12 views

Uploaded by Murali Dharan

d

- Abstracts Sim
- Modelling and Optimization of Process Parameters for Tig
- 6 Confounding
- a1
- Effectiveness of Dose of Drug Applied for the Treatment of Depression
- Experiential vs material
- Gage r&r
- 9-07 Factorial Planning Process
- briquetting
- IJERTV2IS1019
- Dawn on SLeep
- Mock Exam
- UT Dallas Syllabus for psy3392.501.07f taught by Betty-gene Edelman (bedelman)
- Bioelectric Impedance Reliability
- Good Moderator Demo
- Jmp Ttest Anova Examples
- Factorial PDF
- BA
- method of least aquares.ppt
- Overview of a Nova

You are on page 1of 33

E)

5.0. Introduction

In previous designs we considered that are basic designs of experiments which involved with

one factor. But in some situations these types of designs are not useful or fit to analysis. Hence

we need a design in which there is more than one factor of interest. For instance, Let us consider

two fertilizers say Potash (K) and Nitrogen (N). Let us suppose that there are p different varities

of Potash and q different varities of Nitrogen. Here the notations p and q are termed as levels of

the factors Potash and Nitrogen respectively.

To find the various effectiveness of various treatments K and N at different levels . for this we

might couduct two simple experients one for Potansh and the other for Nitrate. A series of

experiments which only one factor is varied at a time would be both lengthy and costly. Further

more , these simple experiemtns do not give us any information regarding dependence or

indepencence of one factor on the other. That is there will be no information about interaction

effect of potash and nitrate (NK).

Hence the only alternative is to develop an design that is to try to investigate the variation in

several factors simultaneously by conduting the above experiment . such a desing is Factorial

experiments .

5.1.

The experimental designs studied so far are useful to study the effect of a

single factor on the response variable. That is, the investigator is concerned

with testing several levels of one factor while keeping all other factors at a

constant level. In some experiments it becomes necessary to study the

effects of two or more factors so that their effect and cross-effects are tested

simultaneously.

Such

experiments

are

commonly

known

as

factorial

experiments.

A factorial experiment is an experiment in which the response variable is

measured at all possible combinations of the levels of the factors

(independent variables) in each complete trial or replication of the

experiment.

The

factorial

experiment

allows

one

to

vary

factors

Page S T A T 2 0 4 3

influence of individual factors. In general, in factorial experiment we are

interested in the effects of two or more factors on the response variable.

A full factorial experiment is an experiment with design consists of two or

more factors, each with levels, and whose experimental units take on all

possible combinations of these levels across all factors. In factorial design

you measure the response using various combinations of factors and levels

[Not vary-one-factor-and-keep-others-constant]. The goal is of factorial

experiment is to determine which factors have the largest effects on the

response, and whether there are interactions between factors.

Main effect and Interaction effects

The effect of a factor is defined as the change in response produced by a

change in the level of the factor. It is called the main effect because it refers

to the primary factors in the study. Consider the simple experiment in Figure

5-1. This is a two-factor factorial experiment with both design factors at two

levels. We have called this levels low and high and denoted them - and

+ respectively.

where

a1

and

a2

b1

B is constant or comparison of

b1

and

b2

b2

and

a1

are levels of

with

a2

while

Page S T A T 2 0 4 3

a1 b1a2 b 1a1 b2 a2 b2

a1 b1a1 b 2 a 2 b 1a2 b2

The

main

effect

or

.

of

is

either

and

the

main

effect

of

is

The main effect of factor A in the above two-level design can be thought of

as the difference between the average response at the low level of A and the

average

response

at

high

level

of

A.

numerically,

this

is

Figure 5.3 plots the response data against factor A for both levels of factor B.

Note that the B- and B+ signs are approximately parallel, indicating a lack of

interaction between factors A and B.

In some experiments, we may find that the difference in response between

the levels of one factor is not the same at all levels of the other factors (i.e.

the effects of factor A may depend on factor B.). In this case, we say there is

interaction between the factors. For example, consider the two-factor

factorial experiment shown in the figure 5-2 and the plot of the response

data against factor A at both levels of B is also shown in figure 5-4

Page S T A T 2 0 4 3

In figure 5-4 we see that the B- and B+ lines are not parallel, indicating

interactions between factors A and B.

The main effects of A and B and the interaction effect AB are:

.

Note that the AB interaction effect is the difference in the diagonal averages.

Advantages of Factorial Designs

1. They are more efficient than one-factor-at-a-time experiments

2. A factorial design is necessary when interactions may be present to avoid

misleading conclusions

3. Factorial designs allow the effects of a factor to be estimated at several

levels of the other factors, yielding conclusions that are valid over a range

of experimental conditions.

5.1

design)

Page S T A T 2 0 4 3

If the order in which the treatments are assigned to the units is not restricted

the design is a CRD. The model for this design is

random

and

and

is the

The hypothesis to be tested is:

Factor A

H 0 : 1= 2= a =0

Factor B

H 0 : 1= 2= b =0

H 1 : atleast one i 0

H 1 : atleast one j 0

H0:

H 1 : atleast one

a

( 1a2) and factor B has two levels

(b1 b2) . We call these levels low and high and denote by positive and

negative respectively. Let

a1

a2

is the

Page S T A T 2 0 4 3

b1

b2

be the

where

and

Treatment

Combinatio

Effects

ns

(1)

a

b

ab

The column headings are the

I

A

B

AB

+

+

+

+

+

+

+

+

+

+

main effects (A and B), the AB interaction, and

I, which represents the total or average of the entire experiment. Notice that

the column corresponding to I has only plus signs. The row designators are

the treatment combinations. To find the contrast for estimating any effect,

simply multiply the signs in the appropriate column of the table by the

corresponding treatment combination and add. For example, to estimate A,

the contrast is

the reactant and the amount of the catalyst on the conversion (yield) in a

chemical process. Let the reactant concentration be factor A, and let the two

levels of interest be 15 and 25 percent. The catalyst is factor B, with the high

level denoting the use of only 1 pound. The data is given below:

Reactant

Temperature Level

2

Concentration

1 Pound

Pounds

18,19,2

15 percent

25 percent

28,25,27

36,32,32

3

31,30,2

Page S T A T 2 0 4 3

9

In the 2 design the low and high levels of A and B are denoted by and +

2

Factor A: 15% (-) and 25% (+)

Factor B: 1 pound (-) and (+)

Factor

A

+

+

Replicate

B

+

+

Treatment

combinati

on

(1)

a

b

ab

I

28

36

18

31

II

25

32

19

30

III

27

32

23

29

Total

80

100

60

90

Note: The high level of any factor in the treatment combination is denoted by

corresponding lower case letter and that the low level of a factor in the

treatment combination is denoted by the absence of the corresponding

letter. Thus,

both factors at the low level. This notation is used throughout the 2k series.

In the two-level factorial design, we may define the average effect of a factor

as the change in response produced by a change in the level of that factor

averaged over the levels of the other factor. Also, the symbols (1), , and

Page S T A T 2 0 4 3

combinations as shown in the figure below.

level of B is

A:

Similarly the main effect of B is found from the effect of B at the low level of

A (i.e.

) as:

effect of A at high level of B and the effect of A at the low level of B. Thus

Page S T A T 2 0 4 3

of B at high level of A and the effect of B at the low level of A.

For the reactant concentration and temperature level example

increasing A from the low level (15%) to the high level (25%) will increase

the yield. The effect of catalyst (B) is negative this suggests that increasing

the amount of catalyst added to the process will decrease the yield.

Consider the sum of Squares for A, B and AB. Note that a contrast is used in

. Similarly

and

. Note that the three contrasts are Orthogonal.

The sum of squares for any contrast can be computed from the following

equation:

contrast squared divided by the number of observations in each total in the

contrast times the sum of squares of the contrast coefficients. Consequently,

Page S T A T 2 0 4 3

glass (Factor A) and temperature level (Factor B) on the lifetimes of batteries

(in hours). The following data is obtained:

Temperature Level

Type of Battery

15

25

20,70,82,

A

130,135,74,18

58

25,70,58,

B

150,188,159,126

45

Analyze the data and draw appropriate conclusions. Use =0.05.

Exercise: The nutritive value of certain edible fruit was measured in a total

of 24 specimens representing six specimens of each two varieties grown in

each two geographic regions taken at random.

Geographical region(A)

Page S T A T 2 0 4 3

1

Varieties

(B)

Total

I

6.9

11.

8

6.2

18.

2

19.

2

6.2

68.

5

2

II

13.4

I

8.9

II

9.1

14.1

13.5

9.2

5.2

13.1

13.2

13

7.7

8.6

12.3

13.7

7.8

5.7

9.8

9.9

80

44.5

63.7

a. Construct the ANOVA table and test the significance of varieties, region

and check whether there is significant interaction effect.

5.2.The advantage of factorial design

5.3. The General Two-Factor Factorial approach

The simplest type of factorial design involves only two factors. There are

Let

levels of factor B.

for the

Page S T A T 2 0 4 3

Example: An engineer is designing a battery for use in a device that will be

used subject to extreme variation in temperature. The only design parameter

that he can select at this point is the plate material for the battery and he

has three plate materials, and he has three choices. The engineer decides to

test all three plate materials at three temperature levels: 15, 70 and 125 0F.

Four batteries are tested at each combination of plate material and

temperature, and all 36 tests are run in random order. The data is given

below

Page S T A T 2 0 4 3

random

and

and

is the

and

and similarly

interest. Specifically we are interested in testing hypothesis about row

treatment effects (Factor A):

H 0 : 1= 2= a =0

H 1 : atleast one i 0

And the equality of column treatment effects (Factor B):

H 0 : 1= 2= b =0

H 1 : atleast one j 0

We are also interested in determining whether the two factors interact. That

is,

H0:

H 1 : atleast one

Page S T A T 2 0 4 3

The total sum of squares has been partitioned into a sum of squares due

to

factor A,

as

; a sum of

; and a sum of

The mean squares are obtained by dividing the sum of squares by the

corresponding degrees of freedom.

Page S T A T 2 0 4 3

The Analysis of variance table for the Two-Factor Factorial Design is shown

below:

Example: Consider the battery design experiment (Table 5.1). Analyze the

data and draw appropriate conclusions. Use =0.05.

Solution:

Page S T A T 2 0 4 3

significant.

The main effects of material type and temperature are also significant.

interaction is appropriate, say

Page S T A T 2 0 4 3

because the presence of significant interaction can have a dramatic impact

on the interpretation of the data.

The statistical analysis of a two-factor factorial model without interaction is

straight- forward and interaction sum of square contribute to the increase in

error sum of square as is given in the following ANOVA table.

5.3

design)

Suppose that three factors A, B and C, each at two levels are of interest. The

design is called a 23 factorial design and the eight treatment combinations

are

Page S T A T 2 0 4 3

geometry (B), and cutting angle (C) on the life (in hours) of a machine tool.

Two levels of each factor are chosen, and three replicates of a 2 3 design are

run. The data are given below:

Factor

Replicate

C

A

+

+

+

+

B

+

+

+

+

+

+

+

+

Treatment

combinati

on

(1)

a

b

ab

c

ac

bc

abc

I

22

32

35

55

44

40

60

39

II

31

43

34

47

45

37

50

41

III

25

29

50

46

38

36

54

47

Total

78

104

119

148

127

113

164

127

To find the main effects, interaction effects and to calculate the sum of

squares we make use of the following table of pluses and minus.

Page S T A T 2 0 4 3

To find the contrast for estimating any effect, simply multiply the signs in the

appropriate column of the table by the corresponding treatment combination

and

add.

For

example,

.

to

Thus

estimate

the

A,

main

the

effect

contrast

of

is

is:

Similarly:

Example: For the cutting speed (A), tool geometry (B), and cutting angle (C)

example find the main and interaction effects:

Solution:

Page S T A T 2 0 4 3

In the 23 design with n replicates, the sum of squares for any effect is:

Example: For the cutting speed (A), tool geometry (B), and cutting angle (C)

example find the sum of squares main and interaction effects:

Similarly:

Page S T A T 2 0 4 3

ANOVA table:

Source of Variation

A

B

AB

C

AC

BC

ABC

Error

Total

Sum

Squares

0.67

770.67

16.67

280.167

468.17

48.167

28.167

482.65

2095.33

of

df

1

1

1

1

1

1

1

16

23

0.67

0.022

770.67

25.548

16.67

0.553

280.167

9.288

468.17

15.52

48.167

1.597

28.167

0.933

30.165

This implies that the main effects of B and C and the interaction effect AC are

significant.

The General Factorial Design

The results for the two-factor factorial design may be extended to the

general case where there are a levels of factor A, b levels of factor B, c

levels of factor C, arranged in a factorial experiment. We can have

treatment combinations. Suppose that each treatment combination is

Page S T A T 2 0 4 3

replicated n times. If all factors in the experiment are fixed, we may test

hypotheses about the main effects and interactions. The test statistics for

each main effect

corresponding

and

interaction

is constructed

by dividing

the

by the mean

square error. The number of degrees of freedom for any main effect is

the number of levels of the factor minus one, and the number of degrees

of freedom for an interaction is the product of the number of degrees of

freedom associated with the individual components of the interaction.

For instance the three-factor analysis of variance mode is given by:

Assuming that A, B and C are fixed the Analysis of variance table is given

below:

Assuming the design is CRD, the sum of squares can be obtained as:

Page S T A T 2 0 4 3

The mean squares are obtained by dividing the sum of squares by the

corresponding degrees of freedom.

Carbon

(A)

10

12

operating

25psi

line

speed

25

200 0

-3

-1

-1

0

0

2

1

1

Pressure (B)

30 psi

line

speed( C)

200

-1

0

2

3

A x B summary table

Carbon

Operating Pre. (B)

250

1

1

6

5

(A)

25

30

10

12

14

Total

-5

4

22

21

1

16

37

54

Total

-4

20

59

75

Page S T A T 2 0 4 3

5

4

14

7

6

7

9

10

11

A x C summary table

Carbonatio

n (A)

10

12

14

Total

Line Speed

20

0

250

-5

1

6

14

25

34

26

49

B x C summary table

Line Speed

Total

-4

20

59

75

Operating

Pressure (B)

25 psi

30psi

Total

200

6

20

26

250

15

34

49

Total

21

54

75

The sums of squares for the two-factor interactions were calculated from

two-way cell totals.

Page S T A T 2 0 4 3

cell totals.

Page S T A T 2 0 4 3

This shows that the percentage of carbonation, operating pressure, and line

speed significantly affect the fill volume.

The

Factorial Design

refers to a special case of the general factorial design in

The two levels are usually called low and high (could be either quantitative

or qualitative).

Assumptions: (1) the factor is fixed, (2) the design is completely

randomized and (3) the usual normality assumptions are satisfied

is particularly useful in the early stages of experimental

work, when there are likely to be many factors to be investigated. It provides

the smallest number of runs with which k factors can be studied in a

complete factorial design.

5.4

The linear statistical model for this design is

Page S T A T 2 0 4 3

Where

particular block is required and a single replicate of a complete factorial

experiment is run within each block. The model now is

where

is the effect of

complete block.

Page S T A T 2 0 4 3

Both ground clutter level and filter type are significant at the 1 percent level

of significance whereas their interaction is not significant only at the 1

percent level.

Factorial experiment with single replication

If there are two factors and only one observation per cell, the effects

model is

Here the error variance is not estimable; that is, the two-factor interaction

effect

Page S T A T 2 0 4 3

there are no tests on main effects unless the interaction effect is zero. If

To test whether there is an interaction, one may use the following possibility

developed by Tukey with one degree of freedom.

Page S T A T 2 0 4 3

Page S T A T 2 0 4 3

is

no evidence

of

interaction

in

5.5

Confounding is a design technique for arranging a complete factorial

experiment in blocks, where the block size is smaller than the number of

treatment combinations in one replicate. The technique causes information

about certain treatment effects (usually high-order interactions) to be

indistinguishable from, or confounded with, blocks.

Confounding the 2k Factorial Design in two Blocks

Suppose that we wish to run a single replicate of the 22 design. Each of the 22

= 4 treatment combinations requires a blocks. For example, if each block is

only large enough for two treatment combinations to be tested. Thus, two

blocks are required and we must assign two of the four treatment

combinations to each block. The treatment combinations having opposite

signs on the higher interaction effect are assigned to different blocks i.e.

block 1 contains the treatment combinations (1) and ab and that block 2

contains a and b. The order in which the treatment combinations run within a

block is randomly determined.

Page S T A T 2 0 4 3

Because the two treatment combinations with the plus sign [ab and (I)]

are in block 1 and the two with the minus sign (a and b) are in block

2, the block effect and the AB interaction are identical. That is, AB is

confounded with blocks.

All treatment combinations that have a plus sign on AB are assigned to block

1, whereas all treatment combinations that have a minus sign on AB are

assigned to block 2.

three- factor interaction ABC with blocks.

Fr

om the table of plus and minus signs assign the treatment combinations

that are minus on ABC to block 1 and those that are plus on ABC to

block 2 and the treatment combinations within a block are run in random

order.

Page S T A T 2 0 4 3

Page S T A T 2 0 4 3

- Abstracts SimUploaded byNikesh Doshi
- Modelling and Optimization of Process Parameters for TigUploaded bybalajigandhirajan
- 6 ConfoundingUploaded byAkshay Goyal
- a1Uploaded byYang Zhou
- Effectiveness of Dose of Drug Applied for the Treatment of DepressionUploaded byIJSRP ORG
- Experiential vs materialUploaded bysasha
- Gage r&rUploaded bymassman02143
- 9-07 Factorial Planning ProcessUploaded byDorababu Adapa
- briquettingUploaded byVishnu Vardhan Sidlagatta
- IJERTV2IS1019Uploaded byrumabiswas853
- Dawn on SLeepUploaded byJeanHaydeeRipdos
- Mock ExamUploaded byEddie Martinez Jr.
- Good Moderator DemoUploaded bySOURASHTRA
- UT Dallas Syllabus for psy3392.501.07f taught by Betty-gene Edelman (bedelman)Uploaded byUT Dallas Provost's Technology Group
- Bioelectric Impedance ReliabilityUploaded byspongebobfishpants
- Jmp Ttest Anova ExamplesUploaded byAjibola Alade
- Factorial PDFUploaded byAndrew
- BAUploaded byMihir Krampp
- method of least aquares.pptUploaded byRichard Obinna
- Overview of a NovaUploaded byHii King Fong
- DOE Wizard - Multilevel Factorial DesignsUploaded byAnonymous FZNn6rB
- 1Uploaded byNikos Konstantinou
- Class 17-2-3 Factorial DesignUploaded byb_shadid8399
- UltimatumUploaded byIsaias Morales
- Optimization of Photochemical MachiningUploaded byatul_saraf001
- Chiradejnant 2Uploaded byJuan Pablo Omonte Vera
- N-P_Stoichiometry-in-atmosphereUploaded byrazvan_matei
- Lecture 5 Final_ Analysis of CovarianceUploaded byteklay
- Hw6 SolutionUploaded byMua Lanh
- RepeatabilityUploaded byArish Adil

- Nerd's Notes_ How We Did the ClinicalTrials.gov Data AnalysisUploaded byMurali Dharan
- Output and Discussion for the Certification ProjectUploaded byMurali Dharan
- Confidence Intervals for Effect Sizes_ Applying Bootstrap Resampling - Practical Assessment, Research & EvaluationUploaded byMurali Dharan
- Bootstrap Sample_ Definition, Example - Statistics How ToUploaded byMurali Dharan
- 6.1.6. What is Process CapabilityUploaded byMurali Dharan
- 7. List in R_ Create, Select Elements With Exampled1p7Uploaded byMurali Dharan
- Statistical Process Control MC QuestionsUploaded byMurali Dharan
- 6.3.2.1. Shewhart X-bar and R and S Control ChartsUploaded byMurali Dharan
- AdvisorUploaded byMurali Dharan
- Worksheets ( Downloads ) - WestgardUploaded byMurali Dharan
- Different Types of Reliability.pdfUploaded byMurali Dharan
- zxzxUploaded byMurali Dharan
- https.docxUploaded byMurali Dharan
- 9783319160610-c2Uploaded byMurali Dharan
- Monitoring the Performance of Bayesian EWMA Control Chart Using Loss FunctionsUploaded byMurali Dharan
- Different Types of ReliabilityUploaded byMurali Dharan
- Montgomery Ch09 Sec02 03Uploaded byMurali Dharan
- Course OutlineUploaded byMurali Dharan
- TQM Review Lecture 2010Uploaded byVicky de Chavez
- Course outline.docxUploaded byMurali Dharan
- Factor Analysis in Social ScienceUploaded byMurali Dharan
- a154b413-793d-4b47-9b43-54c58edc270d (1) (1)Uploaded byMurali Dharan
- Chapter IV Acceptance SamplingUploaded byMurali Dharan
- Course Outline SQCUploaded byMurali Dharan
- Decision Flow Chart for Double Sampling PlanUploaded byMurali Dharan
- Chapter I Introduction.pdfUploaded byMurali Dharan
- Forest plot - Wikipedia.pdfUploaded byMurali Dharan
- lecturenotes1.pdfUploaded byMurali Dharan
- Analysis of Longitudinal and Survival Data Joint MUploaded byMurali Dharan
- Ch Analysing Longitudinal DataIUploaded byMurali Dharan

- Monoolein a Review of the Pharmaceutical ApplicationsUploaded byalinerhc
- Fiber Solutions CatalogUploaded bygladwyn_hallatu
- Mach Number CalculationUploaded byOlusegun Oyebanji
- RDGerroranalUploaded bysjjbio
- 2 Exercises.ppsUploaded bySeverina Mallari
- revmodel.docUploaded bySukddesh Ragavan
- Seal Application Data Sheet | Mechanical Shaft SealUploaded bylubangjarum
- 802_1_1_TRANSMISSION LINE TOWER.pdfUploaded byleua_690001837
- Cloze MetalsUploaded byDeepa Shreenath
- Atomic Structure IitUploaded byArnab Barman Ray
- Pages from STC processing.pdfUploaded byKP
- Inclined Plate ClarifierUploaded byKelvin Siew
- LCI Install GuideUploaded bygonzalo
- BijectionUploaded byHameed Khan
- Electrical Sample ReportUploaded byeric
- insitu combustion projectsUploaded byparthkapadia1234
- liquid chromatography sop - aug 7Uploaded byapi-267902588
- Manual of Met All Ur 00 Maki RichUploaded byAnonymous M5v9mAU
- water structure and properties.pdfUploaded byfajar_setia
- GISPL22Uploaded bynedalina
- Schuster P. - Stochasticity in Processes. Fundamentals and Applications to Chemistry and Biology - (Springer Series in Synergetics) - 2016Uploaded byfictitious30
- 84010292Uploaded byDaniel Manole
- etag-029-april-2013 (1)Uploaded byadnan
- Gelombang Cahaya Tampak (Visible Light)Uploaded byDwipayana Sang Pembunuh
- Tutorial RPSUploaded byJuan Calderas
- 5400348b67d92Uploaded byinaca4
- En 13190Uploaded byzakaria_hsn
- A Transient BEM for Room AcousticsUploaded byLuis de Freitas
- 6.Catalogue - VTS Fan Coil UnitUploaded byHaji Ali
- 1-s2.0-S0017931016342892-mainUploaded byWendy Herrera