adding (Entering) Objectives to the TreeView........................................... 46
Renaming Nodes and Deleting Nodes ...................................................... 47
Trash Can .............................................................................................. 47
Adding Sub-Objectives in the TreeView.................................................... 47
To enter the first sub-objective under an objective: .................................... 47
To enter the second sub-objective: .......................................................... 48
Saving Your Model ................................................................................. 48
Moving Nodes in the TreeView / Hierarchy ............................................... 48
Copying Nodes in the TreeView / Hierarchy .............................................. 49
Copying a Group of Nodes in the TreeView / Hierarchy ............................. 49
Reverting to an Older Version of your Model............................................. 49
Adding (Entering) Alternatives in the Alternatives Pane ............................. 50
Assessments / Pairwise Comparisons Overview ....................................... 51
Changing the Paired Comparison Type .................................................... 51
Making Verbal Judgments for the Objectives ............................................ 52
Inconsistency ......................................................................................... 57
Examining and Improving Inconsistency................................................... 57
Judgments for the Sub-objectives, Using the Graphical Mode.................... 59
Comparing the Alternatives, Using the Numerical Mode ............................ 60
Once All Judgments are Made and Priorities are Derived .......................... 62
Redraw Commands ................................................................................ 63
Autoredraw............................................................................................. 64
Redraw ................................................................................................ 64
Making a Factor Dormant in the Pairwise Comparison Matrix .................... 64
Synthesis - Getting the Results................................................................ 65
Modes of Synthesis................................................................................. 66
ii
Synthesis Commands ............................................................................. 67

Sensitivity Analyses - Graphs .................................................................. 68
Dynamic Sensitivity................................................................................. 68
Sensitivity Command Bar ........................................................................ 70
Performance Sensitivity........................................................................... 71
Gradient Sensitivity Graph....................................................................... 72
Head-to-Head Graph .............................................................................. 73
Two-Dimensional (2D Plot)...................................................................... 74
Lesson 2 Structuring - Another Way to Build A Model ............................. 76

Overview................................................................................................ 76
Theoretical Foundation of Structuring ....................................................... 76
Building a Model Using Top-down Structuring........................................... 78
Adding / Entering Objectives.................................................................... 79
Editing Objectives ................................................................................... 80
Clustering Ideas...................................................................................... 80
Entering Alternatives ............................................................................... 83
Viewing the Hierarchy ............................................................................. 83
Grouping Nodes in the Hierarchy ............................................................. 84
Building a Model Using Bottom-up Structuring .......................................... 85
Entering Alternatives ............................................................................... 86
Entering Pros and Cons of the Alternatives ............................................... 86
Viewing the All Pros and Cons List........................................................... 88
Building the Hierarchy ............................................................................. 89
Converting the Remaining Pros and Cons to Objectives ............................ 90
Information Documents ........................................................................... 92
Notes ..................................................................................................... 92
Object Linking and Embedding ................................................................ 93
How to View or Execute the Linked and Embedded Documents................. 94
Example One: How to Open an Embedded File. ....................................... 94
Example Two: Another way to open an embedded file............................... 94
Example Three: How to Open a Shortcut (link). ........................................ 95
Lesson 3 - Data Grid and Ratings ............................................................ 97

Overview................................................................................................ 97
Why Use the Data Grid? ......................................................................... 99
When to Use Ratings .............................................................................. 99
Resizing and Moving Around the Data Grid ............................................ 102
Viewing the Totals Column.................................................................... 103
Adding an Alternative to the Data Grid ................................................... 104
Entering Ratings ................................................................................... 105
With the mouse:.................................................................................... 105
iii
With the keyboard: ................................................................................ 106

Copying Formula Types ........................................................................ 107
Deleting a Formula Type ....................................................................... 108
Revert Command ................................................................................. 108
Creating a Ratings Scale....................................................................... 109
Diagonal Pairwise Comparisons ............................................................ 111
Modifying an Existing Ratings Scale....................................................... 114
Selecting Alternatives from Data Grid to Include in the
ModelView ........................................................................................... 116
To perform sensitivity analyses on extracted alternatives:........................ 118
To derive refined priorities for the extracted alternatives by
pairwise comparing .................. 118
Entering an Ideal Alternative.................................................................. 118
Lesson 4 - Other Data Grid Functions .................................................... 121

Step Function Overview ........................................................................ 121
Creating a Step Formula From a Ratings Formula .................................. 122
Steps to Create a Step Formula from Scratch......................................... 124
Increasing or Decreasing Utility Curves Overview ................................... 124
Decreasing Utility Curves ...................................................................... 126
Steps to Create a Decreasing Utility Curve from Scratch ......................... 131
Increasing Utility Curves ........................................................................ 131
Steps to Create an Increasing Utility Curve from Scratch......................... 134
Direct Formula ...................................................................................... 134
Formula Type None .............................................................................. 135
Must Constraints................................................................................... 135
Lesson 5 - Group Capabilities ................................................................ 137

Overview.............................................................................................. 137
The Facilitator's Role ............................................................................ 137
Facilitator vs. Participant Functions ........................................................ 138
What the facilitator can do. .................................................................... 138
What the participant can do. .................................................................. 139
Creating a Group Model ........................................................................ 139
Adding Participants to and Group Enabling your Group Model ................. 140
To enter or change participants names .................................................. 141
Entering Demographic Information ......................................................... 141
Copying Judgments and/or Data when Adding New Participants.............. 141
Deleting Participants ............................................................................. 141
Things to Do Before Opening a Model for Group Use.............................. 142
Allowing New Participants to Open a Model Without Having
Been Pre-defined.................................................................................. 142
iv
Opening a Group Model Using Keypads ................................................. 144

What the Facilitator Does After a Model is Opened ................................. 144
Group Decision Making Using the Keypads ............................................ 145
Using Keypads to Make Verbal Judgments............................................. 146
Using Keypads to Make Graphical Judgments ........................................ 147
Using Keypads to Make Numerical Judgments ....................................... 148
Making Individual Judgments Using the Data Grid With Keypads ............. 149
Entering a Ratings with a Keypad .......................................................... 150
Entering a Step Value with a Keypad ..................................................... 150
Entering Data for a Utility Curve with the Keypad .................................... 151
Entering a Priority with Keypads ............................................................ 152
Facilitator Keypad................................................................................. 152
Opening a Group Model On a Network or Internet ................................... 154
Group Decision Making Using a Network or the Web .............................. 155
Participants' Paired Comparison Judgments........................................... 155
Making Verbal Judgments ..................................................................... 156
Making Numerical Judgments................................................................ 157
Making Graphical Judgments ................................................................ 159
Entering a Ratings in the Data Grid........................................................ 161
Entering a Value in the Data Grid........................................................... 161
When to Combine Judgments and Data................................................. 163
Combining Judgments and/or Data ........................................................ 163
Combining from the Participants Window ............................................... 166
Combining From the ModelView ............................................................ 166
Combining from the Data Grid ............................................................... 167
Access and SQL Database.................................................................... 168
Upsizing from a Microsoft Access Database to a Microsoft SQL
Database................. 168
Accessing and Viewing an Expert Choice SQL 7 Database..................... 169
Executing an Expert Choice SQL Database............................................ 170
Lesson 6 - Web Models and the Internet ................................................ 171

Overview.............................................................................................. 171
System Requirements........................................................................... 171
How Web Models and Expert Choice Work ............................................ 171
Setting Up a Model for Access Over the Internet or an Intranet ................ 172
Put an Expert Choice Model on the Server ............................................. 172
Creating a new System Data Source Name ............................................ 172
Accessing a Model Over the Internet ...................................................... 181
Logging into a model on the Server........................................................ 182
Lesson 7 - Resource Allocation ............................................................. 184
v
Overview.............................................................................................. 184
How to Perform Resource Allocation ...................................................... 184
Discrete Alternative Resource Allocation Optimization............................. 185
Dependency Feature ............................................................................ 190
Activity Level Resource Allocation Optimization ...................................... 192
Appendix - Area Validation Exercise ...................................................... 197

Area Validation - Numerical ................................................................... 200
Validation Questions ............................................................................. 201
Area Validation - Verbal ........................................................................ 202
Results & Conclusion............................................................................ 203
Setting Up the Hardware Receiver, Keypads and Other

Equipment ............................................................................................ 206
What You Need to Do Before Connecting the Receiver........................... 206
Connecting the Radio Frequency Receiver ............................................. 206
Programming the Keypads. ................................................................... 207
Determining the Current Internal Keypad Number................................... 207
Programming the Keypad so the * Key Behaves as an Invert Key ............ 208
Common Keypad Troubleshooting Questions ......................................... 208
How the Hardware and Keypads are Activated ....................................... 209
How the Keypads Are Used................................................................... 210
Selecting Keypad Options ..................................................................... 211
Using Keypads Anonymously ................................................................ 212
Default Facilitator Menus ....................................................................... 213
Appendix of Expert Choice Terminology ................................................. 215

Absolute Measurement ......................................................................... 215
Active Alternatives ................................................................................ 215
Active Participants ................................................................................ 215
Alternatives .......................................................................................... 215
Assessment ......................................................................................... 215
Autoadvance ........................................................................................ 215
Autoredraw........................................................................................... 216
Best Fit ................................................................................................ 216
Bottom-Up Structuring........................................................................... 216
Child Node ........................................................................................... 216
Cluster ................................................................................................. 216
ClusterView .......................................................................................... 216
Comparisons - Modes of ....................................................................... 216
Complete Hierarchy .............................................................................. 216
Covering Objective ............................................................................... 217
vi
Criterion / Criteria ................................................................................. 217

Distributive Synthesis............................................................................ 217
Data Grid ............................................................................................. 217
Diagonal Pairwise................................................................................. 217
Formulas Grid ...................................................................................... 217
Gap Analysis........................................................................................ 217
Geometric Average ............................................................................... 218
Geometric Indicator............................................................................... 218
Global Alternatives ................................................................................ 218
Global Priority....................................................................................... 218
Goal Node............................................................................................ 218
Grandchild ........................................................................................... 218
Graphical Comparison Mode ................................................................. 218
Hierarchy ............................................................................................. 219
Incomplete Hierarchy ............................................................................ 219
Inconsistency Ratio............................................................................... 219
Information Document ........................................................................... 220
Intensity Scales .................................................................................... 220
Invert ................................................................................................... 220
Ideal Synthesis ..................................................................................... 220
Judgment ............................................................................................. 221
Local Priority ........................................................................................ 221
ModelView ........................................................................................... 221
Node.................................................................................................... 221
Note..................................................................................................... 221
Numerical Comparison Mode ................................................................ 221
Objective.............................................................................................. 222
Pairwise Comparisons / Paired Comparisons ......................................... 222
Parent Node......................................................................................... 222
Plex ..................................................................................................... 222
Priority ................................................................................................. 222
Redraw ................................................................................................ 222
Relative Measurement .......................................................................... 223
Ratings ................................................................................................ 223
Sensitivity Analysis ............................................................................... 223
Sibling Node......................................................................................... 223
Step Function ....................................................................................... 223
Structural Adjust ................................................................................... 223
Structuring ........................................................................................... 224
Sub-Objective (Sub-Criterion) ................................................................ 224
Synthesis ............................................................................................. 224
TreeView.............................................................................................. 224
vii
Top Down Structuring ........................................................................... 224

Utility Curves ........................................................................................ 224
Verbal Comparison Mode ...................................................................... 224
What-if ................................................................................................. 224
viii
Expert Choice Quick Start Guide
Part I: Expert Choice Quick Start Guide
1
2
Introduction / Installing / Starting

Expert Choice
The purpose of this guide is to help you start using Expert Choice quickly. To learn
more about Expert Choice refer to the Tutorials and other documents available from
Expert Choices Help Menu.
Installing Expert Choice

To install Expert Choice with Windows 95, Windows 98, Windows NT, Windows
2000 and Windows XP:
Insert the CD into CD drive.
If the CD does not automatically start then from Windows:
1. Select Start, and then select Run.
2. Select Browse; then select the launch.exe file from your CD Rom
directory.
Follow the instructions on the screen. When prompted, type your name,
company name and Expert Choice serial number.
Note: If you do not have a serial number, or do not enter a valid serial number,
then only the trial version features will be enabled.
Tip: In this guide we use the words SELECT and CLICK

interchangeably.
3
Starting Expert Choice

From Windows click the Start key; then select Programs .
From the list select Expert Choice.
From the second list, select Expert Choice.
Alternatively, if an Expert Choice icon has been installed on your desk-top you can
click the icon.
4
Building An Expert Choice Model

After starting Expert Choice a Welcome Dialogue box will appear with a blank
ModelView window consisting of three panes in the background.
To create your new model, Choosing a Vacation Spot, perform the following steps:
Using the Welcome to Expert Choice dialogue box, select the Create new
model radio button and click OK.
Figure 1 Welcome Dialogue Box
5
Note: To open an existing model, select Recent; then select a model from the drop-
down list or select Browse existing files.
Tip: If you don't want use the Welcome dialogue box, select Esc; then
from the ModelView, select File, then New.
Type the name of your new model such as MyVacation, and select Open. If
want to select a different directory navigate to the directory before entering the
model name.
Next you must define your goal.
Enter a description of your goal by typing: Choose the best vacation spot.
If you made a typo, select Edit, Edit node , or alternatively, right-click on the
Goal node and click Edit node. Then fix your error and press Enter.
At this point, you will have a model with only the Goal node shown in the TreeView
pane.
Now you will add objectives and alternatives to the vacation model.
ModelView
When you start Expert Choice the first window that appears is a blank ModelView.
The ModelView is divided into three major sections or panes.
The TreeView (the left pane) displays the Hierarchy. Figure 2 shows this; here
we use the sample vacation model.
6
Figure 2 Sample Vacation Model
By default, nodes in the hierarchy with children (objectives) are displayed next to
yellow circles, and nodes with no children are displayed next to black circles. If a
node has children that have not been assessed, a red dot will appear in the center of
the circle. When all objectives in the TreeView have been assessed, the circles are
replaced by squares that graphically display the priority of each objective.
The Alternatives pane (top-right) shows the active alternatives.
The lower-right pane displays the Information Document, if available, for the
current (selected) node. (This feature is explained in the Tutorials Lesson 2
and in Expert Choice Contents and Index.)
Note: The appearance of the ModelView can be altered by accessing the menu
command Tools, Options, View.
7
Build the Model by Entering Objectives in

the TreeView
Figure 2, on the previous page, shows an example of what your model might look
like. To build a similar model, you will insert your objectives (criteria) for your
vacation. We suggest you limit the number of objectives to four. You might choose
from objectives such as cost, climate, nightlife, shopping, recreational activities and
ease of getting to the vacation destination.
Select Edit; then select Insert Child of the Current Node .
Type over the highlighted text and enter the first objective, Activities. When
done press Enter.
The next node will be displayed; type Arts and Culture. Press Enter.
Repeat the last two steps to enter Sights and Low Cost.
After entering Low Cost a blank node will appear, press Enter or Esc to end
the insertion process.
If you get interrupted in the above process, click an objective; then click Edit,
Insert Sibling of the Current Node to continue entering objectives on the same
level as the current node.
Tip: To enter sub-objectives for a node, click an objective; then click

Edit, Insert Child of the Current Node. To delete a node, click Edit,
Delete Node.
Enter the Alternatives in the Alternatives

Pane
Select Edit; then select Alternative , Insert and type the alternative name,
Orlando. Press Enter. Alternatively right-click in the alternative pane and
select Insert.
8
Repeat the above step to enter San Francisco, then New York. (Try clicking the
Add Alternative button.)
Your model is now complete. Save it by selecting: File, Save.
The next step in the modeling process is to make judgments / pairwise comparisons
to derive priorities for the objectives with respect to the goal and for the alternatives
with respect to each objective.
Making Judgments / Pairwise Comparisons

There are three ways to make pairwise comparisons: verbally, numerically and
graphically. Each will be explained in this guide.
Making Verbal Judgments for the Alternatives

You can make judgments / pairwise comparisons starting with the goal and working
down to the alternatives (top-down) or you can make judgments about the
alternatives before making judgments about the objectives (bottom-up). The bottom-
up approach illustrated in this Quick Start Guide is usually better because the
insights you gain about the tradeoffs among the alternatives will help in making
judgments about the importance of the objectives.
Note: Judgments about the relative importance of objectives are made with respect to
the parent node in the hierarchy (either the Goal or a higher level objective).
Judgments about relative preference of alternatives are made with respect to each
covering objective. Covering objectives are the lowest level objectives against
which all alternatives are judged. Thus, in this model, Activities, Arts and
Culture, etc. are considered covering objectives.
We will first make judgments about the preference of the alternatives with respect to
the covering objective, Activities.
1. Click Activities, the first objective in the hierarchy. If need be, refer to Figure
2.
2. From the menu select Assessment; then select Pairwise. You will be taken to
the Verbal comparison window (the default); see Figure 3. Since you are
9
comparing the alternatives with respect to the objective, Activities, the

judgment type is preference.
Figure 3 The Verbal Comparison Window
Now select the judgment that best describes your feeling by dragging the verbal
scale indicator up or down to the appropriate position. Figure 2 shows our
judgment; it says that Orlando is Strongly to Very Strongly preferred to San
Francisco with respect to Activities. Note: If you prefer San Francisco to Orlando,
drag the indicator down.
There are several ways to move the judgment indicator as described fully in on-line-
help or Lesson 1 of the Tutorials. For now, if you do not automatically advance to
the next comparison to be made, either click in one of the cells in the comparison
matrix after making a judgment, or, click on the word or between words to the right
10
of the slider when making your judgment.
Tip: The numerical representations of judgments displayed in black in

the matrix signify that the row element is preferred to the column
element. Judgments shown in red indicate that the column element is
preferred to the row element. A numerical representation of a judgment
does not necessarily mean that one element is preferred to the other by
the numerical amount. (See the Expert Choice Contents and Index for
an explanation of the meaning of the verbal scale.)
3. Repeat the process above until all comparisons for Activities have been made.
Note: The Inconsistency, shown in the bottom left cell of the matrix, is
explained fully in the tutorial lessons. If the Inconsistency is higher than .1, see
"Examining and Improving Inconsistency" in Expert Choice Contents and
Index.
4. After all judgments are made, you will be prompted to "Record Judgments and
Calculate", select Yes; you will be returned to the ModelView.
Calculated Priorities
Priorities for the alternatives with respect to Activities have been automatically
calculated and are displayed in the Alternative Pane of the ModelView. Figure 4
shows our priorities for the alternatives with respect to Activities.
If the resulting relative local priorities do not adequately represent your feelings, you
can repeat the pairwise comparison process using the verbal mode or switch to either
the graphical or numerical mode. To do this, click one of the three pairwise
comparison tabs , , . Once you select a mode that makes
the most sense for you, use that mode to make all judgments for current set of factors
being compared.
11
Figure 4 Derived Priorities of the Alternatives with Respect to Activities
Tip: You can also calculate priorities from any paired comparison mode
by clicking the Calculate Icon. Refer to Lesson 1 of the Tutorials for
more information about this option.
Making Numerical & Graphical Judgments

for the Alternatives
Click the next objective, Arts and Culture, and repeat the Assessment Pairwise
process by clicking , to continue making verbal judgments. If desired,
12
select either the or tab to make judgments using one of the

other modes.
Note: The Numerical mode operates in fashion similar to the Verbal mode. The
Graphical mode operates by dragging your preferred comparison bar with the
mouse.
Repeat the pairwise comparison process for Sights using the Graphical tab.
Repeat the pairwise comparison process for Low Cost using the Numerical tab.
Tip: When comparing the preference with respect to Costs, the

alternative that costs less is more preferable.
Note: You can derive priorities directly from actual costs by using the Data Grids
Utility Function as described in the Tutorials or in the Expert Choice Contents
and Index.
Making Judgments for the Objectives

Click the goal node .
Select Assessment; then select Pairwise to repeat the paired comparison process
to make judgments about the importance of the objectives with respect to the
goal. Use the Verbal mode.
Instead of responding yes to "Record Judgments and Calculate", click No.
Click the calculate button to display the priorities of the objectives.
If these priorities are what you think they should be, then press Enter or click on
the ModelView tab ; otherwise, click on the verbal comparison tab

to redo some or all of your judgments.
Now you should have made judgments for all factors (objectives and alternatives) in
your model. Next we show you how to get your results and perform sensitivity
analyses.
13
Viewing Your Results - Synthesizing

After all the judgments in your model have been made and priorities have been
calculated, a synthesis is automatically performed. When focus is returned to the
ModelView the priorities for the alternatives are shown in the Alternatives pane; see
Figure 5. The priorities of the objectives are also shown in the TreeView in both
graphic and numerical form.
Figure 5 ModelView Showing the Synthesized Results with respect to the Goal
14
To examine the synthesis:
Select Synthesize, With respect to the Goal to produce the display shown in
Figure 6.
To view the details, click the Details tab.
Select Show Totals, By Alternatives to see the breakdown of priorities.
Note: Decision hierarchies with more levels will have additional details. The column
headings of the "Details grid" can be dragged and dropped in different positions to
show different vie ws of the results. If you have a compatible version of Microsoft
Excel, you can use the Edit Pivot Table command to create an Excel Pivot Table.
Figure 6 Synthesis Window
Ideal and Distributive Synthesis Modes

The difference in results obtained using the ideal or distributive synthesis modes is
usually negligible and more of theoretical than practical interest. The Ideal synthesis
mode is the default and should be used, as in this case, when you are choosing one
15
alternative and dont really care about the alternatives you do not choose. For more
information about the Ideal and Distributive modes consult the Expert Choice
Contents and Index.
Click Summary; then click Distributive and if desired look at the details.
Now click Ideal. This step will return you to the default setting.
Now you are ready to examine the graphic sensitivity analyses of your results.
Close the synthesis window and return to the ModelView.
Sensitivity Analysis
The five graphic sensitivity analyses show how the alternatives' priorities change
when the objectives' priorities increase or decrease. Four of the five sensitivity
analyses are shown in Figure 7: Performance, Dynamic, Gradient, and Head to Head.
Figure 8 shows the fifth sensitivity analysis , Two Dimensional Plot.
Figure 7 Sensitivity Graphs: Performance, Gradient, Dynamic and Head to Head
16
Figure 8 Two Dimensional Plot
From the ModelView click Sensitivity-Graphs
Select Dynamic
Drag an objective bar back and forth to see the effect.
Now vary another objective.
Reset your priorities to the original by clicking the Home icon
Click Window; click Open Four.
Click the Performance graph , drag another objective and notice how all
graphs change.
Note: You can dynamically vary three out of five graphs (Dynamic,
Performance and Gradient) by dragging the objective bars.
Experiment changing other graphs. Note that each graph has different menus
and commands.
To view the fifth graph, click the 2D button from the active graph menu bar.
17
For more information about these graphs as well as other commands not
discussed in building a model, making pairwise comparisons, synthesizing
or performing sensitivity analyses, refer to the Expert Choice Contents and
Index. Lesson 1 of the Tutorials will also be helpful.
The Ratings Approach

Overview
If we had a large number of alternatives (possibly hundreds or thousands) to
evaluate, it would not be practical to make pairwise comparisons of all the
alternatives. Instead, a ratings approach could be used to rate the alternatives against
some defined scale(s) using what we call rating intensities. This process of rating
alternatives against intensities, or standards is sometimes referred to as performing
absolute measurement.
We will illustrate this process by:
1. Adding a couple of more alternatives to the model we just created. (The same
process can be used to add hundreds or thousands.)
2. Defining a Ratings formula for each covering objective (lowest level sub-
objective).
3. Defining rating intensities for each Ratings formula.
4. Deriving priorities for the intensities.
5. Rating each of the alternatives using the intensities defined and prioritized in
steps 3 and 4.
18
Create a Ratings Model

If need be start Expert Choice and open your vacation model by:
Selecting File; then from the bottom of the drop-down list that displays the last
four models opened, click 1 for the first file on the list.
Make a copy of this model by:
Selecting File, Save As; then type a new filename, MyVacationRatings.
From the ModelView, select Go, Data Grid, or click the Data grid tab.
Figure 9 shows the Data Grid with the alternatives in the rows and the covering
objectives in the columns. (To change how the Data Grid is displayed, use Tools
Options View command and refer to Expert Choices Contents and Index.)
Figure 9 Data Grid
19
Click the first blank cell below New York in the Alternatives column, and type
another place such as Hawaii. Repeat for one or two more cities.
Defining a Ratings Formula and Ratings

Intensities
Defining a Ratings Formula
From the Data Grid, right-click on the cell with the first covering objective,
Activities.
Select Formula type , Ratings. When asked if is ok to replace any pairwise

comparisons, click OK.
Note: In practice when developing a model you will either make pairwise
comparisons for the alternatives with respect to a given covering objective OR
select a formula type, not both.
You should now be at the Ratings Intensities Definition window where you will
define intensity names and derive priorities for the intensities using pairwise
comparisons.
Defining Rating Intensities

Figure 10 shows three intensities (Many, Some, and Few) that we defined for the
covering objective, Activities.
Tip: It is usually best to have defined at least five intensities per factor
to better reflect a wider range of preferences.
For brevity we have only defined three.
20
Figure 10 Activities Intensities Definition Window
Now type one intensity name per row.
TIP: Intensities, if previously defined in your model, will also appear in

a list to the right of the priority; this is only shown when you are
positioned at a blank intensity name cell. If you want to use an intensity
name previously used in another Ratings formula click an intensity in
the list.
Now that you have defined a Ratings Formula and associated intensities you are
ready to prioritize the intensities.
21
Prioritizing Ratings Intensities

1. Now that you have entered the last ratings intensity in the Ratings Definition
window, click Assess to make pairwise comparisons.
You will be taken to one of the three pairwise comparison modes. For example,
if the last comparison mode used was Graphical it would be displayed. To select
another mode, click the appropriate tab.
2. After making pairwise comparisons for the intensities, you will be asked to
record your judgments, click Yes. You will be returned to the Ratings Intensities
Definition window that will now display the derived priorities. Figure 10 shows
our derived priorities.
If you said No, you would stay at the pairwise window so you could change
judgments. After making changes, click the Calculate icon. Then the new
priorities will be shown; next press Enter to return to Ratings Intensities
Definition window.
Figure 11 shows our derived priorities; now make your own comparisons.
22
Figure 11 Ratings Intensities Priorities for Activities
3. If desired, you can sort the intensities by priority, by selecting the Sort button
(They will be sorted automatically when you are returned to the Data Grid.)
4. Click the Close button to return to the Data Grid.
Now you are ready to rate each alternative with respect to Activities.
Rating Alternatives with Ratings Intensities

Now you will rate each alternative with respect to the covering objective, Activities,
for which a formula type of Ratings has been defined.
Figure 12 shows a sample Ratings Intensity Scale on the Data Grid. The scale is
located above the data entry area. This figure also shows the information document
about the Activities objective; to learn more about this refer to Expert Choice
23
Contents and Index, or Tutorial 2. Figure 12 also shows our ratings that will be
explained shortly.
Figure 12 Sample Ratings Intensities Shown on Data Grid
Use your scale when entering your own Ratings.
First click the cell to be rated.
Enter your rating in any one of the following five ways:
1. Click on a Rating Intensity such as Medium. The Ratings intensity names

are listed in line one of the scale.
2. Type the first letter of the Ratings name; if the first letter is not unique
enter the second letter and so on until the Rating is entered.
24
3. Type a number in the cell to designate an intensity, such as 2 for Medium.

Numbers, short-hand notation for an intensity name, are shown on line two
of the Ratings Scale. The number in parentheses represents the derived
priority.
4. Click and drag a bar in a cell.
5. Type a number between 0 and 1 to assign a priority, such as .345 Don't

forget the decimal point.
Tip: To enter a value between two intensities type a decimal value

between the two priorities shown in the Ratings scale.
Note: In practice you would build a model, make pairwise comparisons for each
objective with respect to the goal and, if applicable, define each covering objectives
formula such as Ratings. Then you would rate each alternative with respect to its
covering objective formula/standard. This is especially true if you are using Expert
Choice Group in a team setting.
Understanding the Ratings Intensities Just

Entered
Look at the cells in the Data Grid where you have just entered ratings. The words
represent the intensity name, while the bars represent the priority. If you entered a
value between zero and one, or dragged a bar then the numeric representation is the
value entered.
Figure 12 on page 24 shows our entries. The ratings entered for Orlando and New
York are the results of dragging a bar. The rating entered for San Francisco was
made by typing "S" for some.
To view all values select View, Values.
Then switch back to words representing Ratings intensities by selecting View,

Data.
You can, of course define ratings formulas for the other covering objectives
columns. You can also copy formulas from one covering objective to another. (For
more information, see Expert Choice Contents and Index as well as the Tutorials).
25
To View Totals
Select View, Totals column to see priorities for the alternatives.
Right-click on the Totals column; select Normalize and examine each of the
normalization options. The ratios of the priorities remain the same regardless of
the normalization.
Make sure the Ideal Mode is shown in the top left cell of the grid; if Distributive
is shown, click the cell to switch to the Ideal mode.
Select Totals, Normalize, Priority and jot down the priorities for one or more
alternatives to compare to what you will see in the section entitled "Examining
Active Alternatives Priorities".
Before ending this Quick Start Guide we explain the difference between global and
active alternatives.
Global versus Active Alternatives

Global Alternatives refer to all alternatives shown on the Data Grid. Alternatives
can only be deleted from the model from the Data Grid. If you want to include a
subset of the global alternatives in the ModelView's Alternative pane you must select
and extract them
Active Alternatives refer to the subset of alternatives shown in the ModelView's

Alternative Pane. When alternatives are added to the model from the Alternatives
Pane they are automatically added to the Data Grid. However, alternatives added to
the Data Grid do not appear in the ModelView until they are made "active".
To Make an Alternative Active

Check-mark the desired alternative by right-clicking (check them all now).
Extract the selected Global alternatives to the hierarchy by selecting Edit,

Extract to Hierarchy.
26
Now you will be returned to the ModelView where you can examine the
extracted alternatives.
To Make an Active Alternative Inactive

Repeat the above process to remove the check-mark, or
From the ModelViews Alternative pane, right-click on the alternative to be

inactivated; then select Inactivate.
Examining Active Alternatives Priorities

From the ModelView, click the Goal.
Observe that all the alternatives priorities are shown in the Alternatives pane. They
will be close to, or exactly the same as those displayed on the Data Grid. They will
be exact if at least one cell in every column of the Data Grid has a priority of 1; see
Figure 13. This is true for pairwise comparisons, Ratings, and the other Data Grid
Formula types. If this is not the case, you can use the Ideal alternative option
explained in Expert Choice Contents and Index and the Tutorials to achieve the
desired normalization.
27
Figure 13 Each Covering Objective has an Ideal Alternative
One reason for extracting alternatives is to perform sensitivity plots, which you can
do now if you like.
Another reason for extracting alternatives is to refine the priorities by making

pairwise comparisons as explained in the Expert Choice Contents and Index, as
well as the Tutorials.
What to Do Next
To learn more about Expert Choice including how to enter data and define formulas
such as Ratings, Step Functions, and Increasing and Decreasing Utility Curves refer
to the Tutorials that can be accessed from the Expert Choice Help Menu; sample
models are found in a directory under the root called ecsamples. A brief description
of each tutorial follows.
Lesson 1: Quickly review how to build a model.
28
Lesson 2: Learn another way to build a model using Structuring.
Lesson 3: Review what you learned about Ratings.
Lesson 4: Learn how to use the other Data Grid formula types.
Lesson 5: Learn how to enable a model for use with a group using Expert
Choice Group. This includes learning how to use Expert Choice
Group with an audio-response system using radio frequency
keypads. This feature provides you with a portable decision
conference room.
Lesson 6: Learn how to access a Web model using Expert Choice Group.
Lesson 7: Learn how to use the Resource Allocation component to select a

combination of alternatives that maximize (optimize) the
attainment of objectives while not exceeding a specified budget.
Only available in the group versions of Expert Choice.
Other modeling techniques especially adapted for use with Expert Choice can be
found in the book Decisions By Objectives, available on the Expert Choice CD.
This ends the Quick Start Guide. To learn more about Expert Choice don't forget to
visit www.expertchoice.com.
29
30
Expert Choice Tutorials
Part II: EXPERT CHOICE TUTORIALS
31
Expert Choice Tutorials
32
Expert Choice Tutorials Introduction
About the Expert Choice Tutorials

The purpose of the tutorials is to teach you how to use the major features of the
Expert Choice software. Consider the decision process involved in purchasing a
laptop computer. When people buy a laptop, they usually have several objectives
they want to satisfy before they will be happy with their selection. Examples of
these objectives may be to maximize performance and reliability, while minimizing
cost. Performance might include sub-objectives of speed, capacity, display quality,
storage capacity, etc. In the following tutorials you will learn how to use Expert
Choice to evaluate alternatives and measure their relative benefits to the decision-
maker(s) objectives.
Lesson 1, Building a Model, will guide you through building a

hierarchical model with objectives, sub-objectives and alternatives,
viewing the overall results (synthesis) and performing sensitivity
analyses.
Lesson 2, Structuring, illustrates other ways to build a model. With
the top-down approach, you cluster objectives and then add
alternatives. With the bottom-up approach you convert alternatives'
pros and cons to objectives.
Lesson 3, the Data Grid, explains the grid and demonstrates
Ratings, the first of several different approaches for applying values
to a model.
Lesson 4 continues to explain the remaining Data Grid approaches
including:
Step Functions, Increasing and Decreasing Utility Curves, and
Entering Priorities Directly.
Lesson 5, is for those who purchased Expert Choice Group. It
shows you how to use the group version in a team setting using
either a local area network or the Internet.
Lesson 6, Web Models and the Internet is a step-by-step guide that
demonstrates how to access the Web models.
Lesson 7, the Resource Allocation component is a way to select a
combination of alternatives that maximize (optimize) the attainment
of objectives while not exceeding a specified budget.
Appendix Area Validation Exercise shows you how Expert
Choice converts seemingly meaningless words into accurate ratio
scale priorities.
33
We recommend that you perform each exercise in order. Lesson 2, Structuring, can
be bypassed although it is highly recommended. You can perform the area validation
exercise after performing Lesson 1. Lessons 3, 4 and 5 describe various functions of
Expert Choice; at the beginning of each section we reference models that are found
in the sample models directory (ecsamples) under the root.
We also recommend that you use the Expert Choice Contents and Index found on the
Help Menu when using Expert Choice as this document contains information not
presented here.
34
Introduction to Expert Choice

Expert Choice (EC) software is a multi-objective decision support tool based on the
Analytic Hierarchy Process (AHP), a mathematical theory first developed at the
Wharton School of the University of Pennsylvania by one of Expert Choice's
founders, Thomas L. Saaty. The AHP is a powerful and comprehensive methodology
designed to facilitate sound decision making by using both empirical data as well as
subjective judgments of the decision-maker.
The AHP assists with the decision making process by providing decision-makers
with a structure to organize and evaluate the importance of various objectives and the
preferences of alternative solutions to a decision.
Following are the steps used in AHP and EC:
Brainstorm and structure a decision as a hierarchical model
Pairwise compare the objectives and sub-objectives for their importance in

the decision
Pairwise compare the alternatives for their preference with respect to the
objectives, or assess them using one of the following:
Utility curves, Ratings or Step function, or enter priorities directly
Synthesize to determine the best alternative
Perform sensitivity analysis
Expert Choice has a unique method of using pairwise comparisons to derive

priorities that can more accurately reflect your perceptions and values than any
other way. Expert Choice synthesizes or combines the priorities that you derive for
each facet of your problem to obtain the overall priorities of your alternatives. By
performing "what-if" and sensitivity analyses, you can quickly determine how a
change in the importance of an objective would affect the alternatives of choice.
If the results of your decision model differ from your intuition; you can modify the
model and/or judgments until the model incorporates your intuition. Then the model
results will either change to conform to your "gut" feeling, or your intuition will
change based upon your modeling. In the former case, you will not only have
verified your "gut" feeling, but you will have a detailed justification, if one is
35
required. In the latter case, you will have learned something and avoided a costly
mistake.
Expert Choice with group capability provides for the synthesis of different peoples'
judgments. Expert Choice and Expert Choice Group are also useful for forecasting,
assessing risk and uncertainty, and deriving probability distributions.
Overview of Expert Choice Features

These are some of the features available in Expert Choice.
User friendly displays that make decision model building straightforward

and simple
ModelView containing either a TreeView or ClusterView presentation of

the decision hierarchy (objectives, sub-objectives), alternatives, pros and
cons, and rich text end-user documentation in the form of Information
Documents.
TreeView with automatic expansion and collapsing of branches including:
Long objectives/names of the criteria to allow for in-depth explanations
Drag and drop manipulation of objectives and sub-objectives in the

TreeView or ClusterView
Unlimited Levels in Hierarchy Build larger, more complex decision

models
Information Documents Easily accessible from most windows, enabling

you to import videos, sound or text or even run other programs.
Notes Another form of documentation, where each group participant can

enter their own comments related to the decision process
Data Grid Approaches
Increasing and Decreasing Utility Curves
Step Function
Ratings
36
Direct Entry of Priorities
Data Conversion The ability to rapidly convert large amounts of raw data
from existing databases to prioritized information to be used for decision-
making.
Printing/Reporting Easily exportable reports to Microsoft Word and

Excel.
Group Features
Internet and Intranet capabilities for group decision-making.
Reply System Keypad capability for group decision-making
Resource Allocation a way to select a combination of alternatives that

maximize (optimize) the attainment of objectives while not exceeding a
specified budget. This module is available for our group enabled versions
for an additional fee.
About the Internet and Intranet Features

In addition to the features described above Expert Choice Group is a decision-
support software tool designed to help groups enhance the quality of their decisions
by bringing structure to the decision making process. It enables group members any
place in the world to solve decisions together through the Internet. The Web
capability works as a virtual meeting place allowing for each participant's input to be
officially heard and included in the final decision. This not only reinforces buy-in
from all members but also saves on travel time and costs.
Expert Choice Group helps participants of a group define objectives, sub-objectives

and alternatives and then organize them into a structured hierarchical tree. Once
participants compare and prioritize the relative importance of the decision objectives,
Expert Choice Group is then used to synthesize the group's judgments to arrive at a
conclusion. Then the group can examine how changing the priorities of any objective
affects the outcome by manipulating one of four dynamic sensitivity graphs. By
keeping the participants focused on the problem, Expert Choice Group contributes to
more efficient use of meeting time.
37
About the Keypads

If Expert Choice Group is keypad enabled you can create a group decision room
where members of your team can brainstorm and answer questionnaires. Then the
results can be structured into a decision hierarchy. Using the electronic keypads, up
to 150 people can make judgments about the objectives or sub-objectives and
alternatives of a decision problem. Like the Internet version, the keypad version
keeps the participants focused on the problem, while contributing to more efficient
use of meeting time.
What Expert Choice is used for:

Expert Choice is designed to help individuals and groups in the analysis, synthesis
and justification of complex decisions and evaluations. Thousands of businesses and
government agencies worldwide use Expert Choice for a variety of applications
including:
Resource Allocation
Source Selection
Human Resource Management
Employee Performance Evaluation
Salary Decisions
Formulating Marketing Strategy
Selecting Alternatives
Predicting Likely Outcomes
Analytical Planning
Facilitating Group Decision Making
Benefit/Cost Analysis
38
Engineering Design Evaluations
Production and Operations Management
Policy Formulation and Evaluation
Evaluating Acquisitions and Mergers
Supplier Evaluation
Credit Analysis
Customer Feedback
IT Portfolio Management
Product Pricing Decisions
Innovation Management
Installing Expert Choice

To install Expert Choice with Windows 95, Windows 98, Windows NT 4.0 or
higher, Windows 2000 and Windows XP:
Insert the CD into CD drive. If the CD does not automatically start then from
Windows:
1. Select Start, and then select Run.
2. Select Browse; then s elect the launch.exe file from your CD Rom
directory.
Follow the instructions on the screen. When prompted, enter your name,
company name and Expert Choice serial number.
Note: If you do not have a serial number or do not enter a valid serial number,
then only the trial version features will be enabled.
39
Tip: In this guide we use the words SELECT and CLICK

interchangeably.
Starting Expert Choice

From Windows click the Start button; then select Programs .
From the list select Expert Choice.
From the second list, select Expert Choice.
Alternatively, if an Expert Choice icon has been installed on your desk-top you can
click the icon.
Conventions Used in the Tutorials

Window refers to one of several EC windows:
ModelView
Data Grid
Formulas
Pane is a portion of a window. A window can have several panes that change as
different portions of Expert Choice are accessed. For example, the ModelView
has 3 panes.
Mouse Conventions
If you have a multiple-button mouse, the left mouse button is the primary mouse
button, unless you have configured it differently. Any procedure that requires you to
click the secondary button will say, right-click.
40
Click Items following the word "Click" will appear in bold typeface.
Select, then Press the mouse button once on the menu bar item. Move the
cursor to the specified item within the sub-menu and press the
mouse button again or slide the cursor to the specified item and
release the mouse button.
For example: Select File, then Open. Then move the mouse to the
word File on the menu bar and click once. Move the mouse to the
word Open on the drop-down menu and click once.
Drag Press and hold the left-mouse button on the indicated item as you
move the mouse to the designated location, dragging the item with
the mouse pointer, and then release the mouse button on the
desired destination.
Type When you are required to enter data, we use the word "Type",
followed by what is to be entered in bold print; for example: Type
Performance
Scroll Bars Click and drag the arrows on a scroll bar, located at the right-side
or bottom of a window, to scroll within a window.
Using the Keyboard to Select a Menu Command

Special keys are in bold and should be pressed to perform an action include: Enter,
Esc, F1, Tab, Alt.
For example, you can press Alt plus F to select the Files command from the menu
bar. This will drop the file menu down and you can then select the item you want by
again pressing one of the underlined keys.
Obtaining Help Press Alt H, or press the function key, F1.
Hot Keys Most menu items have hot keys that are underlined and can be invoked
by pressing the Alt key in combination with the hot key. The hot keys may not be
underlined in Windows 2000 until the Alt key is pressed.
41
Shortcut keys Shortcut keys exist for some, but not all of the menu commands.
You can see the shortcuts when a drop-down menu is selected using the keyboard or
the mouse.
Standard Editing Keys

Text can be edited using the same keys you would use for a typical word processor.
arrow keys move the cursor

horizontally and vertically
(, , , ) through the text
Delete deletes character immediately to the

right of the cursor
backspace deletes character immediately to the

key () left of the cursor
Home moves the cursor to the beginning of

the line
End moves the cursor to the end of the

line
42
Expert Choice Tutorials Lesson 1 Building A Model
Lesson 1 Building A Model

Overview
This lesson describes how to build an Expert Choice model. In this lesson you will
learn how to insert objectives, sub-objectives and alternatives into the decision
model, make judgments (paired comparisons), synthesize to get your results and
view sensitivity analyses.
Building a hierarchy is as much an art as it is a science. Following are guidelines that

you need to keep in mind before beginning to build any model.
Guideline 1: Try not to include more than nine elements in any cluster because
experiments have shown that it is cognitively challenging for human beings to deal
with more than nine factors at one time and this can result in less accurate priorities.
Guideline 2: Try to cluster elements so that they include elements that are
"comparable", or do not differ by orders of magnitude. In other words, try not to
include items of very small significance in the same cluster as items of greater
significance. The purpose of a hierarchy is to cluster the more important elements
with each other and the less important elements with each other.
By keeping these two simple guidelines in mind, you will be able to model complex
decisions correctly and efficiently.
Differentiating Between Objectives and

Criteria
In practice, the words objectives and criteria are used interchangeably. An objective
is something that is sought after or aimed at, whereas, a criterion is a principle or a
standard that an idea or objective is judged by. When dealing with choice models, we
will use the word objectives rather than criteria because it better describes the
purpose of the evaluation to select the alternative that best meets an individuals or
groups objectives.
43
Starting / Creating a New Model (File)

In this lesson we are going to build a new model. Figure 1 shows our ModelView
window consisting of three panes that contains our completed model. If you make
the same judgments that we suggest during this lesson then your model will look like
this.
Figure 1 Lesson One Sample Model
TreeView pane displays the hierarchical listing of the objectives and sub-
objectives. This is where you will enter the objectives and sub-objectives of your
decision. Each element in the TreeView is also referred to as a Node.
Alternatives pane displays the active alternatives. Alternatives can be added to

your model from this pane or from the data grid (the latter to be explained in Lesson
3).
44
Information document pane displays information about the highlighted objective

or alternative in either the TreeView or Alternatives pane.
To Create a Model
Click the windows Start key, select Programs ; select Expert Choice program
group and then select Expert Choice.
Tip: If you have already created an icon for Expert Choice on your
desktop, you may double click the icon to start.
Then a blank ModelView window consisting of three panes will appear with a
Welcome to Expert Choice dialogue box superimposed over the ModelView.
Using the Welcome to Expert Choice dialogue box, select the Create new
model radio button and click OK.
Note: To open a recently opened model, select Recent; then select a model from
the drop-down list or select Browse existing files.
Tip: If you don't want use the Welcome dialogue box, select Esc; then
from the ModelView, select File, then New.
A dialogue box will appear, that prompts you to type a file name. Type Laptop;
then click Open.
Another dialogue box will appear prompting you to "Enter a description for your
goal." Type To purchase a laptop computer; then click OK.
Notice your model has no elements below the goal; this is indicated by the black
circle next to your goal description. Now you will enter the objectives and sub-
objectives of your decision model.
45
Adding (Entering) Objectives to the

TreeView
To enter your first objective:
Select Edit, then Insert Child of Current Node, then a node will appear below
the Goal.
Type your first objective, Performance: a laptop that is fast and holds a lot of
data, then press Enter.
Tip: A quick way to insert an objective is to right-click on the Goal and

select Insert Child of Current Node from the drop down menu. Once
you have entered an objective, you may right-click on it and select
Insert Sibling of Current Node to enter other objectives on the same
level.
Notice that the black circle next to the Goal changed to yellow after you have
entered the first objective; this means that there are items below the goal.
Type your second objective, Reliability: a laptop that is durable, then press
Enter.
Notice when you add the second objective the yellow circle, to the left of the
goal node, now includes a red dot in its center; this means that the objectives
below the goal have not been assessed.
We will now continue entering in our objectives. Notice that we briefly define
our objectives to help clarify what we are trying to accomplish.
Type your third objective, Customer Service: easily accessible Technical

Support, then Enter.
Type your fourth objective, Price: affordable with some bells and whistles;
then press Enter.
Press Enter to stop inserting children.
46
Renaming Nodes and Deleting Nodes

If you make mistakes while entering nodes or feel that you have added too many,
move to the node that you would like to either be renamed or deleted and do either
of the following:
Select Edit, then select Edit Node to rename a node, or
Select Edit, then select Delete Node to delete a node.
Trash Can
Trash receives nodes that have been deleted from the model (Edit, Delete Node) or
copied using Edit, Copy Ple x to Trash Can. The trash only appear on the
ModelView menu bar when nodes are in trash.
From the ModelView, select View, Trash Can or alternatively, click the Trash
button.
Once items are in the trash can they can be dragged and dropped back into the
hierarchy by positioning the cursor over the node that will receive them.
Adding Sub-Objectives in the TreeView

In this tutorial, we will enter sub-objectives for only one of the top-level objectives.
In a "real" decision model you might add sub-objectives for most top-level
objectives. To illustrate how to do this, we use our objective: Performance: a laptop
that is fast and can hold a lot of data.
To enter the first sub-objective under an objective:

Click the objective, Performance, and it will become highlighted.
Select Edit, then select Insert Child of Current Node.
Type Speed, then press Enter.
47
To enter the second sub-objective:

Type Size of Hard Drive , then press Enter.
Press Enter again to stop inserting children.
Saving Your Model

Since some of your model is stored in memory versus the database, it is important
to continuously save your model additions or modifications; to do so:
Select File, then Save. Alternatively, you can click the save icon from the button
bar or use the shortcut keys Ctrl S.
Tip: If you terminate your Expert Choice session now by selecting File,
Exit you can easily retrieve the most recently accessed models by
selecting one of the entries in the bottom portion of the File, Open
menu dropdown list.
Moving Nodes in the TreeView / Hierarchy

You have the ability to re-order the nodes in the hierarchy any way you want. To
move nodes follow the steps below.
Press and hold the left mouse button while positioned on the node (objective) or
its description that you wish to move elsewhere in the hierarchy.
Drag that node over another node in the hierarchy where you would like it to
appear, until the receiving node is highlighted; then release the mouse. The node
will become the first child under the node on which it is dropped. During the
move a white notepad will be dragged from one portion of the hierarchy to
another.
If you want the nodes of a cluster to appear in alphabetical order, select the
parent of the cluster and then select Edit, Sort Cluster.
48
To rearrange nodes within a cluster, select each node within the cluster in
reverse order of the way you want them to appear in the cluster and drop each in
turn on the parent node of the cluster.
Copying Nodes in the TreeView / Hierarchy

Follow the same steps above except press the Shift key, then press and hold the left-
mouse button. Drag the node to the copy location, as you do this, a notepad will be
displayed.
Copying a Group of Nodes in the TreeView /

Hierarchy
Follow the same steps above except press the Control key, then press and hold the
left-mouse button. Drag the node to the copy location, as you do this, a notepad will
be displayed. Then all the nodes below head of the dragged Plex will be copied
below the destination node.
Tip: An easier way to move and copy nodes is to drag and drop with the
RIGHT mouse button. Then a list of options will appear, select one.
Reverting to an Older Version of your Model

We shall assume that you practiced moving and copying nodes in the hierarchy. To
put them back to their original order:
Select Edit, Re vert.
From the displayed Revert to box, double-click on the option "after save ". This
will restore your model to its form right after you executed the File, Save
command in this lesson. (Should you revert to the wrong option, select Edit,
Revert again and choose another option from the box.)
49
Note: Revert is only available during the current session or if there is an

abnormal termination due to causes such as power interruptions.
Adding (Entering) Alternatives in the

Alternatives Pane
To enter the alternatives into your decision model, do the following:
Click Alternative pane. Then a dialogue box will appear which prompts you to
enter an "Alternative name".
Type the Add Alternative button in the right corner of the DaVinci, then
click OK or press Enter.
Note: For legal reasons, we use the names of mathematicians in lieu of computer
manufacturers. You may wish to enter computer manufacturer names such as
IBM, Gateway, Apple, Dell, or Toshiba.
Repeat the above process to enter the remaining two alternatives, Galileo, and
then Turing.
You can also enter alternatives from the ModelView by:
Selecting Edit, then select Alternative. Now select Insert and proceed as
described above.
Tip: When you see the phrase "Click OK", you have the option of doing
just that or pressing the Enter key.
Now your model is complete.
Save it by selecting File, Save .
50
Next you will learn how to make judgments or pairwise comparisons about the
objectives and alternatives in your model.
Assessments / Pairwise Comparisons

Overview
One of the major strengths of the AHP and Expert Choice is the use of pairwise
comparisons to derive accurate ratio scale prio rities, as opposed to using traditional
approaches of "assigning weights" which can also be difficult to justify. Once your
model is built, the next step is to evaluate the elements by making pairwise
comparisons. A pairwise comparison is the process of comparing the relative
importance, preference, or likelihood of two elements (objectives) with respect to
another element (the goal) in the level above. Pairwise comparisons are carried out
throughout this model to establish priorities.
There are three pairwise comparison assessment modes.
Verbal judgments are used to compare factors using the words Equal,
Moderate, Strong, Very Strong, Extreme. Equal means the two items being compared
are of equal importance to you. Extreme means an order of magnitude about 9 or
10 to 1. Judgments between these words, such as Moderate to Strong are also
possible.
Graphical judgments are made by adjusting the relative length of two bars
until the relative lengths of the bars represent how many times more important one
element is than the other.
Numerical judgments are made using a nine-point scale, represent how

many times one element is more important than another.
Changing the Paired Comparison Type

There are three different paired comparison types, Importance, Preference and
Likelihood. Importance is most appropriate when comparing objectives or criteria.
Preference is appropriate when comparing alternatives with respect to their covering
objective. Likelihood is appropriate when comparing the likelihood of uncertain
events or scenarios, such as in risk analysis.
51
To change the comparison type:
Select Assessment, Type, and then select either: Importance, Preference or

Likelihood.
Making Verbal Judgments for the Objectives

Since judgments about the relative importance of the objectives may depend on the
alternatives being considered, we recommend that judgments be made from the
"bottom up", that is first for the alternatives, then for the sub-objectives, and then for
the objectives. However, for illustrative purposes, we will make judgments "top
down" in this tutorial.
From the TreeView, click the Goal (until highlighted).
Then from the menu, select Assessment; then Pairwise, or simply click the
Verbal tab.
The Verbal pairwise window will appear. Figure 2 shows what the Verbal
comparison window will look like if you make the judgments described below.
52
Figure 2 The Verbal Comparison Window
Tip: If you cannot see the lower portion of the window, enlarge it.
Using the verbal comparison mode, you assess the relative importance, preference or
likelihood of two elements at a time using words from Equal to Extreme. You may
also select "between" Equal, Moderate, Strong, and Very Strong. You should use the
word that most accurately represents your feelings.
The upper portion of the Verbal Pairwise window shows two elements being
compared with respect to their parent. On the right side of the Verbal Pairwise
screen is the verbal judgment indicator. By clicking and dragging the indicator,
either up or down, you can select the judgment that best describes your feelings. The
lower portion of the window shows numerical representations of the verbal
judgments; the current judgment is highlighted. Note for example, a judgment of
Strong, while represented by 5, does not necessarily mean that one element is 5 times
another since the priorities are derived from the set of judgments, rather than one
judgment alone.
53
Now you will make judgments with respect to the objectives. To make your first
judgment you will compare Performance to Reliability with respect to your goal.
Click on Moderate, above the word "Equal" on the Verbal Judgment Indicator.
We are saying that Performance is Moderately more important to you than the
computer's Reliability.
As you make your first comparison, by clicking Moderate, notice that a number
representing your verbal judgment is entered in the matrix and you are automatically
taken to the next paired comparison to be made. (In this lesson the Autoadvance
option is on. It automatically advances you to the next paired comparison. To learn
about deactivation of this and other default options see Tools, Options, Calculation
and refer to Expert Choice Help Contents and Index.)
Tip: In our example, the objective on the top is always more important
than the objective on the bottom. In the real world, if the objective on
the bottom is more important than the top, select the word from the
lower portion of the scale.
Now make the remaining five judgments by referring to Table 1. When finished,
read the additional information about entering verbal judgments before proceeding.
Table 1 Remaining Verbal Judgments under the Goal: to purchase a laptop computer
At any time you may alter your judgments by clicking on whichever judgment box in
the matrix that you would like to change; then make your change using the verbal
scale.
There are several ways judgments can be entered with the mouse when using the
Verbal comparison mode; for further information refer to context sensitive on-line
help by pressing F1 from the Verbal comparison window.
54
One way to make Verbal judgments is to type the numerical representation of the
judgment in a matrix cell. If you want to enter a judgment between two words such
as "Very Strong and Extreme" you would type 8. If you do this, 8 is merely a
numerical representation of the wording "Very Strong to Extreme". Refer to Table 2
Numerical Representations of Verbal Judgments. When the column element in the
matrix (objective 2) is preferred to the row element (objective 1) invert the judgment
by pressing Alt I or clicking the Invert button .
Table 2 Numerical Representations of Verbal Judgments
After you enter your last judgment, a dialogue box will appear asking if you
want to "Record Judgments and Calculate".
o Select No, and you will remain at the Verbal window where you may
change judgments. Once finished reviewing and/or changing your
judgments, click the Calculate button to record and calculate your
priorities and display a graph of the results. See Figure 3.
o If you had selected Yes, the priorities for the objectives would have
been calculated and you will be returned to the ModelView window
without displaying the priority graph.
Tip: When entering judgments, you may click the calculate button as
long as it is colored yellow or green. The calculate button will be: Red
if not enough judgments are made to calculate priorities; Green if all
judgments have been made; or Yellow if you have missing judgments
but there are a sufficient number to calculate priorities.
55
Figure 3 Priorities Window
Notice at the top portion of the Priorities window you have the option to:
Sort by Name sorts your objectives in alphabetical order.
Sort by Priority sorts your objectives in ascending order by priority.
Unsort leaves the objectives in the order in which they are shown in the
TreeView.
If the Normalize box is checked the highest priority will be shown as a value of
1.000 while the other priorities will be displayed as a ratio with respect to the
highest priority.
56
Before proceeding select File, Save to save your model.
Inconsistency
The inconsistency measure is useful for identifying possible errors in judgments as
well as actual inconsistencies in the judgments themselves; this is accessed from the
Priorities Window.
Inconsistency measures the logical inconsistency of your judgments. For example, if

you were to say that A is more important than B and B is more important than C and
then say that C is more important than A you are not being consistent. A somewhat
less inconsistent situation would arise if you would say that A is 3 times more
important than B, B is 2 times more important than C, and that C is 8 times more
important than A.
In general, the inconsistency ratio should be less than 0.1 or so to be considered

reasonably consistent. The priority window also shows how many missing judgments
are in the set of objectives being compared.
Examining and Improving Inconsistency

After judgments have been entered, you may also request suggestions for reducing
inconsistency. This can be done from any comparison mode.
To see how Expert Choice helps identify inconsistencies, do the following from the
Verbal comparison window:
Select Inconsistency, 1st from the menu to identify the most inconsistent
judgment.
Select Inconsistency, Best Fit.
Figure 4 shows our Best Fit, where the judgment between Reliability and Customer
Service should be Equal as indicated by the Best Fit indicator of 1.0, located above
the matrix. Notice we entered a judgment of Moderate that is represented by 3. Your
Best Fit might be different if you did not make the same judgments. (You should
only change an inconsistent judgment if you feel that your initial comparison was in
error and did not truly represent your feeling.)
57
Figure 4 Verbal Comparison Window Showing Best Fit to Reduce Inconsistency
Note: If the Best Fit is displayed in red then the column element of the matrix is
preferred to the row element. To enter an inverted element in the matrix, enter your
judgment and if it is not displayed in red then click the Invert Icon .
You can also view the second, third etc., inconsistent judgment by selecting an
option from the Inconsistency menu. Then, if desired, you can alter the judgment.
Inconsistencies can also be suggested when you have missing judgments.
Tip: To view the "best fit" of any cell right-click in a selected matrix
cell.
58
Judgments for the Sub-objectives, Using the

Graphical Mode
Now you will make judgments (paired comparisons) with respect to the sub-
objectives below Performance using the Graphical comparison mode.
If you made any changes to your model while you were learning how to use the
Inconsistency command you might be prompted to record your judgments. Since you
were experimenting, just select No.
Click the ModelView tab.
Click the objective, Performance.
Click the Graphical tab.
The Graphical pairwise window will appear; refer to Figure 5 when you make
your comparison. Note: In this example, there is only one judgment to be made.
The Graphical mode operates by clicking and dragging either comparison bar
(the blue or red bar) either to the left or to the right.
Tip: By dragging the top bar to the right you are saying that the objective
on top is more important than the objective on the bottom, in proportion to
the relative lengths of the bars. By dragging the bottom bar to the right you
are saying that the objective on the bottom is more important than the
objective above. Its representation is entered into the matrix in red
indicating that the column element is preferred to the row element.
When done dragging a bar you are prompted "Record Judgments and
Calculate?" If you select Yes you are returned to the ModelView.
Select No.
Click the Calculate button to view the Priorities Graph, or
Click ModelView tab to return to the hierarchy.
59
Figure 5 Graphical Comparison Window
Comparing the Alternatives, Using the

Numerical Mode
Since we have now made comparisons for all the objectives and sub-objectives, it is
time to make pairwise comparisons with respect to each covering objective (lowest
level objective/sub-objective) and their alternatives. Now we will illustrate how to
make numerical comparisons.
From the TreeView, click the sub-objective, Speed
Click the Numerical tab
60
A window will appear where you will compare your preference of the alternative
computers with respect to Speed. When finished, a dialogue box will appear,
prompting you to Record and Calculate Judgments. Figure 6 shows what we entered.
Figure 6 Numerical Comparison Window
If you click on the number above the slider bar you will autoadvance to the next
paired comparison. Alternatively, you can drag the indicator bar to make the
next judgment. Then click the next empty cell in the matrix. (Refer to on-line
help to see other ways to enter judgments.)
After all comparisons have been made, you will be prompted to record your
judgments and return to the ModelView.
From the TreeView, select Hard Drive , the next sub-objective to be compared.
61
Select one of the modes of comparison of your choice Verbal, Graphical or

Numerical.
Make your own comparisons and return to the ModelView.
Tip: Remember to save your model often; then if you want to go back to
an earlier version you can use the Edit Revert command.
You can repeat the above steps until all remaining three comparisons are made,
or alternatively, you can:
o Select Go, Next Unassesed (to advance to the next unevaluated set of
elements).
o Select one of the modes of comparison of your choice - Verbal, Graphical

or Numerical.
o Repeat the above process until all judgments have been made for both the
last two objectives.
Then return to the ModelView.

Tip: You can tell when comparisons have been made with respect to a
node, as the red dot in the yellow or black circles of the TreeView will
no longer be present.
Once All Judgments are Made and Priorities

are Derived
Once priorities have been derived, the circles in the TreeView are replaced by
Priority Fills ; they graphically portray the derived priorities (Figure 7). These fills
indicate the amount of importance of each objective or sub-objective.
You can also use the View command to switch between assessed indicators (circles)
and priority fills .
62
Figure 7 Hierarchy Showing Priority Fills
Before proceeding save your work. Select File, Save.
Redraw Commands
There are two redraw commands that are available when at the ModelView or Data
Grid windows (lessons 3 and 4 explain the Data Grid). Each command redraws the
TreeView (hierarchy) around the current node. Other redraw commands are found on
the View Menu; they expand and contract nodes and display only the first level of
the hierarchy.
The redraw commands are useful when viewing large models; you can move to a
node in the hierarchy and view what is relevant without losing your position. When
at the Data Grid you can click a node in the hierarchy and you will automatically be
positioned to its corresponding covering objective on the grid. See lesson 3.
63
Autoredraw
Autoredraw, the default option, automatically redraws the hierarchy about the
current node. When you move to a new node in the hierarchy, Autoredraw will
automatically expand the current node displaying its descendants while collapsing
other branch nodes of the hierarchy. (This feature uses Expert Choice's patent
4,613,946.) Autoredraw is represented on the button bar by the left-hand. It is on
when it is depressed.
Redraw
Redraw, is only available when Autoredraw is off. It manually redraws the hierarchy
about the current node and does not collapse other nodes. It is represented by the
right-hand Redraw button.
Tip: When at the ModelView's TreeView, depressing the first letter of a

node name (or a series of characters) will take you to that node.
Making a Factor Dormant in the Pairwise

Comparison Matrix
Before we tell you how to synthesize to obtain the results, we want to explain how to
make an element (objective or alternative) dormant so it can be excluded from being
prioritized. Alternatively, if an alternative has already been prioritized it can be
excluded from the results.
From any pairwise comparison window:
Press Ctrl and click a factor name (any objective or alternative located in the
rows of the matrix). The row for the factor that you have selected will be
blocked. In addition, any comparison made (or to be made) relating to that factor
will also be blocked. Figure 8 shows this.
Now press Ctrl and click on the dormant factor to make it non-dormant; any
judgments previously made will be shown.
64
Figure 8 An Example of a Dormant Objective
Synthesis - Getting the Results

A synthesis can be done for either the entire model or a portion of your model.
From the ModelView, select Synthesize, then With Respect to Goal. The synthesis
window will appear, showing the results (see Figure 9).
65
Figure 9 Ideal Synthesis
Modes of Synthesis
In Figure 9, notice the Ideal and Distributive buttons. In our example, the Ideal mode
is selected.
Ideal Synthesis is the default synthesis mode. The Ideal mode assigns the full
priority of each covering objective to the alternative that ranks highest under it. The
other alternatives receive a percentage of the priority in proportion to the highest
alternative. The Ideal Synthesis should be used when you are interested in only one
alternative and the remaining alternatives are no longer relevant.
Distributive Synthesis is used when you are interested in prioritizing alternatives

from which you may pick more than one. The Distributive Synthesis distributes the
priority of the objectives among the alternatives, thereby dividing each objective's
priority into proportions relative to the percentage of the alternatives.
Because each synthesis mode combines priorities differently, you will notice that
each mode may yield different, although normally very similar, results.
66
Synthesis Commands
There are two synthesis tabs - Summary and Details.
Summary displays bar graphs of the alternatives' priorities with respect to the node
from which you performed the synthesis (in our example, the Goal). See Figure 9
above.
Details displays a grid showing all the priorities and alternatives with respect to the
goal or objective selected. In our example, since we selected the Goal node, we show
all objectives, sub-objectives and alternatives. Figure 10 shows the details.
Experiment with the other detail options and refer to the on-line help for more
information.
Figure 10 Ideal Synthesis Details
67
Sensitivity Analyses - Graphs

Sensitivity analyses from the Goal node will show the sensitivity of the alternatives
with respect to all the objectives below the goal. It can also be performed from the
nodes under the goal if the model has more than three levels to show the sensitivity
of the alternatives with respect to an objective or sub-objective. When performing a
sensitivity analysis you may vary the priorities of the objectives and observe how
the priorities of the alternatives would change. There are five types of sensitivity
analysis.
Dynamic
Performance
Gradient
Head to Head
Two-Dimensional (2D Plot)
You can open four types of sensitivity analyses at once or each one separately. Each
graph has its own unique menu commands and each sensitivity analysis can be
compared to a "what-if" analysis because the results are temporary. The remainder
of this section explains each graph.
Dynamic Sensitivity
Dynamic Sensitivity analysis is used to dynamically change the priorities of the
objectives to determine how these changes affect the priorities of the alternative
choices. By dragging the objectives priorities back and forth in the left column, the
priorities of the alternatives will change in the right column. If a decision-maker
thinks an objective might be more or less important than originally indicated, the
decision-maker can drag that objective's bar to the right or left to increase or
decrease the objectives priority and see the impact on alternatives. Figure 11 shows
a Dynamic Sensitivity graph.
68
Figure 11 Dynamic Sensitivity
From the TreeView, click the Goal
From the menu select Sensitivity-Graphs, then select Dynamic
Experiment changing the priorities of your objectives by clicking and dragging

the objective bars on the graph. Drag one objective bar back and forth and you
will see the alternatives' priorities change. Also try this for the remaining
objectives. To see the original priorities, click the Home button on the button bar
To view a dynamic sensitivity graph, Figure 12, showing the alternative's bars in
multiple sections:
Select Options, Components
69
Figure 12 Dynamic Sensitivity with Component Option Selected
Sensitivity Command Bar

While viewing any sensitivity analysis, you can hold your mouse over each icon
under the menu bar to show you what it represents. Each graph has different icons as
well as different menus.
Note: You can have multiple graphs open at the same time. However, if desired, you
can close all graphs by clicking the X icon.
70
Performance Sensitivity
To see the Performance Sensitivity graph:
Click the appropriate button on the Dynamic graph button bar. Alternatively,
select Sensitivity-Graphs from the ModelView; then select Performance.
Figure 13 Performance Sensitivity
The Performance Sensitivity analysis, displayed in Figure 13, shows how the
alternatives were prioritized relative to other alternatives with respect to each
objective as well as overall.
To see how the best computer compared to the second and third computers, read the
overall priority from the intersection of "right y-axis" and the overall priority for
each alternative. In our example, DaVinci is approximately .43, Galileo is
approximately .34 and Turing is approximately .24. Note the priorities for the
alternatives sum to one.
To read each objective's priority (based on the decision-makers paired

comparisons), use the left y-axis . For example Performance is about .52 while
Reliability is about .25. Customer Service is about .14 and Price is about .10.
To read the alternative priorities with respect to each objective, read from the "right
y-axis ". In our example, using Reliability, DaVinci has priority of approximately .42;
while Galileo is about .33 and Turing is approximately .24.
71
This graph is also dynamic, so you can temporarily alter the relationship between the
alternatives and their objectives.
Drag any one of the objective bars up or down.
Restore the original priorities by clicking the Home button .
Select the Open Four icon from the Performance menu bar .
Gradient Sensitivity Graph

Make the Gradient Sensitivity graph the active window by clicking the appropriate
button; see Figure 14. This graph shows the alternatives' priorities with respect to
one objective at a time. By clicking the menu command " X Axis", you have the
ability to select which objective appears on the x-axis. The red vertical line indicates
the objective's priority (based on the decision-makers paired comparisons). To
indicate where an objective's priority changes drag the red bar to either the left or
right; this is shown as a blue dashed vertical line.
In this figure you can see that increasing the priority of Price from .09 to .30 changes
the choice of the alternative with respect to Price. When viewing a Gradient graph
you are looking for cross-over points of the alternatives.
72
Figure 14 Gradient Sensitivity Graph
Head-to-Head Graph
Make the Head-to-Head graph the active window.
Figure 15 shows how two alternatives compared to one another against the
objectives in a decision. One alternative is listed on the left side of the graph and the
other is listed on the right. The alternative on the left is fixed while the alternative on
the right can be varied, by selecting a different tab on the graph. Down the middle of
the graph are listed the objectives in the decision. If the left-hand alternative is
preferred to the right-hand alternative with respect to an objective, a horizontal bar is
displayed towards the left. If the right-hand alternative is better, the horizontal bar
will be on the right. If the two choices are equal, no bar is displayed. The overall
result is displayed at the bottom of the graph and shows the overall percentage by
which one alternative is better than the other; this is the composite difference. The
73
overall priority can either be shown based on the objective weights (typical) or un-
weighted.
Figure 15 Head to Head Graph
Two-Dimensional (2D Plot)

From the menu bar of the last graph viewed, click the Two-Dimensional (2D
Plot) button.
This graph (Figure 16) shows the alternatives' priorities with respect to two
objectives at a time. By clicking the menu commands "X Axis" and "Y Axis" you
have the ability to change the objectives being displayed.
The area of the 2D plot is divided into quadrants. The most favorable alternatives
with respect to the objectives on the two axes will be shown in the upper right
quadrant (the closer to the upper right corner, the better the alternative). The least
favorable alternatives will be shown in the lower left quadrant (the closer to the
lower left corner, the less favorable the alternative).
Alternatives located in the upper left and lower right quadrants indicate key tradeoffs
where there is conflict between the two selected objectives.
74
Figure 16 2 Dimensional Plot
There is one additional Options command.
Projection - is used to turn on and off the projection of alternatives.
When on, projection not only shows how preferable the alternatives are with respect
to the two selected objectives but it shows a composite projection line indicating the
preference of each alternative taking into account all the objectives' priorities. The
farther to the right on the line, the better the alternative. (Each alternative has two
circles. The large circles represent how preferable the alternatives are with respect
to the two objectives that have been. The smaller circles along the composite line
indicate the overall preference of the alternatives with respect to all the objectives.)
To close all the graphs, click the X icon on the button bar.
Now save your model by selecting File, Save. Alternatively, click the save icon from
the button bar or use the shortcut keys Ctrl S.
75
Expert Choice Tutorials Lesson 2 Structuring
Lesson 2 Structuring Another Way

to Build a Model
Overview
Expert Choices Structuring component is a freestyle, interactive technique for
building a model. Unlike the way we built our model in Lesson 1, you can think in
terms of clustering ideas that are a result of brainstorming, discussion, research and
the solicitation of views and ideas from others.
There are two approaches to model building in Structuring.
Top-down structuring is better suited for decisions of a strategic nature where the
objectives are better understood or known than the alternatives. Top-level objectives
are identified followed by the identification of sub-objectives.
Bottom-up structuring is best suited for situations where the alternatives are better
understood than the objectives. The pros and cons of the alternatives are used to help
identify the objectives that can then be clustered into groups.
Theoretical Foundation of Structuring

When using Structuring individually or with a group, top-down or bottom-up
structuring provides a framework whereby decision-makers use their knowledge of
the problem to identify and organize the factors (objectives and alternatives) of the
decision.
With a group the facilitator tracks the discussion, entering objectives and alternatives
or pros and cons. Then the facilitator clusters the concepts as the participants direct.
These elements are then assembled into a hierarchical model consisting of

objectives, sub-objectives and alternatives.
Each alternative may have advantages (pros) and/or disadvantages (cons), sometimes
complicated by the fact that the pros of one alternative may be cons of another.
Using Structuring, decision-makers will list positive and negative aspects of each
alternative in the decision model.
76
The group of pros and cons is a pool of ideas that helps identify the important
objectives that need to be met in order to resolve the issue. Since objectives
encompass these ideas, the pros and cons will be used for creating the objectives.
The conversion of pros and cons into objectives means organizing the pros and cons
of each of the alternatives in relation to your main goal or objective. The pros and
cons are evaluated for the alternatives and a list of objectives is developed. Often,
you will then continue to cluster the objectives into additional levels of sub-
objectives.
Structuring provides an intuitive interface to help you create a hierarchical decision

model. This interface includes the following familiar concepts for model building:
pros and cons, objectives, and sub-objectives. Pros and cons are translated into
objectives and sub-objectives. These are actually the same as criteria and sub-
criteria. Generally in the Analytic Hierarchy Process objectives are synonymous
with criteria, and sub-objectives are synonymous with sub-criteria.
You, the decision-maker, are the one who turns pros and cons into objectives. Thus,
it is important to remember that converting pros and cons to objectives is a matter of
thinking about which objective(s) the pros or cons point to. More specifically, using
pros and cons to derive objectives means you must re-conceptualize the pros and
cons so that they are 1) described in value-neutral terms, rather than the positive and
negative terms of pros and cons, and 2) specified in a general form. Making this
conversion is usually easier than it may appear. For example, the pros and cons of
being inexpensive or high priced in deciding which car to buy would be generalized
into the objective of saving money. Another way to think about converting pros and
cons into objectives is to consider how the pros and cons of an alternative can serve
as indicators of what you actually want, which can then be expressed as an objective.
There are two Structuring approaches to model building, bottom-up and top-down.
The terms bottom-up and top-down are not based on AHP but Expert Choice users
should not interpret these terms rigidly for two reasons: 1) in bottom-up structuring,
regardless of how hard you try to stick with a strict bottom-up approach, you still
have to start with some kind of goal in mind and the goal is always represented by
the top level of the decision model; and 2) with regard to top-down modeling,
Structuring provides for a freestyle, interactive "brainstorming" of the various
aspects of your decision situation which is a natural, intuitive approach to building a
decision model, where identifying parts of the problem is a matter of inspiration,
intuition and common sense more often than a formal stepwise process, working
from top to bottom.
Bottom-up structuring means that you are starting construction of your decision
model with the alternatives at the bottom and working "up" towards your goal. Top-
down structuring means, essentially, that you are starting by listing and organizing
77
your objectives. The instructions for top-down structuring emphasize the freestyle
interactive approach to top-down structuring because 1) discussions of induction and
deduction applied to AHP model building are beyond the scope of this manual; and
2) if you know how to use formal applications of inductive and deductive logic, it is
straightforward to implement them in top-down structuring once you are skilled in
using Expert Choice (if, for some reason, you think you must). Emphasizing a
freestyle, interactive approach to top-down structuring is the best way to lead you
into becoming skilled, efficient, and comfortable with the Structuring.
Building a Model Using Top-down

Structuring
Top-down structuring provides a simple format for establishing objectives and sub-
objectives for evaluating alternatives.
Start Expert Choice.
From the menu select File, Ne w.
In the dialogue box type a new file name, Vehicle; then click Open or press
Enter to continue.
Type the Goal Description To purchase the most preferred vehicle.
The TreeView pane now contains your goal; now you must turn on Structuring;
From the menu, select Tools, Options; then select the General tab.
In the Structuring box, select the Enable button; then select Close.
From the menu, select View; then select the ClusterView pane or just click
Figure 1 shows the ModelView window that now consists of a ClusterView

pane; it replaced the Treeview pane used in Lesson 1. Additional buttons appear
in the right corners, over both the ClusterView and Alternatives panes.
Drag your mouse over each button to see what they do.
78
Figure 1 ModelView with ClusterView and Alternatives Panes
Adding / Entering Objectives

Click the Objective/ criterion/ group button, located in the ClusterView pane.
When a dialogue box appears, type an objective ; for this lesson, type Style.
A highlighted box will appear in the ClusterView containing your objective. The
red dot indicates where the next objective will be placed.
Tip: To position an existing objective drag it where desired. To position the

placement of a new objective, click anywhere in the ClusterView; drag a
box with the mouse. Next enter the objective.
Now use the list below to enter the remaining objectives, repeating the same
process from above.
79
Engine Performance
Safety
Price
Horsepower
Acceleration
Crash Test Score
Braking Distance
Initial Cost
Resale Value
Note: When using the Structuring feature, it is not necessary to worry about whether
a new entry is an objective or a sub-objective as you will cluster them later.
If you cannot see the entire objectives name within a box select Tools,
Options.
You should be at the View tab. Look at the second box in the tab.
Change the number of characters to 20.
Select Close. Now you see the entire objectives names; that is if you typed
what we suggested.
Editing Objectives
To edit an objective, double click on it.
Clustering Ideas
Once you are satisfied with the list, it is time to cluster your objectives by identifying
those that belong in the same cluster and then repositioning the objectives into that
80
cluster. An objective that does not fit into any cluster may standalone. If you have
several objectives that do not fit into any existing cluster but belong to the same
cluster, simply add another objective into which these objectives can be clustered.
Follow the steps below to cluster the objectives entered in the ClusterView.
Click Engine Performance, the objective you want to move; it will become
highlighted.
Drag Engine Performance to the right of the list of objectives by pressing and
holding the left-mouse button. Release the mouse when done.
Click Horsepower, it will become highlighted.
Drag and position Horsepower over Engine Performance. Once positioned,

release the mouse button. Now Horsepower is clustered in the Engine
Performance box.
Cluster the remaining objectives listed below, using the same steps just described.
Refer to Figure 2 below to see what our clusters look like.
Acceleration to Engine Performance
Crash Test Score and Braking Distance to Safety
Initial Cost and Resale Value to Price

Tip: Another way to enter an objective is to click anywhere in the
ClusterView and drag the mouse until a rectangular shaped box is
drawn, simultaneously, a dialogue box will appear, prompting you to
enter your objective.
81
Figure 2 Clustered Objectives in Vehicle Model
Tip: You can also cluster objectives by dragging a box around one or
more existing objectives. When the mouse button is released, you will
be prompted for the name of the new clustering objective. (You may
need to drag and arrange the objectives to an empty portion of the
screen before doing this.)
After clustering the objectives you are ready to enter alternatives.
82
Entering Alternatives
Now that the objectives and sub-objectives are clustered it is time to enter
alternatives.Click the Add Alternative button or select Edit Insert

Alternative from the menu.
A dialogue box will appear prompting you to enter an "Alternative name"; type
Clydesdale. (Rather than use actual car names we use the names of horses; feel
free to use your preferred vehicle names.)
Repeat the Add Alternative process to enter the remaining two alternatives:
Arabian and Palomino.
Save what you have done by selecting File, Save.
Now we will view the hierarchy.
Viewing the Hierarchy

To view the hierarchy, click the TreeView button. Or select View from
the menu. Figure 3 shows this.
83
Figure 3 Completed Model After Using Top-down Structuring
Note: You can drag/drop the objectives to change the order of the objectives in the
hierarchy as explained in "moving nodes in the hierarchy" in Lesson 1.
From this point on you would continue to evaluate your model by making paired
comparisons as described in Lesson 1, or you could evaluate your model based on
what you will learn in Lesson 3, using the Data Grid.
Before continuing, select File, Save.
Now we want to show you how to draw a box around several objectives to make a
new top-level objective.
Grouping Nodes in the Hierarchy

If you already have your hierarchy structured and wish to create a forgotten top-
level objective you can do the following:
84
From the ModelViews TreeView go to the ClusterView pane by clicking .
Drag Style and Engine Performance close to each other.
With the mouse drag and draw a box around them.
When prompted type an objective name such as Features.
Tip: To create one objective, just drag and draw with the mouse.
To delete what you just did:
Drag Style and Engine Performance from the group, Features.
Click Features, the empty group; then select Delete.
Building a Model Using Bottom-up

Structuring
In Bottom-up structuring you start constructing your decision model with the
alternatives. This is a good approach when more is known about your alternatives of
choice than your objectives.
To use our patented bottom-up structuring approach, you will first, enter alternatives;
second, enter the pros and cons of each alternative; and third, convert the pros or
cons to objectives.
Start Expert Choice and from the menu select File, Ne w.
Type the file name, Vehicle2.
Type Goal Description To purchase the most preferred vehicle.
85
Entering Alternatives
From the ModelView pane click the Add Alternative button.
Enter your alternatives (Clydesdale, Arabian and Palomino), as described in

the previous section, Top-down structuring.
Entering Pros and Cons of the Alternatives
Click the Pros/Cons pane button over the TreeView pane. Alternatively,
from the menu select View; then select Alternative Pros/Cons pane.
The Pros/Cons pane with three buttons will appear in the ModelView window.
Figure 4 shows our pros and cons for the highlighted alternative. Notice the
highlighted alternative in the Alternatives pane also appears as the title above the
Pros/Cons pane.
Figure 4 The Pros/Cons for the First Alternative
86
Now you will enter pros and cons for the first alternative. If you would like to add
pros and/or cons for a different alternative simply click on another alternative to
make it the current alternative.
With Clydesdale selected, click the Add Pro button. A dialogue box will
appear labeled "Add Pro for Clydesdale"
Type High Engine Performance then click OK. If necessary, drag the pro
under the Pro button.
Tip: To align the entry of a future pro or con in the pane, click with the
mouse where you would like the pro or con to appear before selecting
the appropriate button.
Now enter a con for the Clydesdale .
Select the Add Con button. When the dialogue box appears, type Very
Expensive and click OK
Notice that the cons are displayed in magenta while the pros are in blue.
Add another con, Not very safe
Add the following pros, repeating the same process just explained:
Stylish
Crash Test Score
Fast
Now continue adding Pros and Cons for the remaining two alternatives; for brevity
we will only enter one pro and one con for each alternative.
Click on the next alternative, Arabian. Add the pro, Affordable
Add the con, Slow Acceleration
87
Click on the last alternative, Palomino.
Add the pro, Fuel Economy
Add a con, Terrible Braking
Select File, Save to save your model before viewing a list of all your pros and
cons.
Viewing the All Pros and Cons List

Now that you have entered your alternatives as well as their associated Pros and
Cons, you will look at the list of all Pros and Cons (Figure 5). Next you will convert
the pros and cons on the list to objectives using our patented process.
The way to see this is to click the All Pros/Cons button over the Alternatives
Pane .
Note: The Pros are in blue while the Cons are magenta.
88
Figure 5 Alternatives Pane Showing All Pros and Cons
Building the Hierarchy

Now you will start building a hierarchy by dragging and dropping (as well as
redefining) a pro or a con (from the All Pro/Cons pane) to the TreeView. Since order
is not important, you can start anywhere in the Pro/Con list.
Lets begin with the con, Very Expensive, which is associated with the Clydesdale.
Click Very Expensive .
Drag and drop it on the Goal in the TreeView.
When prompted to re-define the con as an objective, type Low Cost; click OK.
The above process places the objective, Low Cost, below the Goal in the TreeView.
You should now be able to see the beginning of the formation of your hierarchy.
89
Notice how the con, Very Expensive, is now grayed out; this indicates that it was
converted to an objective. Gray can be removed from a pro or con by double
clicking.
Tip: Since alternatives are evaluated based on their preference with

respect to objectives, the wording for cons must almost always be
changed to state the objective (or objectives) that the con points to.
For example, expensive points to low cost as an objective.
Looking at the list of pros and cons, you can see that some pros and cons could be
repeated more than once because they are associated with more than one alternative;
if this is the case, you would not need to convert repeats. Those pros and cons not
used in the conversion process will remain white.
Tip: On the other hand, a single pro or con may point to several
objectives and can be dragged and dropped numerous times. For
example, a pro of size for a large car may point to the following
objectives: Comfort; Carrying capacity; Safety; Fuel Economy; Ease of
Parking, etc.
Converting the Remaining Pros and Cons to

Objectives
Refer to Figure 6 and use it as a reference, practicing converting pros or cons to
objectives or sub-objectives.
This is what we did. First we dragged relevant pros and cons to the hierarchy
redefining them. When we were done we had a list of objectives. Second, we clicked
the ClusterView button and started grouping our objectives. Third, we also
added a few more objectives. Fourth, we returned to the TreeView by clicking
and dragged the objectives in reverse order of presentation to place them in the
hierarchy as shown.
We could have added objectives directly in the TreeView by selecting Edit, Insert
either Child or Sibling of the current node. Then we could have dragged existing
objectives beneath the new objectives. Both choices produce the same results.
90
Expert Choice is a flexible decision support system designed to work the way you
think.
Figure 6 TreeView of Converted Pros and Cons
Once your hierarchy of objectives is complete, select the Alts/Children/InfoDocs

button located over the Alternatives pane. Then the Alternatives pane will
replace the All Pros/Cons pane.
Now you are ready to make paired comparisons as outlined in Lesson 1 or use the
Data Grid that you will learn in Lesson 3.
Before this lesson ends, we want to describe three more features that are available:
Information Documents and Notes and Object Linking and Embedding.
91
Information Documents
Information Documents, referenced briefly in Lesson 1, are displayed in the lower
right-hand side of the ModelView. They typically contain text, entered by you, or a
facilitator in the group enabled verison, which may describe the goal, give additional
information as to why particular objectives or sub-objectives were selected, and tell
us how paired comparisons were made. Information Documents are rich text objects
and can include files such as Microsoft Office Files (Word, Powerpoint, Excel,
Access), audio, pictures and video.
Information documents can be created for each objective, alternative or paired

comparison in the hierarchy.
From the Goal, click the Information button that appears on the button bar
and a blank document will appear.
Type your name.
To close, click the Information button.
Note: If an Information Document exists for the current (highlighted) node, the
Information button will appear as an open book. .
When an information document is open you can drag other documents to it

and/or create shortcuts to other files or programs. Information documents can
also be created and accessed from other windows.
Notes
In addition to Information Documents, Notes can be entered for elements in the
hierarchy, alternatives, individual paired comparisons, and cells in the data grid
(discussed in Lesson 3). When working with group models, each participant can
enter their own notes to express their views, rationale, concerns and the like about
the different parts of the decision process.
To create a note from the current (highlighted) node, click the Note button.
Enter your comments and when done, click the Note button to close.
92
If a note exists for the current node, the bottom portion of the Note Icon will be
yellow instead of completely black. When a Note window is open, you can
drag other documents to it and/or create shortcuts to other files or programs.
Like Information documents, Notes are available from other windows.
To learn more about these features, refer to Expert Choices Contents and Index.
Object Linking and Embedding

You can insert shortcuts (links) or embed files within any Expert Choice Information
Document. This very powerful feature enables you to document the rationale of your
decision process as well as execute other programs.
The process to link or embed files is dependent on your operating system; with
certain operating systems you may only be able to embed. This is what we
recommend.
1. With an open Information Document or Note, position the cursor to the insertion
point within the Information Document or Note.
2. With Windows Explorer navigate to the file to be linked or embedded.
a. For Windows 2000 and other versions:
Drag a rich text file icon such as .xls, .ppt, or .doc from Windows Explorer
into the opened Information Document or Note. Then a link (shortcut) to
the file will be created. This saves space, but the file won't be available on
other computers if the .ahp model is copied.
If you want to "embed the object" (the file), select File, Open from the
Information Document or Note menu. Next navigate to the file to be linked
or embedded and select it.
b. For Windows 98 and earlier:
Drag the file into the Information Document or Note. Then the file will be
embedded in the Information Document or Note: Alternatively select File,
Open from the Information Document or Note menu. Next navigate to the
file to be embedded and select it. Windows 98 only supports embedding.
3. Lastly close the Information Document or Note.
93
Note: You can embed several documents or files in one Information Document or
Note, intermixed with text if desired.
How to View or Execute the Linked and

Embedded Documents
There are several ways to open and execute linked and embedded files.
Example One: How to Open an Embedded

File.
1. From an opened Information Document or Note, double-click on the
embedded file to make the Excel commands available within the Expert Choice
window.
2. Then Expert Choices file menu will disappear and it will be replaced by the
embedded program's menu, in this case Excel.
3. Alter the embedded file using Excel.
4. To close the embedded file and executing program, in this case, Excel, you must
click x in the upper-right corner of the window or press Alt-Spacebar.
Example Two: Another way to open an

embedded file.
1. Open an Information Document or Note and right-click on the embedded
document.
2. When a pop-up list appears, select Worksheet Object, Open. You will be taken
to the embedded program.
Now continue as described above.
94
Example Three: How to Open a Shortcut

(link).
1. Fro m an opened or closed Information Document or an opened Note, double-
click on the shortcut (link).
2. Then you will be taken to that program.
3. To close the program click x in the upper-right corner of the window.

4. Then if need be close the Information Document or Note.
95
96
Expert Choice Tutorials Lesson 3 Data Grid and Ratings
Lesson 3 Data Grid and Ratings

Overview
The Data Grid contains data about the decision alternatives with respect to the
decision hierarchy covering objectives. The alternatives appear in the rows of the
Data Grid while the covering objectives appear in the column headers; see Figure 1.
Instead of deriving priorities for alternatives through pairwise comparisons, as
illustrated in Lesson 1, formulas can be defined under any covering objective
(column). Each covering objective's formula converts data about the alternatives to
values (priorities).
There are five different formula types that can be used in the Data Grid. This lesson
discusses how to use one of these formula types, called RATINGS. Lesson 4
discusses the remaining formula types.
Alternatives that are added to a model from the Alternatives pane in the ModelView,
as described in Lesson 1 and 2, are known as active alternatives and are
automatically added to the Data Grid. Alternatives can also be added directly in the
Data Grid. All the alternatives in the Data Grid are referred to as global alternatives.
Active alternatives are check-marked in the grid (see Figure 1). You can mark and
un-mark active alternatives and then extract them to the Alternatives pane to refine
judgments or perform sensitivity analyses. Extracting alternatives from the Data Grid
will be explained later.
While alternatives appear in the first column of each row of the Data Grid the
covering objectives appear in the remaining column headers, while the ratings appear
in the intersecting cells underneath the column headers.
97
Figure 1 Sample Data Grid
The Data Grid also has "Total" and "Cost" columns that can be made visible or
hidden with the View menu command. The value in the total column for each
alternative (see Figure 1) contains the weighted sum of the values (priorities) for the
alternative across all the covering objectives columns.
Each cell in the body of the Data Grid contains (1) a bar representing the value of the
respective alternative (row element) with respect to the covering objective (column
element) and (2) the underlying data or value (you can ask to see either data or
values).
98
The values in the Data Grid body cells are derived from the data based on one of five
formula types:
Ratings,
Increasing utility curve,
Decreasing utility curve,
Step function, and
Direct data entry.
Why Use the Data Grid?

Applications such as personnel evaluation, resource allocation and college admission
decisions, may involve a large number of alternatives. Making pairwise comparisons
in a model that has a large number of alternatives can be very time consuming, as
you would have to perform a large number of comparisons for the alternatives (up to
n*(n-1)/2) with respect to each covering objective. Instead of making pairwise
comparisons for the alternatives, the data grid can be used.
The use of the data grid combines the power of the hierarchy and the pairwise
comparison process with the ability to evaluate hundreds or even thousands of
alternatives. Pairwise comparisons are still used to evaluate the elements in the
hierarchy itself, but not for evaluating the alternatives. Alternatives priorities are
established relative to each covering objective either by (1) using ratio scaled Rating
intensities (explained below), or (2) using other formulas that can be applied to data
(explained in Lesson 4).
When to Use Ratings

People often "rate" alternatives using words such as High, Medium and Lo w; or
Excellent, Very Good, Good, Fair and Poor. This is often the case for subjective
aspects of an evaluation or when there is no hard data. First we will show you how
to do this and derive ratio scale priorities using rating scales that are predefined in a
sample model and then we will show you how to define and prioritize your own
rating scales. We will also show you how to:
Copy formulas,
99
Delete existing formulas,
Revert to a prior version of your model to undo a formula deletion,
Create a new fo rmula, and
Modify an existing formula.
We will start by using the model called Ice Cream Site.ahp that is found in the
sample models folder. Figure 2 shows the model.
Start Expert Choice as you normally would.
Select File, Open; navigate to the ECSamples models folder (found in a root
directory) and select Ice Cream Site.ahp from the list.
Note: Expert Choice files have the extension of .ahp. If you are upgrading from
version 9 your models will have to be converted.
100
Figure 2 Icecream ModelView
Since we will be explaining how to use Ratings by changing this model, we want
you to make a copy of the model first.
Select File, SaveAs and when the save model dialogue box appears:
a. Click the Up One Level button for the Expert Choice folder
b. Click the Create New Folder button and type a folder name, such as
Models
c. Type MyIcecreamSite and then click Save .
The original model will be closed and the new copy of the model will
appear. The filename of the new model also appears in the "Title" bar.
101
The window shows the copied model with four objectives which have already been
prioritized via pairwise comparisons. Notice that there are no alternatives in the
Alternative pane.
Next we will show you your Data Grid and how to move around it.
Resizing and Moving Around the Data Grid

From the ModelView, select Go; then select Data Grid (see Figure 3). This
model contains several global alternatives and no active alternatives.
Tip: Clicking the Data Grid tab from the ModelView will take you
to the Data Grid. Or, use the Data Grid shortcut key shown on the Go
menu.
You can use the standard window keys to resize the grid as well as individual
columns. Notice the TreeView pane to the right; if you click on an objective in the
hierarchy you will move to the corresponding column in the Data Grid. You can use
the arrow keys or the mouse to move around the grid.
Note: If your TreeView pane is not visible, select View from the menu, then select
TreeView pane to make it visible.
When in the covering objectives data entry area, if the Autoadvance option is on
(the default) you will notice a Down Arrow or a Right Arrow on the Data Grid
button bar. The arrow button designates the direction of Autoadvance when entering
data in the columns of the grid. Clicking on the button will change the direction.
102
Figure 3 Icecream Retail Selection Data Grid
Viewing the Totals Column

To make the totals column of the grid visible, select View, Totals column.
As we mentioned earlier, the value in the total column for each alternative contains
the weighted sum of the values (priorities) for the alternative across all the covering
objectives' columns. (If the elements in the decision hierarchy have not yet been
prioritized, the weights, and hence the totals, will be zero.)
The priorities in the total column shown in Figure 4 not only denote rank but also
have interval and ratio meanings. For example, referring to Figure 4, the priority of
Main Street High Rise is .527, while the priority of the Business District is .217.
From this we can infer that the Main Street High Rise is more than twice as
preferable as the Business District.
To see a more meaningful view of the results of this model:
Click any cell in the Total column
Select Edit, Sor t, Descending and the priorities for your alternatives will be
sorted from highest to lowest.
103
Figure 4 Alternatives Sorted by Total Priority
Next we will show you how to enter a new alternative into the Data Grid; then we
will show you how to rate the alternative.
Adding an Alternative to the Data Grid

For this part of the lesson, we will add one additional alternative to the Data Grid.
104
Click on the first vacant cell in the Alternative column.
Type MyStore as one word, then click the cell below or move down using the
keyboard down-arrow key.
To edit your alternative (for practice):
Move to the cell, Mystore, and press F2, the windows standard editing key.
Move the cursor to " S"; press the space bar once so that your alternative is
typed as "My Store" (two words) and then click the cell below or press Enter.
Entering Ratings
There are several ways that Ratings can be entered either using the mouse or
keyboard.
With the mouse:

Move to Data Grid cell for the My Store Row and the Visibility Column.
Notice how a Ratings scale appeared above the alternatives on the Data Grid
when you moved to a cell in the Visibility column.
Since we will be entering ratings for each cell in this row, make sure the
Autoadvance arrow in the button bar is pointing to the right and not down. If it
is pointing down, then click it once.
Tip: The Autoadvance to the next cell option is on by default. If you

do not see an arrow pointing down or right on the button bar, turn on
Autoadvance, by selecting Tools, Options, Calculation; then select
Autoadvance.
Click OK (the second element in the Ratings scale) and that rating will be
entered for My Store with respect to Visibility.
105
Notice the word "OK" appears in the cell; the bar in the cell indicates the priority
value associated with "OK".
Tip: While at the Data Grid, clicking an objective in the TreeView selects
that column in the Data Grid. If there are sub-objectives associated with
that objective you will be positioned at the first sub-objective. To go to
a sub-objective you must click on it in the hierarchy.
For the Competition covering objective, select Intense.
Suppose you change your mind and think that competition is somewhere
between moderate and intense, say about .33. With the left mouse button
depressed in the cell in the Competition column, drag the bar in the cell to a
value of .33, creating a value between moderate and intense.
With the keyboard:

For the Customer Fit objective, try the following four ways of entering ratings into
the cell:
For Customer Fit, press 1, number representing the first rating intensity, "Lots",
or
Type a value between zero and one such as .345, (if you did this, the entry
would be between few (.303) and some (.550), or
Type 0 for a zero rating, or
Clear a rating, by pressing Delete.
Move to Cost and press 2, the number corresponding an "average" rating.

Typing with the keyboard is often the fastest way to enter ratings in the Data
Grid.
Now that you have learned how to enter Ratings, we will next show you how to copy
a formula from one covering objective to another.
106
Copying Formula Types

To practice copying a formula, such as a ratings scale, from one covering objective
to another, do the following:
Click the Formula Grid tab . Figure 5 below shows a partial Formula
grid.
Click in the cell in the Type column of the Competitors row.
Select Edit, Copy Formula
Click in the Customer Fit row.
Select Edit, Paste Formula. Co mpare your results with Figure 5; it shows the
Data Grid before you copied a formula to Customer Fit.
Click the Data Grid tab and go to the Customer Fit. This column will contain
the copied formulas and data.
Now, if desired, you can make entries for the alternatives with respect to the
copied Ratings scale, Customer Fit. Remember to change the Autoadvance
Right Arrow to point down to facilitate entering the ratings in this column and
enter your ratings as you learned in Resizing and Moving Around the Data Grid.
Figure 5 Formula Grid before Copying
107
Next we will show you how to delete a formula type.
Deleting a Formula Type

From the Data Grid, go to the Customer Fit column, right-click on the word
Ratings; refer to Figure 4.
From the pop-up menu showing all the formula types select None.
Select OK to confirm the deletion.
Revert Command
The Revert command can be used to revert to a previous version of your model. For
example, should you decide you do want to abandon what you just did, do the
following:
From the Data Grid select Edit, then Re vert, or alternatively, click the Revert
button.
When the Revert to window is displayed double-click the desired entry; see
Figure 6 below.
In our example, click Original, and the model as originally loaded will be
restored.
Click Revert again; try another selection. Got the idea?
Note: The "revert to" selections are deleted when a model is closed
108
Figure 6 Revert to Window
Now that you know how to enter alternatives, Ratings intensities, and copy and paste
Ratings formulas, we will show you how to create Ratings scales.
Creating a Ratings Scale

From the Data Grid select a cell in the Customer Fit column.
From the menu bar select Formula Type; then select None to delete the existing
formula.
When the dialogue box appears saying "This will replace any existing data.
Proceed?" Click OK.
From the menu bar select Formula Type then select Ratings to create a new
Ratings scale.
Tip: Instead of using the menu bar, right-clicking in the cell above
Customer Fit will display a drop down menu; then select Ratings.
A new window will appear where you will create a new Ratings scale; see Figure 7.
When you are positioned on a blank cell in the left column of this grid, a list of
intensities defined elsewhere in the model will be displayed so you can, if you like,
109
drag one to the left column of the first vacant row. More typically, you will just type
an intensity name.
Now you will re -create the Customer Fit Ratings scale.
Click in the first vacant Intensity Name column; then click some from the list
on the right
Move down to the next row, and click little from the list on the right
Move down to the next row and type few
Figure 7 Ratings Scales Window
110
Tip: To change an intensity name, click on that intensity, then retype or

press F2 to edit using standard editing procedures. To delete, move to
the intensity name and press the Delete key.
There are three intensities in this scale, each with a zero priority. For a more robust
solution, all intensity scales should have at least 5 elements; we use three here for
brevity.
Click the Assess button.
You will be taken to one of the three paired comparison modes (verbal,
numerical, or graphical); if you like, you can switch modes by clicking on one of
the tabs.
Next we explain how to make diagonal pairwise comparisons.
Diagonal Pairwise Comparisons

We will use the Graphical mode to make pairwise comparisons and will do so only
for the cells on the lowest diagonal of the matrix. Pairwise comparing elements on
the diagonal provides you with the minimum number of comparisons necessary to
calculate priorities. (When you do this you have no inconsistency measure and
priorities will not be as accurate as if you completed all the judgments, but it does
save time and the result is still mathematically valid.)
From the Graphical pairwise mode, select Tools, Options, Calculation.
Notice that the Judgment Order box has three options, Next (described in Lesson
1), Randomize and Diagonal.
If the Diagonal button is not selected, select Diagonal.
Click Close.
Figure 8 shows judgments that you will now make using the Graphical mode.
111
Figure 8 Sample Diagonal Graphical Pairwise Judgments
Drag the bars to (approximately) duplicate our entries.
Did you notice that the Calculate button has turned from red, to yellow ,
after making two entries? When the calculate button is yellow this means not all
judgments have been made, but enough have been made to derive priorities. If
desired you could continue making all judgments; however, in this example:
Click the Calculate button to derive priorities for the scale.
112
After reviewing the displayed priorities (Figure 9), click the Priorities tab ,
or press Enter to return to the Ratings definition window. The window now
shows your derived priorities.
Click Close to return to the Data Grid.
Figure 9 Ratings Scales Priorities Window
You should be at the Customer Fit column. The Ratings Intensities for this column,
Some, Little, and Few, appear above the Alternatives section of the grid. Now use
this scale to rate each alternative, following the directions described earlier in this
lesson. For each alternative type:
113
Little, Little, Little, Some, Some, Some, Fe w
Type your own Rating for My Store.
Next we show you how to modify an existing ratings scale.
Modifying an Existing Ratings Scale

We will modify the Ratings scale for competition as an example.
Right-click the top row of the Competition Column and select Ratings. Next,
you can either:
1. Click Assess to revise your judgments made earlier using the paired
comparison process; or rather than clicking the Assess button,
2. Stay at the Ratings definition window and left-click and drag each
intensity priority bar to the right or left. Figure 10 shows that we reversed
the priority of the intensities by making Intense the highest, followed by
moderate and little.
114
Figure 10 Intensities Derived by Dragging Bars
After making modifications, click Close to return to the Data Grid. The values for
cells in this column will be updated based on the new priorities for the intensities.
You will learn how to create and modify other formula types in Lesson 4.
Next we will show you how to "extract" alternatives from the Data Grid. When
extracted, these alternatives will be displayed in the Alternatives' Pane of the
ModelView, and are then known as active alternatives.
115
Selecting Alternatives from the Data Grid to

Include in the ModelView
Any number of alternatives can be made "active" and extracted from the data grid to
appear in the Alternatives Pane of the Model View. This is done for either of two
reasons. First, to perform sensitivity analyses, and second, to refine priorities by
making pairwise comparisons, replacing the priorities derived by ratings or from
data. You can select and extract as many alternatives as desired. However, since
only 9 alternatives can be plotted in sensitivity analyses, we suggest you activate
nine or fewer. You can then select another set of nine or fewer.
Do the following to extract alternatives:
From the Data Grid, right-click on each alternative to be extracted. Refer to

Figure 11.
From the menu, select Edit; then select Extract Selected to Hierarchy.
When you are prompted to "Extract active alternatives?" click Yes. Then you
will be taken to the ModelView showing the hierarchy and active alternatives.
116
Figure 11 Selecting Alternatives for Extraction
Now experiment with the extracted alternatives by doing sensitivity analyses and
refining your alternatives' priorities by making paired comparisons. For additional
information about either topic refer to Lesson 1.
117
To perform sensitivity analyses on extracted

alternatives:
Click the goal node.
Select Sensitivity-Graphs; then select one of the five graphs such as Dynamic.
To derive refined priorities for the extracted

alternatives by pairwise comparing:
From the TreeView, click to one of the covering objectives in the TreeView
such as Location.
From the menu, select Assessment, Pairwise. Click Yes when prompted to
remove the formula and derived priorities with pairwise comparisons.
Switch to the graphical comparison mode and adjust any of the existing
judgments that appear on the diagonal. Then add additional judgments as
desired.
Note: A judgment of 1" means that the two alternatives received the same
rating; however, you might now think that one is a bit more or less preferable
than the other.
Before proceeding to the next section:
Select Edit, and then Re vert to revert back to the model as it existed "Before
Extracting Global Alternatives".
Entering an Ideal Alternative

An ideal alternative is a conceptual alternative that would receive the highest
possible rating under each covering objective in the Data Grid. It is useful to prevent
"rank reversal" (discussed extensively in the book Decision by Objectives) as well as
to examine how far the real alternatives are from the Ideal. The latter is sometimes
called a Gap Analysis. The "gaps" can be viewed in either the Total column of the
118
Data Grid or from the Performance Sensitivity Graph.
To enter an Ideal alternative on the Data Grid:
From the menu select Tools, Options, General
Select Include the Ideal Alternative in Data Grid (Do not select the "Hide"
option). Then the Ideal Alternative will be inserted in the grid as shown in
Figure 11.
If you want to see the same information graphically, extract the ideal alternative and
up to 8 other alternatives; follow the directions below.
Select the Ideal and up to 8 other alternatives by right clicking their names.
Note: If you right click on a selected alternative it will be deselected.
Extract them to the ModelView's Alternatives' Pane by selecting Edit, Extract

to Selected Hierarchy.
Select Sensitivity-Graphs then select Performance.
119
Figure 12 Ideal Alternative Shown on Data Grid
Most of the commands learned in this lesson can also be applied to the other formula
types to be explained in the next lesson.
120
Expert Choice Tutorials Lesson 5 Other Data Grid Functions
Lesson 4 - Other Data Grid Functions

Overview
Lesson 4 shows you how to use the remaining four formula types: Step Function,
Increasing Utility Curve, Decreasing Utility Curve, and Direct. In lesson 3 we
showed you how to use the formula type of Ratings; much of what was learned in
that lesson applies to the remaining formula types. Therefore, this lesson will only
describe new functions.
Note: Any combination of formula types can be used in the same model.
While it is possible to pairwise compare the active alternatives with respect to

certain covering objectives while at the same time defining formulas such as
ratings, step functions, utility curves and extracting from the data grid, it is more
common in practice to either make paired comparisons of the alternatives OR to use
the data grid.
Another common practice is to make judgments using the data grid and then extract
alternatives to the hierarchy, refining judgments by making paired comparisons.
Step Function Overview

The Step Function is similar to Ratings in that it consists of a scale of prioritized
intensities. However the Step Function automatically calculates the appropriate
intensity for each alternative based on data entered in the Data Grid.
Creating the Step Function is similar to creating a Ratings scale with the addition of
specifying a lower data bound for each step. Like Ratings, once you have created a
Step Function intensity scale, pairwise comparisons are made to derive priorities for
the intensities, or, you can drag the bar in the priority column to derive priorities.
For example, if the intensities were Low, Medium and High, we might want Low to
be applied to each alternative for which the data value was less than 50, Medium
between 50 and 199, and High if 200 or more. The data ranges are specified by
Lower Bound values, in this case: 0 for Low, 50 for Medium, and 200 for high.
121
Creating a Step Formula from a Ratings

Formula
From the Samples models directory, open IcecreamSite; this model was used to
demonstrate Ratings. We will use a copy of this model to illustrate how to create a
Step formula from a Ratings formula.
Select File, SaveAs.
When prompted for the mo del name, type MyIcecreamSite-Cost; then click
Save .
Click the Data Grid tab.
Click the Monthly Rental Cost column.
Then right-click to display the drop-down list; then select Step.
A dialogue box will appear "This will replace any existing data. Proceed?" Click
OK.
Then the Step Scales window will appear; see Figure 1. It shows intensity
names, step lower bounds and priorities.
We will first illustrate converting an existing Ratings Formula to a Step Formula

and then show how to define a Step Function formula from scratch.
122
Figure 1 Step Function Scales Window
From Step LB column starting with cheap, type the following lower bounds: 0,
1000, 2000, and 3000.
If desired you could practice dragging the priority bars or select Assess to view
previously made paired comparisons. Then click the priorities tab to return
to the Step Function Definition window.
Now click Close and return to the Data Grid.
Type the rental cost for each store: 500, 999, 1100, 3005, 2500, 1000, and 3500.
Now we will show you how to define a Step Function from scratch.
123
Steps to Create a Step Formula from Scratch

If you were creating a Step formula from the very beginning you would:
1. Select a covering objective column.
2. Right-click to display the drop down list.
3. Select Step.
4. A dialogue box will appear "This will replace any existing data. Proceed?"
Click OK.
5. Type an intensity name and low bound for each member of the scale. Negative
numbers can be used.
6. Then you would either drag priority bars to derive priorities or select Assess to
pairwise compare the intensities.
7. Lastly, you would enter data in the Data Grid for each alternative with respect
to the covering objective's Step formula just defined.
That's all there is to the Step Function. Next we will show you how to define
and use utility curves.
Increasing or Decreasing Utility Curves

Overview
Utility Curves, like the Step Function, translate data into priorities. Whereas the
Step Function consists of discrete steps, a Utility Curve is continuous. Utility Curve
formulas can be increasing or decreasing, linear or nonlinear, and when nonlinear,
concave or convex.
Before showing how to create a Utility Curve from scratch, we will look at some
existing utility curve functions. From the ECSamples models directory, open
Retirement Places Rated. (This model is based on information taken from
Retirement Places Rated by David Savageau, Macmillian Travel (Simon & Schuster
Macmillian Company) 1995.)
124
Since we will be using this model for the remainder of this lesson make a copy of it
as follows:
Select File, SaveAs
For the model name, type MyRetirement; click Save
If desired review the paired comparisons made for the objectives and sub-
objectives by selecting Assessment, Pairwise.
Click the Data Grid tab and familiarize yourself with the content.
This model (Figure 2) uses three different formula types: Ratings, Decreasing and
Increasing Utility curves, as seen by looking at the top of each column in the Data
Grid. All the curves in this model happen to be linear. Data has been entered for each
alternative with respect to each covering objective.
125
Figure 2 Retirement Places Rated ModelView
Decreasing Utility Curves

From the Data Grid, let's look at the first covering objective of Money Matters/State
Income Taxes. See Figure 3.
126
Figure 3 - Retirement Places Rated Data Grid
Right-click State Income Taxes.
When the menu appears select Decreasing.
Figure 4 shows the decreasing curve. Notice the high and low values for the x-
axis; they are 200 and 75 respectively. Observe that Linear is checked.
127
Figure 4 Decreasing Utility Curve
Change the high x-axis to 250 and low x-axis value to 90.
Click Apply to make this change permanent.
Click Close to return to the Data Grid.
Notice that the Low and High Values, now displayed on your Data Grid changed;
compare it to the Figure 3. Also observe that when you change the definition for a
utility curve, the values and priority bars are recomputed automatically.
128
Return to the Decreasing Utility Curve window by clicking on the Utility Curve
tab.
Drag a point on the straight line down and to the left.
Notice that the linear box becomes unmarked; see Figure 5.
To make this change permanent, click Apply.
Click Close and view the changes to the priority bars on the Data Grid.
Return to the Decreasing Utility Curve window and compare your curve with
ours shown in Figure 5.
129
Figure 5 Decreasing Non-Linear Utility Curve
To change this curve to an Increasing curve, select the Increase button.
Click Close; this change (from a decreasing to an increasing curve) did not take
place because you did not select Apply.
130
Steps to Create a Decreasing Utility Curve

from Scratch
If you were creating a Decreasing formula from the very beginning you would:
1. Select a covering objective column.
2. Right-click to display the drop down menu.
3. Select the formula type, Decreasing.
4. A dialogue box will appear "This will replace any existing data. Proceed?" Click
OK.
5. Type high and low x-axis values for the curve. (Negative numbers can be used.)
6. If the curve desired is not Linear drag the line down and to the left.
7. Select Apply to make the changes permanent.
8. Select Close to return to the Data Grid.
9. Lastly, you would enter data (numbers) for each alternative with respect to the
covering objective's Decreasing formula just defined.
Increasing Utility Curves

Now we will describe how to create an Increasing Utility curve.
Look at the Home Mortgage Payment and Property Taxes covering objective that is
defined with an Increasing Utility Curve. (We are not referring to the same named
sub-objective of Money Matters). According to David Savageau, (author of
Retirement Places Rated ) this objective is defined to have a score between 0 and
100 where higher numbers are preferred to lower numbers; no details of this scale
were given in his book.
To see how we entered the latter information in the model:
131
In the TreeView pane to the right of the Data Grid, click the second Home
Mortgage Payment and Property Taxes node. (If need be, expand the TreeView
by dragging.)
The Data Grid will be positioned to this covering objective, Home Mortgage
Payment and Property Taxes.
Click the Information button to view information entered about this

covering objective.
Increasing Utility curves operate in the same fashion as Decreasing.
From the Home Mortgage Payment and Property Taxes column on the Data
Grid, right-click and select Increasing. Figure 6 shows this portion of the Data
Grid.
Review the information shown on the graph as well as the information entered
for the High and Low X-axis.
If desired, experiment changing the low and high values.
To make this curve non-linear and increasing, drag the line up and to the right.
To make the changes permanent select Apply.
Select Close to return to the Data Grid and compare the changes to the
alternatives' priorities with respect to this covering objective.
132
Figure 6 Major Objective of Home Mortgage Payment and Property Taxes
133
Steps to Create an Increasing Utility Curve

from Scratch
If you were creating an Increasing formula from the very beginning you would:
1. Select a covering objective column
2. Right-click to display the drop down menu
3. Select the formula type, Increasing
4. A dialogue box will appear "This will replace any existing data. Proceed?" Click
OK.
5. Type high and low x-axis values for the curve. (Negative numbers can be used.)
6. If the curve is not Linear, drag the line up and to the right.
7. Select Apply to make the changes permanent.
8. Select Close to return to the Data Grid.
9. Now enter data (numbers) for each alternative with respect to the covering
objective's Increasing Utility curve just defined.
Direct Formula
Use the Direct formula type to directly enter priorities (either specified from some
source or perhaps derived in another model) in the range of 0 to 1.
To show you how this works assume that we have priorities for the Movie Theatres
in the area and that we wish to use this information rather than defining a Ratings
scale.
From the Data Grid, click the Movie theatres' node in the TreeView pane. If it is
not shown, click Leisure Living, then click RMGB to display what is beneath
these nodes.
On the Data Grid, right-click the covering objective Movie theatres and select
Direct.
134
You will be told that the data will be removed (as it no longer applies). Select
OK.
Now enter the respective priorities for each alternative retirement location.
Click the first vacant cell under Movie theatres.
Type .42 (that is a decimal number between 0 and 1)
Move down and type 0
Move down and type 1
Move down and type .5
Click the next vacant cell and drag a bar to the right. Stop at the desired priority.
Formula Type None

Use None to remove a defined formu la type and all data associated with it.
From the Data Grid, right-click the desired covering objective
Select None
When prompted, "This will replace any existing data. Proceed?" Click OK.
Must Constraints
Must Constraints are parameters upper and lower bounds that an alternative must
conform to. You can define a Must Constraint for each covering objective.
From the Data Grid, select View, Formulas/Must and then select Musts to turn
on the display of the formu las types that are shown on the Data Grid.
Move to a desired covering objective and enter either a lower and/or upper
bound.
135
Repeat this process for each covering objective that will be subject to must
constraints.
Note: If an alternative does not satisfy one of the must constraints the alternative
is displayed in red with strikethrough. In addition, the value and/or priority for
the covering objective(s) are displayed in red.
To redisplay the Formulas, select View, Formulas/Must and then select

Formulas to display intensities or parameters, if applicable.
136
Expert Choice Tutorials Lesson 5 Group Capabilities
Lesson 5 - Group Capabilities in Expert

Choice
Overview
Expert Choice can be used by a team to enhance the quality of group decisions by
bringing structure to the decision making process and by synthesizing different
points of view. From earlier lessons we learned how to build a model define the
goal, objectives, sub-objectives and alternatives related to the decision, organized in
a hierarchy. Additionally, we learned about the features of the Data Grid and how to
use and apply different formula types to alternatives' data. In this lesson we will learn
about Expert Choice's group capabilities that help synthesize individual judgments to
arrive at a group conclusion.
First, we will show you how to create a group-enabled model. Second we discuss
how the team-enabled version can be used. Third, we describe how to combine
participants' judgments to reach a group conclusion.
We will be referring to a Facilitator and Participants in the following discussion.

The Facilitator builds the EC model and facilitates the group decision-making
process or session. (Sometimes in practice, one person may act as the Facilitator in
building the model and another person to facilitate the group interaction.) Individual
members of the group or team will be referred to as Participants.
The Facilitator's Role

When Expert Choice is used in a group setting, someone must act as the facilitator. It
may also be useful to have a second person act as a technographer who is responsible
for operating the computer and other technology.
Prior to the group session, the facilitator, and perhaps the meeting owner, might meet
to build the model and enter participants' names, demographics, passwords and other
information as applicable. Alternatively, the facilitator and group might build the
model during a group session.
The facilitator or technographer turns on the receiver and keypads; see Setting Up
the Hardware. This can be done from either the computer or from the facilitator's
137
keypad. If the keypad is used the facilitator has the ability to move around the room
and not be tied down to the computer. Any keypad can be set up as the facilitator
keypad. For convenience we suggest number 1. When this is done this keypad can't
be used to enter judgments. For more information about the facilitator keypad see
page 152.
Note: If the keypad hardware is not available, the facilitator can solicit votes orally
and use the computer keyboard to enter participants votes one at a time.
To keep the group focused on the same issues, the facilitator may automatically turn
the keypads off whenever he/she advances to the next judgment. Prior to voting the
facilitator will turn on the keypads.
Other functions of the facilitator include calculating a set of group judgments;

combining all participants' judgments and data to obtain a group result as well as
displaying a synthesis; and performing sensitivity analyses.
Lastly we discuss Expert Choice's data base format Access and SQL.
Facilitator vs. Participant Functions

What the facilitator can do.
1. Only the facilitator can create/modify a model's structure and enter/edit
Information documents.
2. Structuring, a model building technique, is available only for the facilitator.
3. The Edit, Re vert command, useful to revert to prior model states during a
session, is available only for the facilitator.
4. From the participant table, the facilitator enters participants and demographic
information about each participant as well as optional passwords.
5. The facilitator defines Data Grid formula types.
6. The facilitator can also enter judgments and/or data to be combined with other
participants' judgments and data.
138
What the participant can do.

1. Participants can change their password.
2. Participants can make judgments about the various facets of the decision
problem including:
a. making paired comparisons
b. entering Ratings or data in the Data Grid, and
c. creating notes explaining their judgments as well as other issues or

concerns.
3. Participants can use the Edit, Undo command during the current session to
undo judgments or undo editing where appropriate.
4. Participants can review their own decision model; perform a synthesis; display
sensitivity graphs as well as review the contents of the Data Grid. All
information can be printed.
5. If permitted by the facilitator, participants can view and print one or more of
the combined results.
Creating a Group Model

You can use the group-enabled versions of Expert Choice to create and evaluate a
group model either on a local area network, the World-Wide-Web, or with an audio
response system using radio frequency keypads. This lesson discusses using the
group features using all three modalities. Lesson 6 describes how to install models
on a web-server and how to access models over a local area network or the World-
Wide-Web.
To create a group model, the facilitator (the model builder) first builds a decision
model as described in previous lessons, and then adds participants as described
below.
We will use the Area model found in the Samples directory in illustrating how to
group enable a model.
Start Expert Choice
139
Select File, Open; then navigate to the samples models directory selecting
Samples.
Select the file Area.ahp
Select File, SaveAs and type the filename, GroupArea (There is no need to
append the suffix of .ahp.)
This simple model has only five objectives and no alternatives. There is an
Information document for the Goal showing the different shapes that will be pairwise
compared by individuals in the group. There is also an information document for the
Circle node that contains the answers the actual relative areas of the shapes so the
derived results of the group as well as for each individual participant can be
compared to the actual results for the area validation exercise. (You need to scroll or
page down to see the results because we don't want the participants to inadvertently
see the results before the exercise is complete.)
Adding Participants to and Group Enabling

your Group Model
Note: You must be logged on as the Facilitator to add participants to
and group enable your model.
To group enable your model, select Go, then Participants Table or click the
Participants button on the toolbar in the ModelView.
Select Edit, then select Group enable. This will add a COMBINED instance
into the table.
To add participants to the model, select Edit; then select Add N Participants.
Enter the number of participants to add, e.g. 4.
140
To enter or change participants names

Click on a participant such as P2 and type a name. The facilitator can also enter his
or her name by typing over the word facilitator.
Entering Demographic Information

Demographic information such as age, sex, password and keypad number can be
entered as desired. Note: If a password is entered, it is encrypted. Passwords can
also be entered or changed using the ModelView's Edit command.
Now you have a group-enabled model with participants.
Select File, Close to close the Participants Table and return to the ModelView.
Copying Judgments and/or Data when

Adding New Participants
If desired, judgments and or data can be copied from the facilitator or another
participant's instance of the model when adding new participants.
This is done from the Participants Table by using the Edit, Copy to N New
Participants command. Participants will be added just as with the Add N
Participants command, but in addition, the judgments and/or data from the
participant row currently selected will be copied to the new participants.
Deleting Participants
Deleting a participant deletes everything about that participant from the model; this
includes all judgments, associated priorities and data related to that participant.
Before doing this you might want to make a copy of your model for backup and
historical purposes. (You may also want to do a File Save before deleting
participants in case you might want to do an Edit Revert during the session).
From the Participants Table, select the person to be deleted.
Select Edit, then select Delete Participant and follow the prompts.
141
Things to Do Before Opening a Model for

Group Use
Most, if not all, of these items will be performed by the facilitator.
1. The model should be built prior to placing the model on the server. Lesson 6
tells you how to put the model on a web server.
Built means that objectives, sub-objectives and alternatives have been added to
the model. If the Data Grid will be used then the covering objectives formulas
must be defined and assessed.
To save time, it is desirable to do the same for the keypad version (audio
response system) but it is not necessary.
2. The model must be group-enabled with participants before using a model on

the server. It is highly desirable to do the same for the keypad version but it is
not necessary.
3. The facilitator must determine if new participants (guests) on the network or

web will be able to access an instance of the model.
4. If the model will be used on a network, the facilitator must provide instructions
to the group for model evaluation via email. In addition, the facilitator must
email participants their user names and passwords if applicable. If non-
predefined participants are allowed, the guest will be provided instructions for
using the model as well as a registering password.
If the model will be used with keypads see the Hardware Appendix for
installation and activating instructions. This section also has information about
setting up the facilitators keypad and menus as well as setting other options
such as the number of participants, waves, etc. It is recommended that you do
this prior to conducting a group session.
Allowing New Participants to Open a Model

without Having Been Pre-defined
The facilitator can set an option to allow new participants to open a model without
having been pre-defined.
142
From the ModelView or Data Grid, select Tools; then select Options.
Then select the Open tab.
Select the desired option from the File Open Options box.
o One option requires no password from the user (new participant, no

password required).
o Another option requires that a new user supply a password in order to

REGISTER. (The participant can then enter his/her own password for
subsequent access to the model.)
The facilitator can also specify options for copying judgments and/or data to the
new participants. (See the Copy to Participant Options box.) These options
include: None (dont copy any judgments or data); Hierarchy judgments and
priorities only copied; Data grid entries only copied; or prompt the user.
Now we will describe how to open and use a group model first using keypads and
second using a network. Then we describe how to combine a group's results.
143
Opening a Group Model Using Keypads

1. The facilitator will start Expert Choice and select the model name.
2. If you did not change the default facilitator name discussed earlier in this
chapter, select OK. Otherwise, type your facilitator name; then select OK.
3. If the model is password protected you will be required to enter the password.
Tip: When you open a model as the facilitator, you can switch to view a
participants information in any of the following ways: (1) double click
on a person in the Participants window; (2) use the Participant drop-
down list from the ModelView window; or (3) se lect Go, Person and
enter a number or double click a row.
What the Facilitator Does after a Model is

Opened
1. If the combined instance is not displayed in the ModelView (look at the title
bar) the facilitator must open the combined instance of the model by clicking
on the Combined instance from the drop-down list. You will know when the
combined instance is displayed as a magenta border will be displayed in the
TreeView.
2. Then if paired comparisons are to be made the facilitator will:
a. Select Assessment; then select Pairwise Indi vidual. If group

comparisons have been previously made you will be returned to the
original comparison mode. Or you can click any comparison tab from
the ModelView's TreeView pane and then select Assessment, Pairwise
Individual.
b. From the Pairwise Individual Window select and display one of the
individual comparison windows by selecting the Verbal, Numerical, or
Graphical tab.
If the group will be entering Ratings or data values, the facilitator will:
c. Go to the Data Grid. Select Assessment and then select Indi vidual.
144
3. Now is the time to provide instruction and information about the first paired
comparison to be made.
4. Display either the Poll or Vote window.
The Poll window shows which keypads voted. This visual will let the
facilitator know who voted without any participants looking at the
judgments/votes entered.
The Vote window shows the actual judgment entered by each individual.
5. Turn on the keypads so each participant can enter a judgment (vote).
Note: The facilitator may or may not turn the keypads off in between
judgments.
Tip: Turning the keypads off until voting helps the participants focus on
discussion by not voting before discussion has ended.
6. After the participants vote the facilitator moves to the next paired comparison
to be made and repeats steps 4 6.
7. When the comparison process is completed, the facilitator calculates each

individuals priority as well as the group priority.
The facilitator might also choose to configure one keypad as the Facilitator
keypad; see page 152. This feature provides the facilitator with the mobility to
walk around the room rather than being tied to the computer's keyboard. Also
see Setting Up the Hardware appendix.
Group Decision Making Using the Keypads

After the facilitator has provided instruction to the participants and displayed the
appropriate Expert Choice windows, participants can do the following:
1. From the ModelView individual participants can make judgments with respect
to the objectives and alternatives using Verbal, Numerical or Graphical
judgments.
2. From the Data Grid individual participants can enter data for Step Functions or
Utility curves, or make judgments about Ratings intensities with respect to the
covering objectives.
145
Each option is now explained.
Using Keypads to Make Verbal Judgments

Press a number from 1-9 on the keypad to represent a numerical representation
of a verbal Judgment. If the facilitator projects the verbal comparison window,
use the scale at the bottom of the window to determine what button to press
when making a judgment. Also refer to Table 2 Numerical Representation of
Verbal Judgments on page 163 of this manual.
For example, if you feel the paired comparison is equal you would press one.
If you felt the comparison was a magnitude of order in difference you would
press 9.
Remember you are not entering a number but are entering short-hand notation
for the verbal comparison being made.
Note: All keypad entries default to the first (left) factor being compared.
146
To the judgment, switching the dominance of the judgment, from the left to
*
the right or vice versa, you must press (the asterisk key) either before or
after pressing the number that represents the verbal comparison. (To determine
if you want to do this, look at the projected verbal comparison window.)
To clear a judgment, press the 10 key.
Using Keypads to Make Graphical

Judgments
When making graphical judgments, you use the keypads to move the bars either to
the left or right. Use the following figure to help do this.
For example, when making graphical judgments pressing the numbers 1 or 2 on the
keypad causes the bar to move left or right in a large increment. Using 4 or 5 moves
147
the bars moderately, while 7 and 8 move the bars slightly. Lastly the * (the
asterisk key) and the 10 key move the bars a tad.
Each key press causes the bars to move either to the left or right. Data
registration takes a second, so wait until you see the bars move before
depressing another key.
To invert the judgment, switching the dominance of the judgment, from either
the left to the right or vice versa, you must press 6. (To determine if you want
to do this, look at the projected comparison window.)
To clear a judgment, press 9.
To record the graphical judgment, you must press 3.
Using Keypads to Make Numerical

Judgments
Press a number from 1-9 on the keypad to represent how many times one factor
is more important or preferable than another.
148
For example, if you feel the paired comparison is equal you would press one. If you
felt the comparison was 5 times more important you would press 5.
Note: All keypad entries default to the first (left) factor being compared.
To invert the judgment, switching the dominance of the judgment, from the left
*
to the right or vice versa, you must press (the asterisk key) either before or
after pressing the number that represents the verbal comparison. (To determine
if you want to do this, look at the projected comparison window.)
To clear a judgment, press the 10 key.
Making Individual Judgments Using the Data

Grid with Keypads
From the Data Grid window, the facilitator will project the display of each covering
objective. The facilitator might also display information about the covering objective
149
or lead the group in a discussion about the covering objective. Then the group will
enter a ratings or value. When all entries have been made, the facilitator will move to
the next covering objective to be assessed. This process will be repeated for all
covering objectives.
Tip: The Autoadvance to the next cell option is on by default. If you

do not see an arrow pointing down or right on the button bar of the
Data Grid, turn on Autoadvance, by selecting Tools, Options,
Calculation; then select Autoadvance.
Tip: While at the Data Grid, clicking an objective in the TreeView selects
that column in the Data Grid. If there are sub-objectives associated with
that objective you are be positioned to the first sub-objective. To go to
a sub-objective you must click on it in the hierarchy.
Entering a Ratings with a Keypad

Once the facilitator has selected the covering objective to be rated from the Data
Grid and has turned on the keypads, use your individual keypad to either:
Press the number that represents the ratings intensity such as 2 for Very
Good.
Press * (asterisk) to enter a decimal point and then enter a number such as
345. To record the decimal value of .345, you must press * . The sequence of
key presses is our example is *345*.
To enter zero, press the ten (10) key.
To clear a rating, press the asterisk key twice (**).
Entering a Step Value with a Keypad

Once the facilitator has selected the covering objective to be evaluated and turned on
the keypads, use your individual keypad to either:
150
Press the number that represents the data value such as 25; then press *. The
sequence of key presses is 25*.
*
First press 0 (the 10 key); then press and then press a number such as 672.
To record the decimal value of .672, you must press *. The sequence of key
presses in our example is 0*672*.
To enter a decimal in the middle of the sequence, enter for example 56*3* will
cause 56.3 to be entered.
Note the facilitator must turn the decimal option on using the Options
Keypad Decimals command on the Individual Input Window.
* for the minus sign, then either enter the

To enter a negative value, press
number such as 675 and press * to record. The sequence of presses will be
*675*.
To enter 0.675.9 press 0 (the ten key) then * followed by 675* (for the decimal
point) and lastly press 9
To enter zero, press the ten (10) key.
To clear a value, press the asterisk key twice (**).
Each cell, if valued, is assigned a priority based on the bounds of the steps in the
function.
Entering Data for a Utility Curve with the Keypad

Once the facilitator has selected the item to be valued using the Increasing Utility
function and turned on the keypads, use your individual keypad to either:
To enter an integer press the series of numbers, such as 765, then press *
(asterisk)
151
*
Press 0 (the 10 key) to enter zero and press to enter a decimal point. Then
enter a number such as 345. To record the decimal value of .345, you must
press *. The sequence of key presses in our example is 0*345*.
To enter a decimal in the middle of the sequence, enter for example 56*3* will
cause 56.3 to be entered.
Note the facilitator must turn the decimal option on using the Options
Keypad Decimals command on the Individual Input Window.
*
To enter a negative value, press then the number followed by *. If you want
to enter 0.5 you would press *0 (the ten key) *5.
To enter a zero, press the ten (10) key.
To clear a utility curve value, press the asterisk key twice (**).
Entering a Priority with Keypads

Once the facilitator has selected the item that will be assigned a priority from 0. 1.0
and turned on the keypads, use your individual keypad to:
Enter a priority in the range of zero to one (0 1.0).

For example, press * (asterisk) to enter a decimal point and then enter a
number such as 672. To record the decimal value of .672, you must press *.
The sequence of key presses is our example is *672*.
Turn to page 163 to learn when and how to combine a group's results. Then return
here to learn about the facilitator keypad functions.
Facilitator Keypad
The facilitator keypad is used in conjunction with the facilitator menus to help the
facilitator customize Expert Choice for his or her specific use by logically and
conveniently grouping commands that will be frequently used with the facilitator
keypad. Ten different menus can be defined. We have provided several default
152
menus that can be altered and restored at any time. The facilitator can modify our
menus and create others.
From any Individual Pairwise or Individual Data Grid window, select File,
Facilitator Menus. The default menu will appear; you can select another by
changing the Menu number in the window.
Review the default menus and if desired print them. After becoming proficient
with Expert Choice you may desire to change or create new menus.
To name a menu or change the name of a menu, click the gray bar in the top
center of the window and type the desired name.
To add a command to a facilitator menu, move the cursor to the row where you
want to add the command. Then select the desired command from the
command box. Repeat this process as necessary.
When finished editing the facilitator menus click OK to make the changes
permanent.
To select a facilitator menu to be used select File, Facilitator Menus, then

select a number from the drop-down list. Click OK.
Tip: Use the File Keypad Options command from any individual window
to set the facilitator keypad before or after selecting a facilitator menu.
Tip: Pressing 10 on the facilitator keypad will hide or show the current
facilitator menu on the computer and/or projection device. Note: The
menu does not have to be visible to perform an action.
153
Opening a Group Model on a Network or

Internet
1. Prior to the session the model will have to be placed on your network or
Internet.
a. To install on your network contact your network administrator.
b. For Internet model installation see Lesson 6 and contact your network
administrator.
2. Make sure you are connected to your network.
3. Start Expert Choice by selecting the windows Start key; then select Programs
and select Expert Choice. Alternatively you can click the Expert Choice
desktop icon.
4. a. If you are using a network, select File, Open.
b. If you are using the Web (Internet) select File Open Web Model or File,
Open SQL Web Model. (This is dependent on your version only one
option will be displayed.)
5. Then type the path and name of the model (provided to you by the facilitator).
6. When a dialogue box appears asking whether you are opening the model as a
facilitator or participant follow the appropriate directions listed below.
a. As the Facilitator :
If you did not change the default facilitator name, select OK.
Otherwise, type your facilitator name; then select OK.
If a password has been previously assigned, you will be prompted to

enter it.
b. To open the model as a predefined participant:

Type either the participant name or participant id number, referred
to as PID. You will be prompted for a password if one has been
previously assigned. (The facilitator will provide these items.)
154
c. To open the model as a guest (non-predefined participant):

Type your name, and then type the registering password. (The
facilitator will provide you with the registering password.)
Then if desired, use the ModelView's Edit Password command to

change your password to protect your instance of the model.
Once a participant opens a group model, he/she is ready to make judgments for
elements in the decision hierarchy or enter ratings or data for alternatives in the Data
Grid. The facilitator will provide you with instruction on what portions of the model
to evaluate.
Group Decision Making Using a Network or

the Web
Whether the facilitator and participants are in the same or different buildings or using
the World Wide Web they can participate in group decision making by working
together at the same or even different times.
Participants can make judgments about the objectives, sub-objectives and

alternatives as described in Lesson 1; or they can, using the Data Grid, enter ratings
or data for the alternatives as described in Lessons 3 and 4. The facilitator can
request that participants do either or both of these for one part of the model at a time,
or for the entire model.
Once judgments and/or data have been entered, the facilitator can combine some or
all of the judgments and/or data, for some or all of the participants. (Each participant
will notify the facilitator with email, Net-Meeting or America On-lines Instant
Messaging or a similar product that they have finished their assigned tasks.)
Participants' Paired Comparison Judgments

Participants can make paired comparisons in one of three ways (Verbal, Numerical,
and Graphical) from the ModelView. You will be at the ModelView once you open
your instance of the model. Each is now explained.
155
Making Verbal Judgments

The Verbal Comparisons window is divided into two sections. Verbal judgments are
made in the top pane. Two elements are compared with respect to their parent. What
makes Verbal comparison unique is that words are used to represent the magnitude
of the scale. The slider bar on the right side of the pane is used to indicate which
element is preferred and the strength of that preference is represented by a
corresponding word. The two opposing sides of the scale represent each element
being compared.
The comparison matrix is displayed in the lower pane. The numerical representations
of the verbal judgments are displayed here as numbers from 1 to 9. If the row
element (on the left) is preferred, then the judgment is displayed in black. If the
column element is preferred, then the judgment is "inverted" and displayed in red.
When enough judgments have been made to calculate priorities, they will also be
displayed as bar graphs that overlay the row elements.
Judgments can be made any of the following ways:

Drag the slider bar with the mouse.
Click on a statement (i.e. Moderate, Strong) next to the bar (or between two
statements).
156
Right-click on a statement to automatically advance to the next comparison.

This is the fastest way!
To advance to the next judgment

Expert Choice is automatically configured to advance to the next judgment. You can
change this with the Tools, Options, Calculation command.
If Autoadvance is off, click on the cell in the comparison matrix that represents
the next comparison to be made.
To invert a judgment (to select the other element in the

comparison)
Click the Invert icon to select the other side of the comparison scale.
To enter judgments directly in the comparison matrix

Click on the cell representing the comparison you want to judge and type a
number from 1 to 9; see the numerical representations of verbal judgments.
When using this option it is highly recommended that you use the Numerical
mode since you are probably saying that you prefer, for example, Apples to
Oranges 3 times more with respect to Craving.
Making Numerical Judgments

The Numerical Comparison Window is divided into two sections. Numerical
judgments are made in the top pane. Two elements are compared with respect to
their parent using a numerical scale. The slider bar is used to indicate which
judgment is preferred and the strength of that preference. The two opposing sides of
the scale represent each element being compared.
157
The comparison matrix is displayed in the lower pane. The numerical equivalents of
the judgments are displayed here as numbers from 1 to 9. If the row element (on the
left) is preferred, then the judgment is displayed in black. If the column element is
preferred, then the judgment is "inverted" and displayed in red. When enough
judgments have been made to calculate priorities, they will also be displayed as bar
graphs that overlay the row elements.

Drag the slider bar with the mouse.
Click on a number above the bar.
Right-click on a number to automatically advance to the next comparison. This

is the fastest way!

Expert Choice is automatically configured to advance to the next judgment. You can
change this with the Tools, Options, Calculation command.
If Autoadvance is off, click on the cell in the comparison matrix (in the lower
pane of the window) that represents the next comparison to be made.
Notice that as you make judgments, the numerical equivalents of your

judgments will appear in the comparison matrix.
158
To invert a judgment (to select the other element in the

comparison)
Click the Invert icon to select the other side of the comparison scale.

Click on the cell representing the comparison you want to judge and type a
number from 1 to 9.
Making Graphical Judgments

The Graphical Comparison view is divided into two sections:
Graphical judgments are made in the top pane. Two elements are compared with
respect to their parent with bar graphs. The lengths of the bars indicate the relative
dominance of the elements. If they are equal length, then the elements are equally
important. If one bar is twice as long as the other, then it is twice as important.
Relative dominance is also represented with a pie chart on the right side of the pane.
The comparison matrix is displayed in the lower pane. The numerical representations
of the graphical judgments are displayed here as numbers. If the row element (on the
left) is preferred, then the judgment is displayed in black. If the column element is
preferred, then the judgment is "inverted" and displayed in red. When enough
judgments have been made to calculate priorities, they will also be displayed as bar
graphs that overlay the row elements.
159

Drag either the blue or red bar with the mouse.
Right-drag one of the bars to automatically advance to the next comparison.

Expert Choice is automatically configured to advance to the next judgment. You
can change this with the Tools, Options, Calculation command.
If Autoadvance is off, click on the cell in the comparison matrix (in the lower
pane of the window) that represents the next comparison to be made.
Notice that as you make judgments, the numerical representation of your

graphical judgments will appear in the comparison matrix.
To invert a judgment (to select the other bar in the

comparison)
Click the Invert icon to select the other bar.

Click on the cell in the matrix representing the comparison you want to judge
and type a number from 1 to 99.
When using this option it is highly recommended that you use the Numerical
mode since you are probably saying that you prefer, for example, Apples to
Oranges 3 times more with respect to Craving.
160
Entering a Ratings in the Data Grid

There are several ways that Ratings can be entered either using the mouse or
keyboard.
With the mouse :

Move to Data Grid cell.
Notice how a scale appeared above the alternatives on the Data Grid when you
moved to a cell.
Click on a ratings intensity. (You can either click a ratings word located in the
first row in the scale or a ratings number located in the second row in the scale).
Then the ratings intensity will appear in the cell and you should move to the
next cell to be evaluated.
With the left mouse button depressed in the cell, drag the bar in the cell to a
desired value. This function lets you create a ratings value between two discrete
values.
With the keyboard:

Type a number representing a rating number.
Type a ratings word.
Type a value between zero and one.
Type 0 for a zero rating, or
Clear a rating, by pressing Delete.
Entering a Value in the Data Grid

With the mouse :
161
Notice how a scale appeared above the alternatives on the Data Grid when
you moved to a cell.
Drag a bar in the cell.
With the keyboard:

Type a value; you can refer to the scale that appears at the top of the Data Grid.
Note: To type a value between zero and one dont forget to enter a decimal
point.
To enter zero type 0.
Clear a value, by pressing Delete.
162
When to Combine Judgments and Data

Participants' judgments and data will automatically be updated to the combined
instance when keypads are used. Participants can post their judgments and data
during a group session on a network by using the Assessment, Post command
available from either the ModelView or Data Grid..
The facilitator should combine participants' judgments and data after a group
session using a network especially if participants are working in a different place
and time environment. Do this by using the File Refresh command that is
available from either the ModelView's Pairwise Individuals window or the Data
Grid's Individual window.
Additionally the facilitator might want to combine certain participants to see a

subset of the decision. The next section describes this.
Combining Judgments and/or Data

Judgments and/or data can be combined for a set of active participants. The
facilitator determines who the active participants are by entering and applying an
SQL (Structured Query Language) query from the Participants table. See the figure
below.
One way to do the above is to select and deselect who will participate by checking
and un-checking the participating column in the table and then selecting the
Combined button to invoke the selection.
A second way is to select a predefined SQL query from the drop-list at the bottom of
the Participants table and then select the Combined button.
A third way is to define a new SQL query and then select the combined button.
Additionally the facilitator can combine judgments and data from the ModelView or
data from the Data Grid. When one of these choices is selected the active query
defined in the Participants table will be used to combine the information, replacing
what was previously in the combined instance.
Now we explain some queries followed by the instructions to combine from the
Participants table, ModelView or Data Grid.
163
Figure 1 Partial Sample Participants Window
Two pre-defined SQL queries, "all", and "participating" can be applied by clicking
on command buttons located at the bottom of the Participants window. The SQL
queries generated and applied for these command buttons are:
All: Select * from People order by PID (Person ID) This query selects all
participants in the model and includes those with and without checkmarks in the
Participating column of the Participants table.
Participating: Select * from People where Participating order by PID This query
only selects those instances (participants) whose participating column is checked.
Note: The Combined instance contains the results of the model and has both the
combined and participating columns checked.
The facilitator can enter and apply any valid SQL query to select the set of active
participants. Some examples are as follows:
Select * from People where age >=15 order by PID
164
Select * from People where age >=15 and Participating order by PID
Select * from People where NOT Participating order by PID
SQL queries can also be saved in the model (for later retrieval) using the Save button
and then specifying a name to identify the query.
Tip: It is suggested that the Facilitator examine the Participants table to

ascertain that the desired set of participants are active before
combining judgments and/or data.
All participants entered into the database are initially considered active. If you want
to make someone inactive, you must click on his or her participating column to
remove the checkmark.
Combining judgments and/or data can only be done by the facilitator. This process
can be accomplished from one of three menus and windows.
From the ModelView
From the Data Grid
From the Participants Window
Judgments, as well as data, for the active participants will be combined and placed in
the "combined" section of the database with a PID of 1, and loaded into memory
where the results can be reviewed, synthesized, or analyzed in sensitivity graphs. In
addition, the Combined Column of the Participant table will be checked to prevent
this combined model from being included in future combinations.
In the newly copied combined model's Goal node, you may wish to include
information about the active participants used to form the combination.
Tip: The facilitator can combine different parts of a model with different
sets of active participants. For example, people with expertise in one
area might make judgments about one cluster of objectives in the
decision hierarchy, while another group of people expert in another
area might do the same for a different cluster. Still another group might
enter ratings or data in all or parts of the data grid.
165
Tip: In the participant table, the judgments/values in "Combined" (PID

= 1) can be copied and saved under a new participant name by using
the Edit Copy to N Participants command, by specifying 1 for N and
copying from the PID=1 row. After doing this, the new person name
should be changed to something meaningful like "Combined for
Civilians"; then checkmark the Combined column to designate that this
is a combination.
Combining from the Participants Window

Click the Combine Individuals button at the bottom of the Participants
window.
A dialogue box will appear asking you to select one of the following:
o Judgments (in hierarchy only),
o Data only, or
o Both.
Combining From the ModelView

Select Combined from the Participant drop-down list in the ModelView.
Select Assessment; then select Combine Participants Judgments/Data.
Select Entire Hierarchy (other options include Plex or Current node). A

dialogue box will appear asking you to select:
Judgments (in hierarchy) only,
Data only, or
Both
Then select your option. Selecting judgments will combine active participants
judgments in the combined instance of the model. You can determine active
166
judgments from the Participants table. Selecting Data will only combine the data
for the active participants.
Combining from the Data Grid

When combining from the Data Grid, only data (not judgments) for the decision
hierarchy are combined.
Select Assessment, then Combine Data for and then select one of three
options:
o All Covering Objectives
o Covering Objectives of Selected Parent Plex (Click the parent node in the
TreeView before choosing this option)
o Only Selected Covering Objective
167
Access and SQL Database

Our individual user and group versions of Expert Choice store your models in an
Access database; these files have the extension of .AHP.
Expert Choice Group can use either an Access or SQL database. To use SQL you
must have both Microsoft Access 2000 and Microsoft SQL 7.0 installed on your
computer. The SQL server will determine the file's naming convention and
extension.
Upsizing from a Microsoft Access Database to a

Microsoft SQL Database
1. Start MS Access.
a. Select File | Open and navigate to the drive\ path that contains the
Expert Choice.ahp model.
b. In the File name dialogue box, type *.AHP to view all AHP models.
c. Select an Expert Choice model. This file is in Access 2000 format.
d. Select Tools | Database Utilities | Upsizing Wizard.
e. Select Create new database and click Next.
f. In the next window:
1) Type the database server name or select one from the

dropdown list;
2) Type the Login ID; (usually Administrator)
3) Type the Password; (usually sa)
4) Type the name of your new SQL Server database; and
5) Then click Next. Note: If an error message appears, click

OK. When the dialogue box appears click Trusted
Connection and continue.
168
g. In the next window, when asked "Which tables do you want to export
to SQL Server?" Click the >> button (the double angle brackets) to
include all the database tables and click Next.
h. Since nothing needs to be changed on the next two windows, click

Next twice.
i. Now click Finish to complete upsizing to SQL 7.0. When this process
is complete, the Upsizing Wizard Report will be displayed.
j. Close Access.
2. Start Expert Choice.
a. Select File | Open SQL Model.
b. Type the database server name, username (Login Id) and Password
given in step 1f.
Accessing and Viewing an Expert Choice SQL 7

Database
1. First create an access Expert Choice model (.AHP) using the File, Open
command.
2. Upsize the database from Access to SQL 7 using Access' Upsizing Wizard as
described above.
3. Start SQL by Selecting Microsoft's Service Manager from the Window

Programs group.
4. Select Microsoft's SQL Enterprise Manager to view your database.
a. Move down the tree structure from SQL Server Group to the name of the
SQL file created.
b. To edit this file use SQL commands. It is strongly recommended that you
use Expert Choice to add or delete information from the database.
169
Executing an Expert Choice SQL Database

2. Select File, Open a SQL Model; or to open a SQL web model, select File,
Open a SQL Web Model.
3. Then enter the following:
a. The models location such as http://models.expertchoice.com
b. SQL Server Name
c. Model Name
d. User ID and
e. Password
The facilitator and participants should login with their assigned username and
password, if applicable; then proceed as directed by the facilitator.
For more information about SQL databases see Access and SQL Database
170
Expert Choice Tutorials Lesson 6 Web Models and the Internet
Lesson 6 - Web Models Accessible

Over the Internet
Overview
Expert Choice's Web mo dels allow members of a group in various parts of the world
to simultaneously open the same model over the Internet to solve problems and make
decisions. In this lesson you will learn how to enable Expert Choice's Web Model
feature. We also describe how to put a model on an Intranet.
System Requirements
Internet Information Server (IIS) running on Windows NT 4.0, 2000 or XP, or
Personal Web Server (PWS) running on Windows 95, 98 or Me (or equivalents).
Expert Choice software should be installed on client computers for both the
facilitator and participants. Note: Expert Choice software does not need to be
installed on the web server.
Overview How Web Models and Expert Choice Work

1. The facilitator creates a group model as described in Lesson 5, either on the
destination Web server, or on a client computer. If the model is created on the
client's computer it must be uploaded to the server.
2. To access the model on the server, the facilitator must create an ODBC
System Data Source Name (DSN).
3. The facilitator, using Expert Choice on his/her machine, can access the model
via a Web connection and can add to or revise the model structure. This must
be done when no participants are accessing the model. Note: The Edit Revert
commands are, however, NOT available when editing a model via a Web
connection.
4. Participants using Expert Choice on their computers can access the model via
a Web connection to make judgments, enter data, and examine their results.
171
5. The Facilitator combines the groups' judgments and or data (for all or active
participants) and makes the results available to the participants.
Setting Up a Model for Access Over the

Internet or an Intranet
If you have questions about the following contact your system administrator.
Put an Expert Choice Model on the Server

You must create a new Expert Choice model or transfer a previously created model
to the server computer.
If you use an Intranet put the model on the shared drive and tell the participants the
location of the model; for example, F:\ models \ExpertChoiceModelName. Then
continue reading on page 182.
If you use the Internet follow the directions below to put a model on the Internet
server.
Creating a new System Data Source Name

Contact your system administrator for the URL or IP address of the web server on
which the model will be stored. We will use http://models.expertchoice.com as an
example in the following discussion. When you have completed the steps outlined
below, the Expert Choice model that is on the web server can be accessed if you
have a copy of Expert Choice on your computer.
From the web server computer, create a new System Data Source Name (DSN) that
points to an Expert Choice model as follows :
1. From the Start menu select Settings, then Control Panel.
2. For Windows 95/98 or Me systems:

Note: For Me systems you must first select the "View All Control Panel
Options" and then follow the directions on the next line.
Select ODBC32 Data Sources or ODBC DATA SOURCES (32bit).
172
For NT systems: Select ODBC Data Source.
For Windows 2000: Select Administrative Tools, Data Sources (ODBC).
3. Select the System DSN tab, see Figure 1.
Figure 1 ODBC Data Source Administrator
4. Click the Add button
From the list, select Microsoft Access Driver; see Figure 2.
Click Finish
173
Figure 2 Create New Data Source
5. Type a Data Source Name to designate the model to be accessed with Expert
Choice; see Figure 3. Then click Select.
174
Figure 3 Entering a Data Source Name
6. Now either navigate to or type the path/file name that contains the desired
Expert Choice model. The path/file name must include the extension .ahp.
Note: If desired you can type *.ahp to see a list of all Expert Choice models.
7. Click OK to go back to ODBC Data Source Administrator; see Figure 4. Then

click OK to complete the System DSN process.
175
Figure 4 ODBC Data Source Administrator
An additional step may be needed to host web models using a Windows XP

machine. It is described below.
1. From the Windows Start menu select Run and type INETMGR.
2. Double click on your local computer; double click on Web Sites.
3. Right click on the Default Web Site and create a new virtual directory by
selecting New | Virtual Directory. See the figure below.
176
Figure 5 Window showing Local Computer Name and Default Web Site
4. When the Virtual Directory Creation Wizard appears click Next.
177
Figure 6 Virtual Directory Creation Wizard
5. When prompted for the Virtual Directory Alias type MSADC and click Next.
178
Figure 7 Virtual Directory Alias
6. When prompted for the Web Site Content Directory, type the MSADC system
directory (usually at the location shown below) and click Next.
179
Figure 8 Web Site Content Directory
7. Make sure the following access permissions are set as shown below and click
Next.
180
Figure 9 Access Permissions
8. Then click Finish to complete the web hosting process.
Now that the model is on the server, it can be accessed from Expert Choice that is
installed on either the facilitator or participants' computers.
Accessing a Model over the Internet

Before following the directions below, the following information should be
obtained from the system administrator by the facilitator and provided to each
participant.
The URL Location - IP address or hostname of the web server
Group users will need the SQL server name if their database is SQL
The DSN name
181
The User name assigned to the database
The Password for the user name.
In addition, the facilitator should provide each Participant with his/her Expert
Choice model participant user name and password, if applicable. The facilitator
should also provide directions to the participants for making judgments.
1. Connect to the Internet and verify network connection to the web server.
3. Select File, then Open Web Model. Then a dialogue box will appear.
4. In the Location (URL) field, type the IP address or hostname of the web
server.
For example: http://models.yourhostname.com or http://12.34.56.78
Step 5 is only for Group users using SQL. Others go to Step 6.

5. In the SQL Server field, type the SQL server name .
6. In the Model Name field, type the database name selected when upsizing.
For example: samplemodel
7. In the UserID field, type the user name assigned to ODBC connection or SQL
database.
8. In the Password Field, type password for username (in step 7 above).
Now that you have permission to access the model over the Internet or Intranet
you are ready to open your model.
Logging into a model on the Server

The facilitator should login by clicking OK if the facilitator name has not been
changed. If it has, then the facilitator must enter his or her assigned name. Then
if applicable, the facilitator will be prompted to enter a password.
182
Participants should login with their assigned participant username and password,
if applicable. These are probably different names than those used in steps 7 and
8 above. Then proceed as directed by the facilitator.
183
Expert Choice Tutorials Lesson 7 Resource Allocation
Lesson 7 - Resource Allocation

Overview
Expert Choice's Resource Allocation module provides a way to select a combination
of alternatives that maximize (optimize) the attainment of objectives while not
exceeding a specified budget. Other constraints can also be specified. There are two
forms of Resource Allocation Discrete and Activity Level.
This feature is only available in the group versions of Expert Choice.
How to Perform Resource Allocation

The Resource Allocation module (RA) is accessed from the Tools | Resource
Allocation menu in the Data Grid. This module extracts information from Expert
Choice with group capability and formulates the spreadsheet that will be used to
perform the resource allocation calculations.
Before using RA you should have completed your model's evaluation. In addition
you must enter costs for each alternative in the Data Grid's Cost column. (If the Cost
column is not visible, use the View | Cost command). Once this is done you can
invoke the optimization. Since decision-making is not static remember you and your
group can always adjust your Expert Choice model and re -execute the optimization.
You can also download the Resource Allocation chapter of Decision by Objectives
published by World Scientific Publishing or available at
http://www.expertchoice.com/dbo/ or http://mdm.gwu.edu/forman (where a TV
Video and screen capture videos are also available).
The following sections describe how to perform resource allocation.
184
Discrete Alternative Resource Allocation

Optimization
We explain, by example, how to use and perform a discrete alternative resource
allocation optimization. This module uses algorithms that are exceptionally fast,
even for very large problems.
1. Load the Hawthorne Valley model from the Expert Choice Samples directory.
When prompted for a participant name, just press Enter.
2. From Expert Choice the menu select Go | Data Grid.
3. Select Tools | Resource Allocation.
4. Expert Choice will formulate the optimization problem and you will be taken to
an Excel spreadsheet where you can control the optimization process. Expert
Choice determines if a discrete or activity level optimization should be
performed based on the absence or presence of activity levels (To designate
activity levels you must conform to the following convention; alternative
name:activity level where a COLON is used to separate the alternative name
from the activity level.) Since there are no activity levels in this model, a
Discrete Alternative Resource Allocation will be performed and you will be
taken to an Excel spreadsheet showing information that has been extracted from
Expert Choice.
5. The initial display shows a solution where every alternative is funded at the
highest benefit level (note the 1s in Column B), and the total amount required
to fund these levels below the red/orange cell in the Costs column.
185
Figure 1 Initial Discrete Alternative Solution
6. Type a budget limit in the red/orange cell and press Enter. In this example
enter 50000.
7. From the Excel Menu select Resource Allocation. From the drop down menu
select Solve.
8. The results show that all but two alternatives are funded and how much funding
is expended.
186
Figure 2 Discrete Alternative Results
Min and Max Columns in the worksheet; see the figure above. If you enter 1 in the
Min Column for an alternative it must be funded. If you enter 0 in the Max Column,
for an alternative it will not be funded.
After viewing your result you can choose to Optimize for Increasing Budgets by
selecting this option from the Resource Allocation menu. Next either accept the
defaults for this function or change them; use the figure below as a reference.
187
Figure 3 Optimizing Options Window
Minimum Cost uses the lowest cost activity/alternative for example, 1,000.
The Minimum Difference of Costs uses the delta (change) to increase the
budgeted amount, for example 100.
User Specified The default value is 1000. You can enter any value in this cell.
Approximate Number of Increments; the default is 25 and the
recommended option. However you can enter a number from 0 to the
maximum budget. The number of increments is used to display the next point on
the Pareto curve.
You can Add Constraint Columns to the spreadsheet by selecting this option from
the Resource Allocation menu. The figure below shows a new column that has been
added and it also shows the results of the optimization based on the new entry.
188
Figure 4 Constraint Column Added Showing New Results
To add a constraint use the figure above and follow the instructions below:
From the Resource Allocation menu select Add Constraint Column.
1. When prompted, type the number of columns to add, in this example type 1.
New constraint columns will be added to the right of the Obj (objectives) and
Costs columns.
2. Ne xt you must enter a title (text heading) for each new column. In this example
add the heading labeled Sup Wks for supervisor weeks.
3. Now enter the number of weeks required of the Supervisor for each alternative
as shown in the figure below. Notice some rows do not require supervisor time.
4. Suppose that only 4 weeks of supervisor time are available. Enter a 4 in the
max row of this new constraint column; see the figure above.
5. From the Resource Allocation menu, select Solve . If desired you can select
increasing budgets. The results with the new constraint will be displayed. Note:
189
The salmon color cells show the total cost and in this example the total
supervisor hours.
6. When you are finished using the optimizer you will be given the opportunity to
save your spreadsheet and then you will be returned to Expert Choice. We
highly recommend that you rename this spreadsheet to preserve the audit trail.
Dependency Feature
Use the Dependency tab to specify dependencies between alternatives in the resource
allocation. The dependency table for the Hawthorne model is shown below.
Figure 5 Dependency Table
To add a dependency from Excel, you can either click Resource Allocation | Add
Dependency from the menu or from the Dependency table right-click in a cell (row,
column or column, row) that will be dependent. We suggest the latter as it is only
one keystroke! When the dependency options appear, click the option or type a D, M
or X.
D - the row element depends upon the column element. The optimization assures
that the row element is not funded unless the column element is funded.
M - the row and the column are mutually dependent; both must be funded or neither
is funded.
190
X - the row and the column are mutually exclusive; funding one precludes funding
the other.
Three sample dependency options for the Hawthorne model are shown below. The
Dairy Chart Recorder and Solar Greenhouse are mutually exclusive. The Solar
Greenhouse and Dairy Maintenance Work are mutually dependent. The Bulk Tank
Frig depends on the Dairy Structure Work.
Figure 6 Sample Dependency Options
To delete a dependency, move to the cell (row, column or column, row) and press
the Delete key.
To clear the dependency table, from Excel select the Resource Allocation | Clear
Dependency Table.
191
Activity Level Resource Allocation

Optimization
We explain, by example, how to perform an activity level resource allocation
optimization.
1. Load the Hotspots model from the Expert Choice Samples directory. When
prompted for a participant name, just press Enter.
2. From Expert Choice the menu select Go | Data Grid.
3. Each alternative can have as many possible funding levels as desired each a
separate row in the data grid. The (common) alternative name is followed by a
COLON which is then followed by text describing the funding level for
example, bare bones, status quo, and gold plate. To designate activity levels you
must conform to the following convention; alternative name:activity level; for
example Suitland Nuisance Abatement:3000.
Review the names of the activities in the Data Grid and column one below. Note
each alternative with an activity level has the same alternative name followed by
a COLON and then the activity level. (If the alternatives column is not wide
enough make it larger by dragging.) Each activity level is associated with a
different cost.
4. The cost for each level is entered in the cost column corresponding to the row
for the level. It is customary to enter the levels in either increasing or decreasing
sequence. Typically (but not always), the total priority for levels with more
funding will have a higher total overall contribution toward objectives; see the
Total column.
5. Select Tools | Resource Allocation. You will be taken to an Excel spreadsheet

showing information that has been extracted from Expert Choice. Expert Choice
will formulate the optimization as an Activity Level optimization if Activity
Levels are defined as described above. The following figure shows the
information extracted.
192
Figure 7 Data Extracted from Expert Choice
6. You will be asked "Must one level be funded for each activity?" Select Yes.
(Select No if you do not want to allo w any levels to be funded for one or more
activities).
7. The initial display shows a solution where every activity is funded (note the
1s in Column B), and the total amount required to fund every alternative below
the red/orange cell in the Costs column.
8. Type your Budget Limit, in this example type 75000 in the red cell. Then press
Enter.
9. From the Excel Menu, select Resource Allocation. From the drop down menu
select Solve.
The following figure shows the results.
193
Figure 8 Results of Activity Level Optimization
Min and Max Columns in the worksheet; see the figure above. If you enter 1 in the
Min Column for an activity it must be funded. If you enter 1 in the Max Column for
an activity it will not be funded.
After viewing your result you can choose to Optimize for Increasing Budgets by
selecting this option from the Resource Allocation menu.
You can add constraint columns to the spreadsheet by selecting this option from
the Resource Allocation menu. The figure below shows a new column that has been
added and it also shows the results of the optimization based on the new entry.
194
Figure 9 New Column/Constraint Added to an Activity Level Optimization
To add a constraint use the figure above and follow the instructions below:
1. From the Resource Allocation menu select Add Constraint Column.
2. When prompted, type the number of columns to add, in this example type 1.
3. Next you must enter a title (text heading) for each new column. In this example
add the heading labeled Sup Wks for supervisor weeks.
195
4. Now enter the number of weeks required of the Supervisor for each activity as
shown in the figure below. Note some columns can be blank; none are in this
example.
5. Suppose that only 5 weeks of Supervisor time are available. Enter a 5 in the
max row of this new constraint column; see the figure above.
6. From the Resource Allocation menu, select Solve . If desired you can select
increasing budgets. The results with the new constraint will be displayed. Note:
The salmon color cells show the total cost and in this example the total
supervisor hours.
7. When you are finished using the optimizer you will be given the opportunity to
save your spreadsheet and then you will be returned to Expert Choice. We
highly recommend that you rename this spreadsheet to preserve the audit trail.
To learn about the Dependency Feature go to page 190.
196
Expert Choice Tutorials Appendix Area Validation Exercise
Appendix - Area Validation Exercise

Expert Choice provides decision-makers with the framework to derive ratio scale
priorities or weights from experience, insight and intuition as well as from hard data.
Expert Choice not only assists decision-makers in structuring complexity and
exercising judgment, but also incorporates both objective and subjective factors used
in the decision.
One of the advantages of the Expert Choice pairwise comparison process is that
weights, or priorities, are not arbitrarily assigned, but are derived from verbal,
numerical or graphical judgments. We know this method for deriving priorities is
reliable since it is based on a sound mathematical foundation. The AHP foundation
has been proven through numerous validation studies. Refer to
www.expertchoice.comfor books and technical references.
The following exercise demonstrates the validity of the results that Expert Choice
derives by showing that your results, obtained by using subjective judgments, come
close to known values.
This exercise should take approximately 15 minutes of your time and is most
worthwhile. It will help you actually visualize how Expert Choice assists decision-
makers in making more accurate decisions. The objective of the exercise is to make
judgments on the relative sizes of the areas (Figure 1 - Area Example) in pairs to find
the percentage that each shape contributes to the total area of the five figures.
What do we mean by relative sizes? In the example below area B is twice as big as
area A, area C is three times as big as area A. This is what is meant by relative size:
how many times bigger one shape is than another. Measuring things in relative terms
has a great advantage. The relative comparison is the same, no matter what scale is
used: inches, centimeters or something else.
Figure 1 Relative Size Example
197
Notice that the ratio of these decimals are meaningful; the two squares are twice as
big as the single square therefore, B/A = .333/.166 = 2. The three squares are three
times as big as the single square: C/A = .500/.166 = 3. So when you compare how
much bigger B is over A, you would use the number 2 from fundamental scale of
Expert Choice.
Table 2 Fundamental Scale of Expert Choice
The numbers in the fundamental scale are absolute numbers and independent of the
scale used. If you enter 2, you mean the larger shape is two times the smaller one. If
area B is 2 times bigger than area A, it is always two times bigger, regardless of the
units you use to measure the areas.
The purpose of this validation example is to find the decimal proportion of each area
in Figure 2 to the total area, using Expert Choice and your judgments. You will
perform a series of pairwise relative comparisons, first using the numerical mode and
then using the verbal mode. From your judgments Expert Choice will calculate the
relative areas of the shapes. You can then validate the results by comparing how
close they are to the actual measured areas.
In the course of the validation exercise you will fill out information in Table 1. In
column 1 you put your own estimate as described in step 1 below. In colu mn 2 you
will put results obtained from using the numerical comparison mode. In column 3
you will put results obtained using the verbal comparison mode. In column 4 you
will write the actual results.
Make a copy of Figure 2 and Table 1.
Jot down your unaided estimate (expressed as a decimal) of the proportion each
area is to the sum of the areas in Column A of Table 1. (To help you, assume
you are allocating a total of $1.00 to the shapes according to the size of their
respective areas.)
198
C
D
Figure 2 - Shapes Used in Area Validation Exercise
Table 1 Area Validation Chart to List Judgments
After performing the validation exercise compare the accuracy of your unaided
estimates with those that Expert Choice derived from your pairwise judgments.
199
Area Validation - Numerical

2. Select File Open and navigate to the Expert Choice Samples directory that
contains the Area model. Select Area to open the model.
3. Select File Save As to save the model under another name such as
MyAreaModel. This way the original model will be available should you want
someone else to perform this exercise.
4. Go to the Numerical Comparis on window by selecting . Then the first

two elements to be compared will be presented in the window. (There should be
no judgments in the matrix that appears at the bottom of the screen. If judgments
exist return to the ModelView and select Assessment, Erase Nodes for the Entire
Plex. After erasing return to the Numerical Comparison window)
Looking only at the two shapes being compared and using Figure 1 as an aid
visually estimate the ratio of the larger to the smaller. Keep in mind that for the
purpose of this area validation example, the comparative adjective "Importance"
is analogous to bigger.
For example, if you estimate that one area is 3 times larger than the other you
would click the number on the side of the indicator bar for the larger area.
(Alternatively, you could enter the judgment in the matrix below such as 4.5. If
you do this, remember to select the invert icon if the column element is
more important than the row element then the judgment will appear in red.)
Keep in mind that you must select the dominant side of the numerical indicator
when making judgments.
5. If you did not advance to the next judgment to be made, click the next vacant
cell in the matrix at the bottom of the window. Then make the next judgment.
6. Repeat this process until all judgments have been made.
7. After the last judgment is made you will be asked if you want to record your
judgments, select No.
8. Now select the calculate icon to display your derived numerical priorities.
200
9. Note: If your inconsistency is high (greater than 0.10) you might want to re-
evaluate your judgments as you probably made a clerical error. To do this, click
the Numerical tab and adjust your judgments changing only what you are
convinced is in error. For more information about inconsistency see
Inconsistency in Lesson 1.
10. Enter these derived priorities in Column B of the Area Validation Table that you
printed.
11. Return to the ModelView by clicking the ModelView tab .
Before continuing with the Verbal portion of this exercise we answer a few
commonly asked questions about comparisons and judgments.
Validation Questions
Q: Why are we interested in validating verbal comparisons?
A: Because almost every decision contains subjective or qualitative factors that do

not readily lend themselves to numerical judgments. It is difficult to justify
numerical judgments for the relative importance of clean air compared to clean
water. To overcome this impasse, Expert Choice uses the words EQUAL,
MODERATE, STRONG, VERY STRONG, and EXTREME. Words are often
easier to justify than numbers. For example, if you say that given our current
environment, cleaning up the air is 3 times more important than cleaning up the
water, can you justify why it is exactly 3? Why not 2.5, or 4.0? But if you
said, instead, that clean air is moderately more important than clean water, this
can be justified with a variety of arguments, including, perhaps, some hard
facts.
Q: Although verbal judgments are often easier to make and justify, how can
we combine them and be assured of their validity?
A: True, anyone can put arbitrary numbers behind words in a computer program
and then add the numbers. But with Expert Choice the same fundamental
numerical scale is used to derive ratio scale results for each group of
comparisons, and these ratio scale numbers can be combined to give
meaningful results.
201
Q: Even if the decision maker specifies numerical equivalencies for words,

will he/she, or a group of his/her colleagues consistently remember the
assignments accurately enough to insure that the results reflect their
judgments? Will errors due to the use of imprecise words be a problem?
A: Not with Expert Choice! Even if different individuals or a group attached

different interpretations to the verbal scale, the results will still be accurate, as
you will now see.
Area Validation - Verbal

1. From the ModelView, save your numerical example by selecting File, Save.
2. Select File, SaveAs and type MyAreaVerbal.
Now erase your numerical judgments and derived priorities.
1. From the Goal node, select Assessment.
2. Select Erase Judgments for Entire Plex and when prompted select Yes.
3. Go to the Verbal Comparison window by selecting . Then the first two

elements to be compared will be presented in the window.
Before entering any verbal judgments, read the following:
EQUAL, on the Verbal scale, means that neither alternative is more important than
the other, (i.e., bigger in this example).
EXTREME means an order of magnitude or more. Scientists define "an order of

magnitude" to be about 10 to 1. For example, if you were to compare the size of the
circle to the size of this letter "O", then the circle would be much more than an order
of magnitude larger. Extreme does NOT mean the largest compared to the smallest
(because the items being compared could all be about the same size). It means about
10 times more.
You may select judgments between the words on the verbal scale. For example:
between EQUAL and MODERATE, or between STRONG and VERY STRONG.
202
Keep in mind that you must select the dominant side of the verbal indicator when
making judgments.
Looking only at the two shapes being compared and using Figure 1 as an aid visually
estimate the dominance of the larger to the smaller only using WORDS.
1. Now enter a verbal judgment by clicking on the word that represents your
feelings.
2. If you do not advance to the next comparison, click the next vacant cell in the
matrix below.
3. Repeat this process until all judgments have been made.
4. When asked to calculate judgments, select No.
5. Now click the calculate icon to calculate and display the derived priorities.
Note: If your inconsistency is high (greater than 0.10) you might want to re-
evaluate your judgments as you probably made a clerical error changing only
what you are convinced is in error. To this, click the Verbal tab and adjust your
judgments. For more information about inconsistency see Inconsistency in
Lesson 1.
6. Enter these priorities in Column C of Table 1 that you printed.
7. Then click to return to the ModelView.
Results & Conclusion

Copy the actual relative areas shown below to Column D of Table 4 that you printed.
A = .471
B = .050
C = .234
D = .149
E = .096
Now compare your results as follows:
203
1. First compare your unaided estimates (Column A to Column D). Are they as
accurate?
2. Next compare your pairwise numerical results (Column B) with the actual
results (Column D). They should be very close, usually within 5% for every
figure.
3. Finally, check your pairwise verbal results (Column C) with the actual
measurements (Column D).
Pretty close? Probably amazingly close, given that they were derived from fuzzy
words! Your pairwise verbal results are probably not quite as good as your pairwise
numerical results because this is, after all, a numerical problem. But the accuracy
should be close enough to convince you of the valuable results you can obtain from
verbal comparisons.
When you performed the validation experiment with the verbal comparison mode,
you might have expected that your initial unaided estimates would be better than the
priorities derived from your pairwise verbal judgments. Were your verbal pairwise
results more accurate than your unaided estimates? If your results are typical of the
many people who have already tried this and similar validation experiments, you will
be surprised to find how accurate the priorities derived from your verbal judgments
actually are!
Conclusion
What does this show? That you can derive reliable results using judgments.
When you estimated the relative areas of these shapes without using Expert Choice,
your thought process probably included a scale of measure for each figure, perhaps
in combination with a mathematical formula. The unit of measurement might be
square inches, square centimeters, or perhaps your thumbnail. If you were to
actually use a measuring device, for example a ruler, your results would (aside from
measurement error) be perfect. However, you are able to derive relative measures
almost as well using your judgments.
Incidentally, direct measurement error is usually considerable, as you will find if you
try it using a ruler and some geometric formulas. Often both Expert Choice methods,
numerical and verbal, are better. Measurement error is often a factor when you are
dealing with data, too. Your judgments may give you results that are closer to
reality.
204
In solving this area problem using verbal judgments with Expert Choice, you did not
use a scale. The significance of being able to derive priorities from pairwise
judgments (verbal or numerical) is that accurate priorities can be derived for any
aspect of a decision no matter whether it involves quantitative, qualitative, objective
or subjective elements! If you can estimate physical quantities so well with Expert
Choice, based only on your judgment, it is likely that you are also able to estimate
priorities in situations that do not involve physical quantities for which no scale
exists, such as market appeal, satisfaction, risks, and intangible benefits.
205
Expert Choice Tutorials Appendix Hardware and Keypads
Setting Up the Hardware Receiver,

Keypads and Other Equipment
This section provides basic information for connecting and activating the hardware
(receiver and keypads) used in Expert Choice Group. We use the Reply Cordless
Response System. For detailed information consult the hardware documentation that
accompanied your hardware.
The Radio Frequency Receiver and keypads are produced differently for different
countries. You must specify the countries where you plan to use the receivers as the
radio frequency bands differ by country. A schematic is packed with the hardware.
Individual Wireless keypads come with the system. Each participant uses a keypad
to enter judgments or values. A maximum of 250 keypads can be used with the
system.
Projector There are numerous high quality projection systems available today. It
is important to test the software with the projector prior to a group session to
ascertain that the screen resolution of the computer and projector are compatible and
that the colors are displaying properly. In particular, check to see that the projector
contrast is set so that gray colors are visible and are not washed out.
What You Need to Do Before Connecting the Receiver

1. You will need one free serial port on your computer to connect the Reply
Cordless Receiver.
2. You will also need a video port if you use an LCD display panel, and a port
for the mouse. The cable for an external modem can be removed to free-up
one COM port. (Typically COM1 is configured for a mouse or an external
modem; check your computers hardware instructions to add and configure
additional serial ports.)
Connecting the Radio Frequency Receiver

1. Connect the RF receiver nine-pin female serial port labeled RS-232 (TO PC) to
an available serial port usually located on the back of a computer and typically
labeled COM. You must use the null modem cable that is supplied with your RF
206
receiver. If you have a 25 pin port on your computer you must obtain a DB9 to
DB25 (i.e., 9 to 25 pin) adapter from a computer supply store.
2. The cable coming from the interior of your RF receiver plugs into the front of
the RF receiver into the port labeled TX/RX DATA. This should already be
connected for you when the equipment arrives.
3. Plug the RF receiver into a 120 Volt electrical outlet. (At this time you may
desire to plug in your projection device). Turn on the power. Once the unit
receives power a red light will be displayed on the receiver.
Note: You can turn on the receiver before or after starting Expert Choice Group.
Programming the Keypads

Each keypad must be uniquely identified with its own internal number to enable
Expert Choice Group to identify each participants input. These numbers are easy to
program or change as explained below.
You should make certain that your keypads are properly labeled and uniquely
identified BEFORE conducting a group session. For example, if you have ten
decision-makers in the meeting, you will need ten keypads numbered from 1 to 10
internally, with corresponding external labels. It is also possible to have, for
example, 10 decision makers use 8 keypads by selecting the "wave" option in the
software and have eight people in the first wave, and two in the second wave. The
two in the second wave would use a neighbors keypad to make their judgments.
Determining the Current Internal Keypad Number

1. Press and hold down the CLR key on the keypad until the LED display on the
keypad flashes "C". This will take a few seconds.
2. Release the CLR key. Then you should notice that three numbers flash
consecutively on the LED display. If 0, then 1, then 5 appear, the keypad's
internal number is 015, or 15. Internal numbers always have three digits.
To Change the Internal Keypad Number
1. Press and hold down the CLR key on the keypad until the LED display on the
keypad flashes "C". This will take a few seconds.
207
2. Release the CLR key. Then you should notice that three numbers flash
consecutively on the LED display. If 0, then 1, then 5 appear, the keypad's
internal number is 015, or 15. Internal numbers always have three digits.
3. After the last number has flashed, enter the keypad number you desire. Keypad
numbers must have 3 digits. Zeros are entered using the ten key.
For example, if you want the internal number of the keypad to be number 5,
press the ten key once for the first 0 then again for the second 0, then press the 5
key. To enter the internal number 10, press the ten key, then the 1 key, then the
0 key to enter 010.
4. To check that the number you entered is correct repeat steps 1 and 2 above.
5. Now that you have changed the keypad number dont forget to change the label.
Programming the Keypad so the * Key Behaves as an

Invert Key
*
The asterisk, , on the keypad must be programmed so it can invert the order of a
judgment when making paired comparisons. Check each keypad making sure that the
LCD displays an "A" when the * is depressed. If this does not happened use the
above steps to program the * substituting the code 931 for a keypad number.
Common Keypad Troubleshooting

Questions
Problem: The keypads and receiver do not work.
Solution: Deactivate programs such as the Palm Pilot Hot Sync that use the
same Com port; then reboot and start Expert Choice.
Note: You can only have one instance of Expert Choice or Expert Choice's
Questionnaire and Brainstorming module running at one time. Close one instance
and reboot.
208
Problem: The invert key * isnt working.

Solution: Make sure it is programmed as described above. Note: The invert key
switches the dominance of a judgment from one factor to another.
Problem: Keypad appears to be "haywire"..
Solution: In the rare event that the keypad is not responding or appears to be
sending signals improperly, press and hold the clear key while
sequentially entering the three digits 981. This will reset the keypad
to the factory setting.
Note: If the keypad is dead, check the 9 volt battery. An indicator of

E on the keypad could also signal a weak battery.
Problem: There is a gap in the keypad number sequence or two or more

decision-makers appear to be using the same keypad number. (You
know the latter has happened when one decision announces that
someone is erasing his/her judgments.)
Solution: Check the keypad numbers and you will find a gap in the numbers or
that one or more keypads have been set to the same number. Change
the keypad numbers as described above.
Problem: Message: Keypads not available.
Solution: Make sure the receiver is connected to the computer and turned on.
Then determine if your Com port is set correctly. See Selecting
Keypad Options.
How the Hardware and Keypads are

Activated
Once the receiver is connected to the computer, you can turn on the Radio Receiver
either before or after starting Expert Choice Group.
209
It is good practice to group-enable the model prior to the session. If you have not
done this do it now.
Then either select:
o From the ModelView Assessment Pairwise Indi vidual or
o From the Data Grid Assess Indi vidual.
Then one of the two individual windows will appear. This is where the team
will enter judgments or data using the keypads.
From either individual window select File then select Keypad Options. Use
this to designate the number of keypads used in the session and to select the
facilitator keypad. It is a good idea to do this before the session starts.
Now that the keypad options have been set, make sure the software can see the
receiver. The status line at the bottom of the window should say Receiver. If
not turn on the receiver by selecting File, Recei ver; then select On.
Turn on the keypads by selecting File, Keypads On/Off. Look at the status
line: If the Keypads are on it should say Keypads; otherwise it will say mouse.
How the Keypads Are Used

Use to make YES or NO responses; select from A, B, C, D or E, or 1 through 9
multiple choice questions in Questionnaire and Brainstorming.
Use more importantly to enter paired comparisons from the ModelView; this
includes dragging bars to make graphical judgments.
Use to enter Ratings and other data values in the Data Gr id.
Designate a keypad as the Facilitator Keypad; this capability lets the facilitator
move around the room and not be tied down to the keyboard. See Selecting
Keypad Options.
210
Selecting Keypad Options

This window is used to set various keypad options. Use this window prior to each
session.
From this window you can set:
Designate the maximum number of keypads that will be used in your session.
Designate a Com port; the default is 1. You may need to reconfigure your
system to use Com 2, 3 or 4.
Designate a Facilitator keypad.
Designate the number of waves that can be used with the keypads. The default is
1.
If, for example, there are twelve participants and only eight keypads, you could
set the number of Keypads to 8 and the Wave Number to 1. When the first eight
participants have made judgments, four of the keypads will be passed to the next
"wave" of participants (persons 9 12). Before they made their judgments, the
facilitator would increase the wave number to 2. (This also requires that you
change persons 9 12 to wave number 2 in the Participant Table.)
Keypads off when you move If this box is checked the keypads will be turned
off when you move to a new judgment to be made or a new cell in the Data Grid
to be rated or valued. This feature permits the facilitator to control when
judgments or data will be entered.
Autokeypad If this box is checked then software will always attempt to turn on
the keypads when a model is loaded.
Anonymous If this box is checked then participants names, if entered are

masked to person 1, person 2 etc. Future participants will also be masked.
Templates button displays and prints the keypad templates as a participant aid.
One template can be used when making verbal or numerical paired comparisons
as well as entering Ratings or data values when at the Data Grid. The other
template can be used when making graphical judgments.
211
You can reach the Keypad Options window from any paired comparison window or
the Data Grid.
From ModelView select Assessment then select Pairwise.
Then select Assessment Indi vidual.
From the Data Grid select Assessment.
Then select Indi vidual.
Then from either individual window select File Keypads Options. You will be
taken to a window where you will enter the information just specified.
Tip: You may select any valid keypad number as the facilitator keypad;
for simplicity we suggest using number 1. If you select this option
many features can be accessed using this keypad rather than using the
keyboard. Note: The facilitator keypad remains active (on) whenever the
Receiver is on. (Should you use this option the facilitator keypad is not
available for participant use.)
Using Keypads Anonymously

If this option is selected then Person 1, Person 2, etc. will appear in a window
masking the actual participants names. This is by design so each participant can see
what judgment/value or vote was entered when all judgments/values/votes are
displayed.
This may not be sufficient for some groups as people will be able to tell who is using
what keypad. If this is the case, we suggest the following procedure:
Prior to the session change the keypad label for each keypad from numbers to
alphabetic characters. Mix up the characters; note for example, that keypad 2 is
now Q while JJ is 7.
Then pass out folded pieces of paper with each keypad telling the person what
their person ID really is. (Of course, participants can always identify the internal
212
number of their keypad by pressing and holding down the CLR button, but if
you tell them how to do this it might lead to them experimenting and changing
the IDs themselves.)
An even better way is to remove all keypad numbers. Pass out the keypads
randomly. Then in the middle of the session (or when ever desired) have the
participants switch keypads. This has two negatives no one will be able to
determine if they entered their votes correctly and you will not be able to get
accurate group demographics.
Default Facilitator Menus

Several facilitator menus for the ModelView's Assessment Pairwise Individuals
Window and the Data Grid's Assessment Individuals Window have been defined.
These menus are used by the facilitator with his or her keypad. We show the first
menu for each view.
Figure 1 ModelView's Assessment Pairwise Individuals Facilitator Menu
213
Figure 2 Data Grid's Assessment Individuals Facilitator Menu
214
Expert Choice Tutorials Appendix Expert Choice Terminology
Appendix of Expert Choice

Terminology
Absolute Measurement
Absolute measurement is used when the alternatives are compared against previously
established scales rather than relative to one another. See Relative Measurement.
Active Alternatives
Alternatives displayed in the ModelView's Alternatives pane are known as active
alternatives. They can only be inactivated but not deleted. See: Global Alternatives
Active Participants
Those team members who participate in the decision-making process. They have
been selected as a result of an SQL query and then the selection was applied to the
database. This feature enables one to define many participants and have different
participants make judgments about various facets of the model.
Alternatives
The alternatives are the options from which a choice is to be made. In a decision
model the alternatives can be evaluated in two ways:
Relative comparisons - alternatives are directly compared to one another using the
pairwise process.
Data Grid - alternatives are evaluated against a standard or scale that has been
developed.
See Active Alternatives and Global Alternatives.
Assessment
Assessment involves making pairwise comparisons between objectives, sub-
objectives, and alternatives of the decision. These are all made relative to the node
above.
Autoadvance
Automatically advances to next judgment to be made.
215
Autoredraw
One of two TreeViews Redraw commands; it is the default. When you move to a
node in the hierarchy, the tree will automatically be redrawn about this node showing
its descendants.
Best Fit
This command is used to examine inconsistency. It identifies the judgment that is
the most inconsistent and suggests what changes should be made to reduce
inconsistency from a mathematical perspective. As a decision-maker you must
determine if this is logical and what changes in judgment should be made versus
what is recommended.
Bottom-Up Structuring
A structuring process used to construct a model by first identifying alternatives;
second identifying pros and cons for the alternatives; third converting pros and cons
to objectives and sub-objectives. This approach is useful when more is known about
the alternatives of choice than the objectives. See Top Down Structuring
Child Node
A child node is a sub-objective of the node above it.
Cluster
A cluster consists of a node and its peers.
ClusterView
The ModelView's ClusterView pane is used to enter and cluster objectives and sub-
objectives into categories. Any structural changes made in the ClusterView will be
updated simultaneously in the ModelView's TreeView pane.
Comparisons - Modes of
There are three ways to enter pairwise comparisons: verbally, numerically and
graphically.
Complete Hierarchy
A complete hierarchy (model) has the same alternatives under each covering
objective (lowest level objective) in the hierarchy. This is the default. See
Incomplete Hierarchy.
216
Covering Objective
A covering objective is defined as the lowest level sub-objectives in the hierarchy
(TreeView). In the Data Grid, the covering objectives are displayed as columns
while the alternatives are displayed in the rows.
Criterion / Criteria
A criterion is a principle or a standard that an element is judged by and is often
represented as a lowest level sub-objective. Throughout Expert Choice, the words
criterion and objective are used interchangeably. The plural of criterion is criteria.
Distributive Synthesis
The distributive mode dis tributes the priorities of each covering objective among all
the alternatives, thereby dividing its global priorities proportionately to the priorities
of the alternatives beneath it. This method is used when all alternatives matter.
Data Grid
The Data Grid lists all alternatives in the model. Alternatives can be selected for
extraction to the Alternatives pane of the ModelView. The Data Grid is also used to
evaluate alternatives with respect to each covering objective in the model using one
of five different formulas: Ratings, Step function, Increasing or Decreasing utility
curves, and Direct entry of priorities.
Diagonal Pairwise
When making pairwise comparisons, one option is to make judgments on the
diagonal of the matrix. This feature uses the minimum number of comparisons
necessary to calculate priorities.
Diagonal Pairwise can also be used with the Data Grid to assess intensity scales that
were created with the formula types of Ratings or the Step function.
Formulas Grid
This grid displays the Data Grid's formulas per covering objective as well as its
parameters. Use this grid to copy and paste formulas from one objective to another.
Gap Analysis
By establishing the distance that every alternative ranks from the ideal alternative
you can perform a gap analysis ; this can be viewed from the Data Grid. In addition,
you can extract the ideal alternative and up to eight other alternatives from the Data
Grid to the Alternatives Pane of the ModelView and then view the performance
sensitivity graph for another form of gap analyses.
217
Geometric Average
The geometric average is the nth root of a product of n numbers. If you have two
numbers 6 and 24, the geometric average is 12 (6x24=144 raised to the power). If
you have three numbers, 1, 3, and 9, the geometric average is 3 (271/3 ).
Geometric Indicator
The Geometric Indicator on the Data Grid shows the amount of variance in a set of
group judgments. The more inconsistent the judgments the higher the indicator.
Global Alternatives
All alternatives displayed in the Data Grid are known as global alternatives. Only
global alternatives can be deleted from the model. If you want to include a subset of
the global alternatives in the ModelView's Alternative pane you must select and
extract them. See: Active Alternatives
Global Priority
The priority of each node relative to the Goal is called the global priority. The global
priorities of a node represent the portion of the parent's priority inherited by the
child. The global priorities of all the children equal the parent's global priority. The
global priority of a child equals the local priority of the child times the global priority
of the parent. See Local Priority and Priority.
Goal Node
The goal node represents the problem to be solved and appears at the top of the
hierarchy.
Grandchild
A grandchild node is a sub-sub-objective of two nodes above it.
Graphical Comparison Mode

When making assessments, you can use one of the three pairwise comparison modes.
When making graphical comparisons, comparisons for two elements are made by
moving judgment bars to the left or right. The upper portion of this window
graphically displays the two elements being compared with respect to the node
above. The lower portion of the window shows the numerical transformation of the
graphical judgment.
218
Hierarchy
A hierarchy is a tree-like structure that is used to decompose a decision problem. It
has a top-down flow, moving from general categories (objectives) to more specific
ones (sub-objectives and sub-sub-objectives).
Incomplete Hierarchy
An incomplete hierarchy (model) has different alternatives under each covering
objective. To convert a complete hierarchy model to an incomplete hierarchy you
would first select File, Convert to Incomplete Hierarchy; this will remove the
alternatives that are displayed in the Alternatives pane and place them below each
covering objective in the TreeView. Second delete any alternative that does not
belong under each covering objective. See Complete Hierarchy.
Inconsistency Ratio
A measure of inconsistency used in Expert Choice when making paired comparisons.
An inconsistency ratio is calculated for each set of judgments. Inconsistency follows
the transitive property, for example, if you were to say that A > B, and B > C, then
say that C > A, you would have been inconsistent.
Theoretical Information
The Inconsistency INDEX, not ratio, is calculated for each node (and its cluster of
children), and multiplied by the priority of the node, and summed over the entire
model. A similar calculation is done for the Inconsistency INDEX for random
judgments. The Overall Inconsistency Ratio is the ratio of these two weighted sums.
It has been shown that for any matrix small perturbations in the entries imply similar
perturbations in the eigenvalues; thus the eigenvalue problem for the inconsistent
case is:
A w = lambdamax w,
where lambdamax will be close to n (actually greater than or equal to n) and the
other lambdas will be close to zero. The estimates of the weights for the activities
can be found by normalizing the eigenvector corresponding to the largest eigenvalue
in the above matrix equation.
The closer lambdamax is to n, the more consistent the judgments. Thus the
difference, lambdamax - n, can be used as a measure of inconsistency (this difference
will be zero for perfect consistency). Instead of using this difference directly, Saaty
defined a consistency index as:
219
(lambdamax - n)/(n-1)
since it represents the average of the remaining eigenvalues.
In order to derive an accurate interpretation of either the difference or the

consistency index, Saaty simu lated a very large number of random pairwise
comparisons for different size matrices, calculating the consistency indices and
arriving at an average consistency index for random judgments for each size matrix.
He then defined the consistency ratio as the ratio of the consistency index for a
particular set of judgments to the average consistency index for random comparisons
for a matrix of the same size.
Since a set of perfectly consistent judgments produces a consistency index o f 0, the

consistency ratio will also be zero. A consistency ratio of 1 indicates consistency
akin to that which would be achieved if judgments were made at random rather than
intelligently. This ratio is called the inconsistency ratio in Expert Choice, since the
larger the value, the more inconsistent the judgments.
Information Document
Used to enter information about the goal, a particular objective, sub-objective or
alternative in the hierarchy. When using a group model this feature is only available
to the facilitator.
Intensity Scales
When using the Data Grid, Intensity Scales are used with either Ratings or the Step
function to measure data. These scales are often made up words such as Very Good,
Good and Poor. The scale is then pairwise compared deriving priorities by which the
alternatives are then compared.
Invert
This command switches the order of dominance between nodes in the pairwise
comparison process.
Ideal Synthesis
The ideal mode of synthesis distributes the full priority of an objective to the
alternative that ranks highest under that objective. The other alternatives are given a
priority in proportion to each alternative and the highest alternative. Use the Ideal
when you are only concerned about one alternative and the others will no longer be
relevant.
220
Judgment
A judgment expresses the strength of importance, preference or likelihood of one
element over another with respect to the node above. Usually the judgments are
made through a series of pairwise comparisons, although judgments can also be
made by entering priorities directly, using Assessment, Direct.
Judgments are also made from the Data Grid, by comparing each alternative with
respect to each covering objective's formula type.
Local Priority
The local priority represents the percentage of the parent node's priority that is
inherited by the child. The local priorities of the children of a node also sum to one.
See: Global Priorities and Priority
ModelView
The main Expert Choice window shows either the TreeView (hierarchy of
objectives), alternatives and an Information Document for the current node, or the
ClusterView. When in the ClusterView you can see all the objectives and
alternatives (top down structuring) or all the pros and cons and the alternatives that
they point to (bottom up structuring). For more information see the Tutorial
Structuring Lesson.
Node
Nodes represent the elements of a decision. A node may represent the goal, an
objective, a sub-objective or an alternative.
Note
Each node of a model can have a corresponding note. The note may include detailed
information that isn't in the definition, such as the information that was considered
for a judgment. It is most useful when the group feature is enabled because each
participant can enter his/her own comments. Notes can also be defined for each cell
in the Data Grid.
Numerical Comparison Mode

When making assessments, you can use one of the three pairwise comparison modes.
The Numerical Comparison Mode uses numbers (1-9) to compare two elements to
the element above. For example, we would say A is 3 times more important than B.
221
Objective
An objective is something that is sought or aimed at in a decision. The word,
objective, is sometimes interchangeably used with criterion or attribute. However it
is Expert Choice's position that an attribute will point to a criterion and a criterion
will point to an objective.
Pairwise Comparisons / Paired Comparisons

One of the major strengths of AHP is the use of pairwise comparisons to derive
accurate ratio scale priorities, instead of using traditional approaches of assigning
weights. This process compares the relative importance, performance or likelihood
of two elements with respect to another element in the level above. A judgment is
made as to which is more important and by how much. Pairwise comparisons are
carried out throughout an Expert Choice model to establish priorities. There are three
modes for pairwise comparisons: numerical, verbal, and graphical.
Parent Node
A parent node is any node in hierarchy that has children.
Plex
A Plex is defined to be the current node and all of its descendants.
Priority
The priority of a node is a numerical value represented as a percentage of one. It is
derived from pairwise comparisons with respect to the parent node. The local
priorities of the children of a node add up to one. The global priorities of a node
represent the portion of the parent's priority inherited by the child. The global
priorities of the children also sum to the parent's global priority. The global priority
of a child equals the local priority of the child times the global priority of the parent.
Redraw
The redraw commands redraw the hierarchy, in the TreeView, about the current
(highlighted) node showing its descendants.
Autoredraw, the default, will automatically redraw the hierarchy about the current
node when you move to it; it is displayed on the button bar with the left-hand icon.
Redraw only works if Autoredraw is off; it is shown on the button bar with the right-
hand icon. You must move to a node and double-click to expand or collapse that
portion of the hierarchy.
222
When at the ModelView's TreeView, depressing the first letter of a node name (or a
series of characters) will take you to that node. This feature is not available from the
Data Grid's TreeView.
Relative Measurement
In contrast to absolute measurement, relative measurement refers to the measurement
of the priorities of the objective, sub-objective and alternatives as determined by
comparing elements to each other. See Absolute Measurement.
Ratings
Ratings is a Data Grid method used to evaluate a large number of alternatives. Scales
of intensities are developed and prioritized; these measures are then used to judge
each alternative. Also see Intensity Scales.
Sensitivity Analysis
Sensitivity Analysis is used to investigate the sensitivity of the alternatives to
changes in the priorities of the objectives. There are five types of sensitivity
analyses: Performance, Dynamic, Gradient, 2D plot, and Head-to-Head. Analyses
can be performed from the Goal node or from the current node in the hierarchy such
as an objective.
Sibling Node
Nodes that appear at the same level in the model, relative to the parent node, are
known as siblings.
Step Function
The Step Function is one of five Data Grid Formula types and is similar to Ratings in
that it consists of a scale or prioritized intensities. However the Step function
automatically calculates the appropriate intensity for each alternative based on data
entered in the Data Grid rather the using a word from a scale. Also see: Intensity
Scales.
Structural Adjust
When making paired comparisons, the Structure button is used to structurally adjust
a set of priorities based on the number of the current node's grandchildren. This
feature should be used when you wish to prevent the dilution of the global weight of
a grandchild simply because it has many siblings.
223
Structuring
This feature of Exp ert Choice facilitates the hierarchy building process by: 1)
clustering objectives, or 2) defining alternatives and their associated pros and cons
and then converting them to objectives.
Sub-Objective (Sub-Criterion)
A sub-objective is a lower level objective of a parent.
Synthesis
This process shows the overall solution by synthesizing (combining) all global
priorities for each alternative and then presents the results in graphic format. A
synthesis can also be performed for other nodes in the hierarchy; this is known as an
intermediate synthesis.
TreeView
The term TreeView in this manual refers to an Expert Choice hierarchy containing
objectives and sub-objectives.
Top Down Structuring

The model is built, from the top, starting with the most general objectives, then the
more specific (sub-objectives), and finally the alternatives of choice. See Bottom Up
Structuring.
Utility Curves
Two utility curves available from the Data Grid include either Increasing or
Decreasing curves,) Like the step function, utility curves translate data into priorities.
Where as the step function consists of discrete steps, a utility curve is continuous. In
addition to being increasing or decreasing, either type can be linear or nonlinear, and
when nonlinear, concave or convex.
Verbal Comparison Mode

One of three paired comparison modes where you compare factors with respect to its
parent using words, such as: Equal, Moderate, Strong, Very Strong, and Extreme.
What-if
What- if analysis can be performed with the Sensitivity Analyses Graphs to
determine how the overall result would change if the priorities of the objectives were
changed.
224
Expert Choice Tutorials Index
Index
2
2D Plot ............................................................................... 74, 223
A
About the Expert Choice Tutorials................................................ 33
Absolute Measurement ......................................................215, 223
Access and SQL Database........................................................ 168
Accessing a Model Over the Internet or Intranet.......................... 181
Acitivity Level Resource Allocation ............................................. 185
Activating Hardware .................................................................. 209
Activating Keypads ................................................................... 209
Active Alternatives ....................................................... 26, 116, 215
Active Participants .................................................................... 215
Adding Alternatives .................................................................... 83
Adding / Entering Objectives........................................................ 79
Adding Alternatives in the Alternatives Pane................................. 50
Adding an Alternative to the Data Grid ....................................... 104
Adding Objectives to the TreeView .............................................. 46
Adding Participants to your Group Model.................................... 140
Adding Sub-Objectives in the TreeView........................................ 47
Advanced Optimization & Activity Level Resource Allocation ....... 192
Advanced Optimization & Discrete Activity Resource Allocation... 185
All Pros and Cons List................................................................. 88
Allowing New Participants to Open a Model Without Having Been
Pre-defined .............................................................................. 142
Alternatives .......................................................................215, 221
Alternatives in Data Grid ........................................................... 104
Alternatives pane ..........................................................................7
Area Model Validation Questions ............................................... 201
Area Validation - Verbal ............................................................ 202
Area Validation Exercise ........................................................... 197
Area Validation Numerical ......................................................... 200
Assessment ......................................................................215, 221
Assessment Menu Commands ...........................................157, 159
Assessments / Pairwise Comparisons Overview ........................... 51
Autoadvance ............................................................................ 215
Autoredraw.................................................................. 64, 216, 222
225
B
Best Fit ......................................................................... 57, 58, 216
Bottom-up Structuring ......................................................... 85, 216
Build the Model by Entering Objectives in the TreeView .................8
Building an Expert Choice Model ................................................. 43
Building the Hierarchy ................................................................. 89
C
Calculated Priorities .................................................................... 11
Child Node ............................................................................... 216
Cluster ..................................................................................... 216
Clustering Ideas.................................................................... 76, 80
ClusterView .......................................................................216, 221
Combining from the Data Grid ................................................... 167
Combining Judgments and or Data ............................................ 163
Combining Judgments for the Group from the Data Grid.............. 167
Combining Judgments for the Group from the ModelView ........... 166
Combining Judgments for the Group from the Participants Window
................................................................................................ 166
Common Keypad Troubleshooting Questions ............................. 208
Comparisons - Modes of ........................................................... 216
Complete Hierarchy .................................................................. 216
Conventions Used in the Tutorials................................................ 40
Converting Pros and Cons to Objectives ...................................... 90
Converting to an Incomplete Hierarchy ....................................... 219
Copying Formula Types ............................................................ 107
Copying judgments and/or data when adding new Participants .... 141
Covering Objective ................................................................... 217
Create a New Model .....................................................................5
Create a Ratings Model............................................................... 19
Creating a Group Model ............................................................ 139
Creating a Group Model & Adding Participants ........................... 139
Creating a Ratings Scale........................................................... 109
Creating a Step Formula From a Ratings Formula ...................... 122
Creating Decreasing Utility Curve .............................................. 131
Creating Increasing Utility Curve ................................................ 134
Criterion ................................................................................... 222
Criterion / Criteria ..................................................................... 217
D
Data Grid .............................................. 25, 102, 215, 217, 221, 223
226
Data Grid Overview..................................................................... 97

Decreasing Utility Curves .......................................................... 126
Default Facilitator Menus ........................................................... 213
Defining a Ratings Formula and Ratings Intensities....................... 20
Deleting a Formula Type ........................................................... 108
Deleting Nodes ........................................................................... 47
Deleting Participants ................................................................. 141
Dependency Feature ................................................................ 190
Diagonal Pairwise Comparisons ................................................ 111
Differentiating Between Objectives and Criteria ............................ 43
Direct ....................................................................................... 217
Direct Formula .......................................................................... 134
Discrete Alternative Resource Allocation .................................... 185
Distributive Synthesis.......................................................... 66, 217
Document Linking and Embedding............................................... 93
Dynamic Sensitivity.................................................. 68, 69, 70, 223
Dynamic Sensitivity Analysis ....................................................... 16
E
Ecsamples ......................................................................... 28, 100
Embedded Documents................................................................ 94
Enter the Alternatives in the Alternatives Pane................................8
Entering a priority with keypads ................................................. 152
Entering a Rating ...................................................................... 105
Entering a Ratings .................................................................... 161
Entering a Ratings with a Keypad .............................................. 150
Entering a Step Value ............................................................... 161
Entering a Step Value with a Keypad ......................................... 150
Entering a Value in the Data Grid............................................... 161
Entering Alternatives ................................................................... 83
Entering Alternatives in the Alternatives Pane............................... 50
Entering an `Ideal' Alternative .................................................... 118
Entering Data for a Utility Curve with the Keypad ........................ 151
Entering Objectives to the TreeView ............................................ 46
Entering Pros and Cons of the Alternatives ................................... 86
Entering Sub-Objectives in the TreeView...................................... 47
Entering Verbal Judgments for the Objectives............................... 52
Examining Active Alternatives Priorities ........................................ 27
Excel's Pivot Table..................................................................... 14
Expert Choice Group Overview.................................................. 137
Extracting Selected Alternatives from Data Grid to ModelView ..... 116
227
F
Facilitator Keypad..............................................................152, 153
Facilitator vs. Participant Functions ............................................ 138
Facilitator's Role ....................................................................... 137
Figure 1 - Area Example ........................................................... 199
Formula Type -Copying or Deleting............................................ 107
Formula Type None .................................................................. 135
Formulas Grid .......................................................................... 217
Fundamental Scale of Expert Choice ..................................198, 199
G
Gap Analysis............................................................................ 217
Geometric Average ................................................................... 218
Geometric Indicator................................................................... 218
Getting the Results ..................................................................... 65
Global Alternatives ....................................................... 26, 215, 218
Global Priority....................................................................218, 222
Global versus Active Alternatives ................................................. 26
Goal..................................................................................218, 221
Goal Node................................................................................ 218
Gradient Sensitivity............................................................. 72, 223
Gradient Sensitivity Analysis........................................................ 16
Grandchild ............................................................................... 218
Graphical Making Graphical Judgments .................................. 159
Graphical Comparison Mode ..................................................... 218
Graphical Judgments .................................................................. 59
Group Decision Making Using a Network or the Web .................. 155
Group Enabling a Model.....................................................139, 140
Group Model ............................................................................ 155
Group models opening ........................................................... 154
Grouping Nodes in the Hierarchy ................................................. 84
H
Head-to-Head..................................................................... 16, 223
Head-to-Head Graph .................................................................. 73
Hierarchy ................................................... 216, 217, 218, 219, 224
Hierarchy building..................................................................... 224
How to Activate Receiver and Keypads ...................................... 209
I
ideal alternative ........................................................................ 217
228
Ideal Synthesis .............................................................. 66, 67, 220

Improving Inconsistency.............................................................. 57
Incomplete Hierarchy ................................................................ 219
Inconsistencies ......................................................................... 216
Inconsistency ....................................................................... 57, 58
Inconsistency INDEX................................................................ 219
Inconsistency Ratio................................................................... 219
Increasing or Decreasing Utility Curves ...................................... 124
Increasing Utility Curves ............................................................ 132
Information Document pane ..........................................................7
Information Documents ...........................................92, 93, 220, 221
Information Documents and Notes ............................................... 92
Installing Expert Choice............................................................... 39
Installing Expert Choice.................................................................3
Intensity ................................................................................... 223
Intensity Scales ........................................................................ 220
Internet - Models on the Internet ................................................ 171
Introduction to Expert Choice....................................................... 35
Inverting Judgments.......... 55, 58, 147, 148, 149, 156, 157, 159, 208
inverting pairwise comparisons .................................................. 220
J
Judgment ................................................................................. 221
L
Linked Documents ...................................................................... 93
Local Priority .....................................................................221, 222
Logging into a Model on the Server............................................ 182
M
Make Verbal Judgments............................................................ 146
Making a Factor Dormant in the Pairwise Comparison Matrix ........ 64
Making Graphical Judgments .................................................... 159
Making Grapical Judments Using Keypads ................................. 147
Making Individual Judgments Using the Data Grid With Keypads . 149
Making Judgments for the Objectives ........................................... 13
Making Numerical & Graphical Judgments for the Alternatives ....... 12
Making Numerical Judgments.................................................... 157
Making Numerical Judments Using Keypads .............................. 148
Making Paired Comparisons ...................................................... 223
Making Pairwise Comparisons ..................................................... 51
229
Making Pairwsie Comparisons .......................................................9

Making Verbal Judgments ......................................................... 156
Making Verbal Judgments for the Alternatives ................................9
Making Verbal Judgments for the Objectives Overview ................. 52
Making Verbal Judments Using the Keypads .............................. 146
ModelView ................................................................ 6, 7, 216, 221
ModelView's Alternatives ........................................................... 215
Modifying an Existing Ratings Scale........................................... 114
Moving and Copying Nodes in the TreeView / Hierarchy................ 48
Moving Nodes in the TreeView /Hierarchy .................................... 48
Must Constraints shown on Data Grid ........................................ 135
N
Node........................................................................................ 221
Notes ................................................................................. 92, 221
Numerical - Making Judgments.................................................. 157
Numerical Comparison Mode .................................................... 221
Numerical Comparisons .............................................................. 60
Numerical Representations of Verbal Judgments.......................... 55
O
Object Linking and Embedding .................................................... 93
Objective................................................................... 221, 222, 224
Opening a Group Model ............................................................ 154
Opening a Group Model Using Keypads ..................................... 144
Opening a Team (Group) Model ................................................ 142
Optimization ............................................... 184, 185, 188, 190, 192
Other Data Grid Formula Types Overview .................................. 121
Overview of Expert Choice Features ............................................ 36
P
Paired Comparisons ................................................................. 224
Pairwise Comparisons ......................... 215, 216, 217, 218, 220, 221
Pairwise Comparisons / Paired Comparisons - Making Paired
Comparisons ............................................................................ 222
Parent Node............................................................................. 222
Participants .............................................................................. 221
Performance Sensitivity....................................................... 71, 223
Performance Sensitivity Analysis ................................................. 16
Plex ......................................................................................... 222
Post Command......................................................................... 167
230
Prioritizing Rating Intensities........................................................ 22

Priority ...................................................................... 218, 221, 222
Priority Fills .......................................................................... 62, 63
Programming the asterisk * key ................................................ 208
Programming the Keypads ........................................................ 207
R
Rating Alternatives with Rating Intensities .................................... 23
Ratings ............................................... 101, 105, 106, 217, 220, 223
Ratings - When to Use Ratings .................................................... 99
Ratings Approach Overview ........................................................ 18
Ratings Model ............................................................................ 19
Ratings Model Create ................................................................. 19
Receiver and Keypad Hardware Overview.................................. 206
Redraw ......................................................................... 63, 64, 222
Redraw Commands .................................................................... 63
Reduce Inconsistency ................................................................. 57
Refresh Commnad.................................................................... 167
Relative Measurement .......................................................215, 223
Renaming Nodes ........................................................................ 47
Resizing and Moving Around the Data Grid ................................ 102
Resource Allocation.................................................................. 184
Results of area validation .......................................................... 203
Revert .................................................................................. 49, 50
Revert Command ..................................................................... 108
Reverting to an Older Version of your Model................................. 49
S
Sample Models.............................................................. 28, 34, 100
Saving Your Model ..................................................................... 48
Selecting Alternatives from the Data Grid to Include in the
ModelView ............................................................................... 116
Selecting Keypad Options ......................................................... 211
Sensitivity Analyses Graphs ................................................ 68, 224
Sensitivity Analysis ............................................................. 16, 223
Setting Up a Model for use on an Internet or Intranet................... 172
Sibling Node............................................................................. 223
SQL
All Query .................................................................................. 164
SQL database .......................................................................... 168
SQL Databases .................................................................168, 170
SQL Particiapting Query ............................................................ 164
231
SQL Queries ............................................................................ 163

Starting / Creating a New Model (File) .......................................... 44
Starting Expert Choice ................................................................ 40
Starting Expert Choice ..................................................................4
Step Function ..................................................... 121, 217, 220, 223
Step Function Overview ............................................................ 121
Steps to Create a Decreasing Utility Curve from Scratch ............. 131
Steps to Create a Step Formula from Scratch............................. 124
Steps to Create an Increasing Utility Curve from Scratch............. 134
Structural Adjust ....................................................................... 223
Structured Query Language ...................................................... 163
Structuring .................................................................... 77, 78, 224
Theoretical Foundation................................................................ 76
Structuring Overview................................................................... 76
Sub-Objective........................................................................... 221
Sub-Objective (Sub-Criterion) .................................................... 224
Sub-sub-objective ..................................................................... 218
Synthesis ................................................................................. 224
Synthesis - Getting the Results.................................................... 65
Synthesize to Get Results ........................................................... 14
T
Table 1 - Area Validation........................................................... 199
Theoretical Foundation of Structuring ........................................... 76
Things to Check Before Opening a Team (Group) Model............. 142
Top-down Structuring.......................................................... 78, 224
Totals Column in Data Grid ....................................................... 103
Trash Can .................................................................................. 47
TreeView................................................................... 216, 221, 224
TreeView pane .............................................................................6
Two Dimensional Plot ........................................................... 16, 17
Two-Dimensional (2D Plot).......................................................... 74
U
Understanding the Ratings Intensities Just Entered....................... 25
Upsizing from Access to SQL .................................................... 168
Using Keypads Anonymously .................................................... 212
Using Keypads to Make Graphical Judgments .....................146, 147
Using Keypads to Make Numerical Judgments ........................... 148
Using Keypads to Make Verbal Judgments................................. 146
Utility Curves .....................................................................217, 224
232
V
Verbal Comparison Mode .......................................................... 224
Verbal judgments...................................................................... 156
Viewing the Hierarchy ................................................................. 83
Viewing the Totals Column........................................................ 103
W
Web Models and the Internet ..................................................... 171
Welcome dialogue box ................................................................ 45
What Expert Choice is used for.................................................... 38
What the Facilitator Does .......................................................... 144
What to do next? ........................................................................ 28
What-if ..................................................................................... 224
When to Combine Judgments and Data..................................... 163
233
234

Manual

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Manual

Uploaded by

Copyright:

Available Formats

Advanced Decision Support Software for the Millennium

Part I: Quick Start Guide

Software Copyright 1982-2002, Patents: 4,613,946; 5,995,728; 6,067,719; other

System Design: Ernest H. Forman, D.Sc.

Programming: Ernest H. Forman, D.Sc.

Deployment: Michael Korpics

Technical Support: John Zottola

Documentation: Mary Ann Selly

Due to our commitment to constantly improve our software, as well as the

Introduction / Installing / Starting Expert Choice ........................................ 3

Part II: EXPERT CHOICE TUTORIALS .................................................... 31

About the Expert Choice Tutorials............................................................ 33

About the Internet and Intranet Features .................................................. 37

Lesson 1 Building A Model ................................................................... 43

Synthesis Commands ............................................................................. 67

Lesson 2 Structuring - Another Way to Build A Model ............................. 76

Lesson 3 - Data Grid and Ratings ............................................................ 97

With the keyboard: ................................................................................ 106

Lesson 4 - Other Data Grid Functions .................................................... 121

Lesson 5 - Group Capabilities ................................................................ 137

Opening a Group Model Using Keypads ................................................. 144

Lesson 6 - Web Models and the Internet ................................................ 171

Lesson 7 - Resource Allocation ............................................................. 184

Appendix - Area Validation Exercise ...................................................... 197

Setting Up the Hardware Receiver, Keypads and Other

Appendix of Expert Choice Terminology ................................................. 215

Criterion / Criteria ................................................................................. 217

Top Down Structuring ........................................................................... 224

Part I: Expert Choice Quick Start Guide

Introduction / Installing / Starting

Installing Expert Choice

Insert the CD into CD drive.

If the CD does not automatically start then from Windows:

1. Select Start, and then select Run.

Tip: In this guide we use the words SELECT and CLICK

Starting Expert Choice

From the list select Expert Choice.

From the second list, select Expert Choice.

Building An Expert Choice Model

Figure 1 Welcome Dialogue Box

Next you must define your goal.

Figure 2 Sample Vacation Model

The Alternatives pane (top-right) shows the active alternatives.

Build the Model by Entering Objectives in

Select Edit; then select Insert Child of the Current Node .

Tip: To enter sub-objectives for a node, click an objective; then click

Enter the Alternatives in the Alternatives

Add Alternative button.)

Your model is now complete. Save it by selecting: File, Save.

Making Judgments / Pairwise Comparisons

Making Verbal Judgments for the Alternatives

comparing the alternatives with respect to the objective, Activities, the

Figure 3 The Verbal Comparison Window

of the slider when making your judgment.

Tip: The numerical representations of judgments displayed in black in

Figure 4 Derived Priorities of the Alternatives with Respect to Activities

Making Numerical & Graphical Judgments

select either the or tab to make judgments using one of the

Tip: When comparing the preference with respect to Costs, the

Making Judgments for the Objectives

Instead of responding yes to "Record Judgments and Calculate", click No.

Click the calculate button to display the priorities of the objectives.

the ModelView tab ; otherwise, click on the verbal comparison tab

Viewing Your Results - Synthesizing