Preparing the Sample

1
Preparing the Sample
2
What is a sample?

• The target population, defined by each site, is
  the population to which the results of the survey
  should be applicable.
• A sample is a carefully selected subset of the
  target population.

3
Why select a sample of the target population?

• It is not logistically possible to survey every
  individual within your target population.
  Therefore, a sample of this population is
  selected for the survey. The results of this
  sample are then "applied" to the target
  population.

4
Determining the Sample Size

• STEPS recommends a minimum sample of at least
  2,000 individuals for sample of 2564-year-olds
  stratified by age and sex.
• The sample size should be calculated specifically
  for your survey.

5
Identifying the Sample Frame

• A sampling frame is a list of units or elements
  that defines the target population.
• A sampling frame should cover all of the target
  population.
• Your sampling frame should provide the highest
  possible level of detail.

6
Stages of Identifying the Sampling Frame
Site

• A sampling frame is most often a combination of
  many different lists that contain different
  levels of information.
• Create a flowchart that details the different
  levels/clusters/groups that describe your site.

Province
District
Town
Village
Household
Individual
7
Stages of Identifying the Sampling Frame
Site

• Select the lowest level in the chart that
  contains information on every individual in the
  top level. This determines the sampling design
  selected.
• (e.g. Does the town level contain information on
  every individual in the site?)

Province
District
Town
Village
Household
Individual
8
Choosing a Sample Design

• The sample design describes how the sample is
  selected.
• STEPS recommends and supports five different
  sample designs.
• The type of sample design selected is determined
  by the sampling frame available.

9
Sample Design 1

• Information is available for all the individuals
  in the site.
• Sampling frame consists of a complete list of
  every individual in the site.
• Ideal scenario, but extremely rare.

Site
Province
District
Town
Village
Household
Individual
10
Sample Design 2

• Information is available for all the households
  in the site.
• Do not have information available on individuals
  within the households, but do have information on
  the households.

Site
Province
District
Town
Village
Household
Individual
11
Sample Design 3

• Information is not available for individuals or
  households at the site level.
• You need to select part of the sample before you
  can get more detailed information.
• Information is available at for individuals once
  you get to a lower level in the sampling frame
  (e.g. Town or Village)

Site
Province
Select a subset 1st part of sample
District
Town
Village
Household
Individual
12
Sample Design 4

• Information is not available for individuals or
  households at the site level.
• You need to select part of the sample before you
  can get more detailed information.
• Information is available for households once you
  get to a lower level in the sampling frame (e.g.
  Town or Village).

Site
Province
Select a subset 1st part of sample
District
Town
Village
Household
Individual
13
Sample Design 5
Site

• Information is not available for individuals or
  households at the site level.
• You need to select part of the sample before you
  can get more detailed information.
• There is no information available on the
  households or the individuals at the lower levels
  in the sampling frame (e.g. Town or Village),
• Perform a census of the village and select
  households or select every household,

Province
Select a subset 1st part of sample
District
Town
Village
Household
Individual
14
Probability Proportional to Size Sampling (PPS)

• Probability proportional to size sampling is used
  for every level of sampling except for household
  and individual selection.
• The larger clusters/groups have a greater chance
  of inclusion in the sample.

15
Basic Sampling Terminology
Target Population
PSU

• PSU Primary Sampling Unit
• SSU Secondary Sampling Unit
• TSU Tertiary Sampling Unit

SSU
TSU
Participant
Participant
Participant
Participant
Participant
Participant
Participant
Participant
Participant
16
Selecting the Sample

• Once the sample design is selected, you are ready
  to proceed with sample selection.
• There is an Excel workbook entitled STEPS
  Sampling that includes spreadsheets for every
  stage of the sample selection STEPSsampling.xls
  will do the following
• Randomize your sample selection.
• Provide probability proportional size.
• Provide information for the weighting.

17
STEPSsampling.xls
18
STEPSsampling.xls
19
Using STEPSsampling.xls

• Helps automate the selection process and provides
  weights for the dataset.
• Plug information directly into the green sections
  of the spreadsheet and you will draw the first
  level of your sample.
• Draw the subsequent levels by putting in
  information from the selected units.

20
Documenting the Sample

• The documentation is used during
• data analysis and
• writing the site report.
• The documentation needs to include information on
• sampling frame,
• sample selection method at each sampling stage,
  and
• probability of selection at each sampling stage.
• Without this information, you will not be able to
  weight your data!

21
Data Collection Forms

• There are two data collection forms used by
  STEPS
• Interview Tracking Form
• Kish Household Coversheet
• You always need to use the Interview Tracking
  Form.
• The sample design selected determines if you need
  the Kish Household Coversheet.

22
Common Questions

• If the interviewer knocks on a selected house and
  no one responds, but there is someone next door
  who wants to take part, can the interviewer ask
  them the questions?

23
Common Questions

• If the interviewer knocks on a selected house and
  no one responds, but there is someone next door
  who wants to take part, can the interviewer ask
  them the questions?
• NO, if the interviewer includes the neighbour in
  the sample then the sample is no longer random
  and it can not be weighted.
• If the interviewers deviate from the selected
  sample, the sample will no longer be
  representative of the target population and will
  not be able to be weighted during data analysis.

24
Common Questions (contd)

• If the interviewer knocks on the door during
  their second attempt at an interview and no one
  is home, does the interviewer need to record this
  information on the Interview Tracking Form?

25
Common Questions (contd)

• If the interviewer knocks on the door during
  their second attempt at an interview and no one
  is home, does the interviewer need to record this
  information on the Interview Tracking Form?
• YES! This information is important. It is used
  to calculate the household response rate. The
  household response rate provides valuable
  information during data analysis.

26
Common Questions (contd)

• If a potential participant refuses to
  participate, do you need to record this
  information on the Interview Tracking Form?

27
Common Questions (contd)

• If a potential participant refuses to
  participate, do you need to record this
  information on the Interview Tracking Form?
• YES! This information is used to calculate the
  response rate and it is also used to weight the
  data.

28
Steps in Selecting a Sample

• Create a chart that describes the structure of
  your site.
• Determine what information is available for each
  level of the structure (sampling frame).
• Select the sampling scenario that corresponds to
  the available sampling frame.
• Select the sample.
• Document the sample.
• Prepare the data collection forms.

29
Using the Generic STEPS Tools

• Every STEPS tool builds onto the next STEPS tool.
  For example, the analysis tools use information
  gathered and formatted into the data entry tool
  and into STEPSsampling.xls.
• You do not have to use the generic tools.
• If you do not use the generic tools, however, you
  will need to use the Mapped Instrument Tool if
  you wish to use the generic data analysis tools.