You are on page 1of 117

STEPSsampling.

xls is an excel workbook that will guide you through the process of selecting the sample for you
to use this spreadsheet you will need to gather some supplementary information, including:
1. Population information for the top level of sampling (the PSUs).
2. Population information information about all SSUs, TSUs, and so forth until you reach the househo

General Guidelines: This workbook used macros and other complex equations. These equations must not be changed o
cause the spreadsheets to malfunction. Please follw the simple guidelines below when using the spredsheets:
1. Only enter or delete information in the green sections of the workbook. (The same color as this cell.)
2. Do not move or delete columns in the spredsheets, most empty columns actually include formulas and if you enter infor
delete the formula.
3. If you need to add more information you need to insert a column into the workbook. If you add a column you need to tu
can be easily identified.
4. You can never add another row in the spreadsheets.
5. Do not worry about #Value or #N/A errors in the spreadsheets. The sheets are anticipating 500 sampling units. If thes
then an error is returned. This will have no impact on your information.
6. You can only enter 500 PSU/clusters/households on the spreadsheets and from these select only 100 PSUs/SSUs/hou
enter or select more than the designated number, contact the STEPS team for assistance.
7. If you have any problems at all contact the STEPS team at steps@who.int .

Using STEPSsampling.xls:
1. Fill in the PSU spreadsheet as indicated with the names and population sizes of all possible PSUs.
2. Follow the directions on the PSU spreadsheet to select the PSUs for your sample.
3. For each selected PSU, you will create an 'SSU' spreadsheet which will be used to select the SSUs from each PSU. Fo
the SSU spreadsheet for duplicating this spreadsheet for all selected PSUs to select the SSUs within each PSU.
4. For each selected SSU, you will create a 'RandHold' spreadsheet which will be used to select individuals or households
information available) within each SSU. Follow the instructions on the 'RandHold' spreadsheet for duplicating this spreads
SSUs and selecting individuals/households within each.**(See note below)
5. Follow the instructions on the Info for Weighting spreadsheet to calculate the weights for your data. The 'Format for Epi
spreadsheet entitled 'IndWeight'. You will need to complete the Stratum and PSU columns in this spreadsheet.
6. Fill in the Population Info spreadsheet.

After completing the above steps, fill in the header of your Interview Tracking Forms and, if needed, the Kish hous
7. For every interview tracking form, record the appropriate Cluster Number from the Rand Hhold spreadsheet in the top ri
form. That is, the Cluster Number copied to the form should correspond with the cluster in which all participants on the for
8. If the Kish method will be used, copy the Cluster Number to all Kish household coversheets for each cluster.

**Note: It is possible that your sampling design will require more levels of sampling than are provided for here. If this is the
the STEPS team at STEPS@who.int
of selecting the sample for your STEPS survey. In order
n, including:

h until you reach the household or village level.

quations must not be changed or altered, doing so will


ing the spredsheets:
r as this cell.)
de formulas and if you enter information into them you will

ou add a column you need to turn the column green so it

ating 500 sampling units. If these values are not there

select only 100 PSUs/SSUs/households. If you need to

sible PSUs.

ct the SSUs from each PSU. Follow the instructions on


SUs within each PSU.
select individuals or households (depending on the
heet for duplicating this spreadsheet for all selected

r your data. The 'Format for Epi Info' button will create a
s in this spreadsheet.

s and, if needed, the Kish household coversheets:


Hhold spreadsheet in the top right-hand corner of the
which all participants on the form reside.
eets for each cluster.

re provided for here. If this is the case, please contact


ONLY ENTER INFORMATION INTO THE DESIGNATED GREEN AREAS. DO NOT ADD OR REMOVE COLUMNS/ROWS. IF YOU NEED ASSISTANCE CONTACT THE STE

Directions on using this spreadsheet:


1. Type the names of the PSUs in column B in the green space below.
2. Type the estimated population for each PSU in column C in the green space below.
3. Enter the number of PSUs you want to select in D11.
4. Type the random number from G10 into J10 ONLY ONCE (the random number is generated each
time the spreadsheet is activated, this means that G10 and J10 will not always match).
5. The selected PSUs will have a YES in column D.

Random
Total population 27,218 555
number (r)
Number of PSUs to Sampling
7 3,888
select (maximum 100) interval (k)

PSU Names (or abbreviations) of Estimated size of probability of


Selected
Number your primary sampling units sampling units inclusion
1 La Ceiba 1967 YES 0.5059
Choluteca 965
2 Choloma 1195 YES 0.3073
La Lima 720
Olanchito 406
Puerto Cortes 891
El Progreso 1261
3 Danli 417 YES 0.1072
Comayagua 622
4 San Pedro Sula 9090 YES 1.0000
5 Santa Rosa de Copan 319 YES 0.0820
6 Tegucigalpa 7210 YES 1.0000
7 Tela 1048 YES 0.2695
Trujillo 485
Villanueva 622
. IF YOU NEED ASSISTANCE CONTACT THE STEPS TEAM.

Type the random number


3
from G10 (only 1 time)

Number of PSUs available 15

2,785
ONLY ENTER INFORMATION INTO THE DESIGNATED GREEN AREAS. DO NOT ADD OR REMOVE COLUMNS/ROWS.
IF YOU NEED ASSISTANCE CONTACT THE STEPS TEAM.

more spreasheets. To do this follow the


You have selected: 7 PSUs to sample in the PSU spreadsheet. Complete this 6 instructions titled "Copying Spreadsheets" in
spreadsheet for each cluster selected. This workbook contains Instructions workbook
one "Clustering SSU" spreadsheet, you need to create:

Directions on using this spreadsheet:


1. Insert the name of the first PSU selected in the PSU spreadsheet in B15.
2. Type the names of the SSUs for this PSU in column A in the green space below.
3. Type the estimated population for each SSU in column B in the green space below.
4. Enter the number of SSUs you want to select for the cluster in D14.
5. Type the random number from G13 into J13 ONLY ONCE (even if later on G13 and J13 do not match).
6. The selected clusters will have a YES in cloumn C.
7. If you need to select units from another cluster, click the Duplicate spreadsheet button and repeat steps 1-5 until all the
selected PSUs from the PSU spreadsheet have been used.

Type the random


Total population within cluster 1,967 Random number (r) 120 number from G13 (only
1 time)
Number of SSUs
Number of SSUs to select 1 Sampling interval (k) 1967
available
Insert the name of this PSU
from PSU spreadsheet La Ceiba
Names (or abbreviations) of Estimated size of probability
Selected
secondary sampling units sampling units of inclusion
La Ceiba 1967 YES 1.00
sheets. To do this follow the
titled "Copying Spreadsheets" in the
workbook

1
ONLY ENTER INFORMATION INTO THE DESIGNATED GREEN AREAS. DO NOT ADD OR REMOVE COLUMNS/ROWS. IF YOU NEED ASSISTANCE CONTACT THE STE
NOTE: Remember you need to fill out this spreadsheet for each selected SSU by creating a duplicate of this spreadsheet.

Directions on using this spreadsheet:


1. Enter the address (or identification number) of each household/individual in
column A. Number of households/Individuals to select 6
2. Enter the number of households/individuals you want to select in F5 and the
Cluster Number in B9. Number of households/Individuals in list 6
3. If you need to select households (individuals) from another SSU or age/sex
group, click the Duplicate spreadsheet button. Probability of selection for households/Individuals : 1.00000

Create Cluster Number (will be used as identifier on


interview tracking form and Info for Weighting worksheet) 7
Address/ Indentifier of the households/Individuals Selected
La ceiba YES
choluteca YES
danli YES
san pedro sula YES
santa rosa de copan YES
tegucigalpa YES
TANCE CONTACT THE STEPS TEAM
ONLY ENTER INFORMATION INTO THE DESIGNATED GREEN AREAS. DO NOT ADD OR REMOVE COLUMNS/ROW
CONTACT THE STEPS TEAM

This spreadsheet is used to enter the estimated population structure. It is used during the weighting process.

Directions on using this spreadsheet:


1. Fill out the table below (do not use any punctuation, just type the estimated population size e.g. 195648).
2. Click on the Format for Epi Info button.

Estimated Population Size proportion of


Sex AgeRange (target population) population
Male 25-34 #DIV/0!
Male 35-44 #DIV/0!
Male 45-54 #DIV/0!
Male 55-64 #DIV/0!
Female 25-34 #DIV/0!
Female 35-44 #DIV/0!
Female 45-54 #DIV/0!
Female 55-64 #DIV/0!
total population 0
DO NOT ADD OR REMOVE COLUMNS/ROWS. IF YOU NEED ASSISTANCE

is used during the weighting process.

d population size e.g. 195648).


ONLY ENTER INFORMATION INTO THE DESIGNATED GREEN AREAS. DO NOT ADD OR REMOVE COLUMNS/ROW

Directions on using this spreadsheet:


1. Enter the Cluster Number for each Rand Hhold spreadsheet in column A.
2. Enter the cluster name, if available, in column C.
3. If the SSU spreadsheet was used, enter the "probability of inclusion" (in column D of the associated SSU
spreadsheet) for each SSU.
**IMPORTANT: Make sure to use the SSU spreadsheet that is associated with the Cluster Number.
4. Enter the "Probability of selection for..." from F7 from the associated Rand Hhold spreadsheet in column G.
** IMPORTANT: Make sure to use the Rand Hhold spreadsheet that is associated with the Cluster Number
5.
6. Once all thenumber
Type in the of the
weighting associated
information hasPSUbeeninentered
columnfor
B.allThis
the number is in column
cluster numbers, clickAon
ofthe
the"Format
PSU spreadsheet.
for Epi Info"
button.
7. You will need to manually assign the values for the sampling design in the database. Use the spreadsheet IndWeight
to create these values.

Cluster Associated Prob. of Selection


Number PSU Number Name of Cluster, if available from SSU
La ceiba 1 La ceiba
choluteca 2 choluteca
danli 3 danli
san pedro 4 san pedro sula
santa rosa 5 santa rosa de copan
tegucigalp 6 tegucigalpa
ADD OR REMOVE COLUMNS/ROWS. IF YOU NEED ASSISTANCE CONTACT THE STEPS TEAM

of the associated SSU

with the Cluster Number.


preadsheet in column G.
ociated with the Cluster Number
nkAonofthe
the"Format
PSU spreadsheet.
for Epi Info"

e. Use the spreadsheet IndWeight

Prob. of Selection Prob. of Selection for Ind Prob. Of Selection


from Rand Hhold PSU Minus Int Tracking
0.505878462781983 0.5058784628
0.307333382320523 0.3073333823
0.107245205378793 0.1072452054
1 1.0000000000
0.082041296201044 0.0820412962
1 1.0000000000
HE STEPS TEAM
Instructions for assigning Stratum and PSU values for analysis:
(examples follow instructions below)

FIRST, SORT YOUR DATA BY PSUchance:


Step Action
1 Scroll down until you find the end of your data.
Click the mouse in the left-most cell of the last line of your data and drag across to the right-
2 most cell of your data. Continue holding down the mouse button and drag the mouse back up
until the cursor is over D1.
Release the mouse button. You should now see a big purple/blue box that contains all of your
3
data spanning columns A through D.
4 From the Menu select "Data", "Sort".
Select PSUchance in the "Sort by" drop-down list and select the option "Descending" to the
right of this drop-down list. Next select "SamplePSU" from the "Then by" drop-down list and
5 select the option "Descending" to the right of this drop-down list. Finally, select "Header row"
at the bottom of the window then click "OK".

***IMPORTANT:
How the Stratum and PSU values are assigned is dependent upon the ordering of the sampling units entered
into the PSU spreadsheet. If the list of units was ordered either by size, location, or some other
characteristic that may make the sampling units more similar to units placed closer on the list, then follow
Option A below. If there was no order to the list of sampling units entered into the PSU spreadsheet, then
follow Option B below. Option B is a more conservative option and should be followed if there is any
doubt about the ordering of the sampling units in the PSU spreadsheet.

NEXT, ASSIGN THE STRATUM:


OPTION A
STEP Action
For each SamplePSU with a PSUchance equal to 1, assign a unique Stratum value, starting with
1
the number 1.
For the remaining SamplePSU values, assign a unique Stratum value, continuing with the
numbering in step 1 above, to each pair of SamplePSU values.
2 If there is one SamplePSU value remaining at the end of the list that cannot be paired, assign it
the same Stratum value as the last pair of SamplePSU values.

OPTION B
STEP Action
1 Assign a unique Stratum number to each SamplePSU value, starting with the number 1.

FINALLY, ASSIGN THE PSU:


OPTION A
STEP Action
For each SamplePSU with a unique Stratum value, assign each ClusterN value within the
1
SamplePSU a unique PSU value, starting with the number 1.
For each SamplePSU that shares a Stratum value with another SamplePSU, assign all ClusterN
2
values within each SamplePSU the same PSU value.

OPTION B
STEP Action
1 Assign a unique PSU number to each ClusterN value, starting with the number 1.

EXAMPLES:
OPTION A
ClusterN SamplePSU PSUchance IndProbSel Stratum PSU
5-1 5 1.000000 0.120000 1 1
5-2 5 1.000000 0.160000 1 2
3-1 3 1.000000 0.100000 2 3
3-2 3 1.000000 0.180000 2 4
3-8 3 1.000000 0.160000 2 5
6-1 6 0.118792 0.004752 3 6
2-1 2 0.253691 0.060886 4 7
2-2 2 0.253691 0.031711 4 7
2-6 2 0.253691 0.018266 4 7
4-2 4 0.206376 0.075583 4 8
4-3 4 0.206376 0.046228 4 8
1-1 1 0.247651 0.081130 4 9
1-2 1 0.247651 0.039624 4 9

OPTION B
ClusterN SamplePSU PSUchance IndProbSel Stratum PSU
5-1 5 1.000000 0.120000 1 1
5-2 5 1.000000 0.160000 1 2
3-1 3 1.000000 0.100000 2 3
3-2 3 1.000000 0.180000 2 4
3-8 3 1.000000 0.160000 2 5
6-1 6 0.118792 0.004752 3 6
2-1 2 0.253691 0.060886 4 7
2-2 2 0.253691 0.031711 4 8
2-6 2 0.253691 0.018266 4 9
4-2 4 0.206376 0.075583 5 10
4-3 4 0.206376 0.046228 5 11
1-1 1 0.247651 0.081130 6 12
1-2 1 0.247651 0.039624 6 13
ClusterN SamplePSU PSUchance IndProbSel Stratum PSU
La ceiba 1 0.505878463 0.505878463
choluteca 2 0.307333382 0.307333382
danli 3 0.107245205 0.107245205
san pedro sula 4 1 1
santa rosa de 5 0.082041296 0.082041296
tegucigalpa 6 1 1
Instructions for assigning Stratum and PSU values for analysis:
(examples follow instructions below)

FIRST, SORT YOUR DATA BY PSUchance:


Step Action
1 Scroll down until you find the end of your data.
Click the mouse in the left-most cell of the last line of your data and drag across to the
2 right-most cell of your data. Continue holding down the mouse button and drag the mouse
back up until the cursor is over D1.
Release the mouse button. You should now see a big purple/blue box that contains all of
3
your data spanning columns A through D.
4 From the Menu select "Data", "Sort".
Select PSUchance in the "Sort by" drop-down list and select the option "Descending" to the
right of this drop-down list. Next select "SamplePSU" from the "Then by" drop-down list
5 and select the option "Descending" to the right of this drop-down list. Finally, select
"Header row" at the bottom of the window then click "OK".

***IMPORTANT:
How the Stratum and PSU values are assigned is dependent upon the ordering of the sampling units entered
into the PSU spreadsheet. If the list of units was ordered either by size, location, or some other
characteristic that may make the sampling units more similar to units placed closer on the list, then follow
Option A below. If there was no order to the list of sampling units entered into the PSU spreadsheet, then
follow Option B below. Option B is a more conservative option and should be followed if there is any
doubt about the ordering of the sampling units in the PSU spreadsheet.

NEXT, ASSIGN THE STRATUM:


OPTION A
STEP Action
For each SamplePSU with a PSUchance equal to 1, assign a unique Stratum value, starting
1
with the number 1.
For the remaining SamplePSU values, assign a unique Stratum value, continuing with the
numbering in step 1 above, to each pair of SamplePSU values.
2 If there is one SamplePSU value remaining at the end of the list that cannot be paired,
assign it the same Stratum value as the last pair of SamplePSU values.

OPTION B
STEP Action
1 Assign a unique Stratum number to each SamplePSU value, starting with the number 1.

FINALLY, ASSIGN THE PSU:


OPTION A
STEP Action
For each SamplePSU with a unique Stratum value, assign each ClusterN value within the
1
SamplePSU a unique PSU value, starting with the number 1.
For each SamplePSU that shares a Stratum value with another SamplePSU, assign all
2
ClusterN values within each SamplePSU the same PSU value.

OPTION B
STEP Action
1 Assign a unique PSU number to each ClusterN value, starting with the number 1.

EXAMPLES:
OPTION A
ClusterN SamplePSU PSUchanc IndProbSel Stratum PSU
5-1 5 1.000000 0.120000 1 1
5-2 5 1.000000 0.160000 1 2
3-1 3 1.000000 0.100000 2 3
3-2 3 1.000000 0.180000 2 4
3-8 3 1.000000 0.160000 2 5
6-1 6 0.118792 0.004752 3 6
2-1 2 0.253691 0.060886 4 7
2-2 2 0.253691 0.031711 4 7
2-6 2 0.253691 0.018266 4 7
4-2 4 0.206376 0.075583 4 8
4-3 4 0.206376 0.046228 4 8
1-1 1 0.247651 0.081130 4 9
1-2 1 0.247651 0.039624 4 9

OPTION B
ClusterN SamplePSU PSUchanc IndProbSel Stratum PSU
5-1 5 1.000000 0.120000 1 1
5-2 5 1.000000 0.160000 1 2
3-1 3 1.000000 0.100000 2 3
3-2 3 1.000000 0.180000 2 4
3-8 3 1.000000 0.160000 2 5
6-1 6 0.118792 0.004752 3 6
2-1 2 0.253691 0.060886 4 7
2-2 2 0.253691 0.031711 4 8
2-6 2 0.253691 0.018266 4 9
4-2 4 0.206376 0.075583 5 10
4-3 4 0.206376 0.046228 5 11
1-1 1 0.247651 0.081130 6 12
1-2 1 0.247651 0.039624 6 13

You might also like