How to start up a good survey?
Step 1: Identification of the target group
- Decide whether the impact survey will be about all your beneficiaries or only beneficiaries under a specific programme or in a specific region. This will be your target group for the impact measurement survey.
Step 2: Design of the sampling frame
- Make a list with the names of the target group members;
- Define subgroups that need to be distinguished (specific target groups, regions etc.);
- Define what proportion of the total target group is in each subgroup.
If you have specific target groups, follow the steps below per subgroup.
Step 3: Defining the appropriate sample size
In statistics, a sample is a subset of a population. Determine an appropriate sample size using the following sample size calculator: surveysystem.com . For the World Citizens Panel we will work with a confidence level of 95% and a confidence interval of 5.
Step 4: Drawing the sample
- Divide the total number of the target group by the sample size – as obtained in step 3 -. The result is ‘X’.
- Take the list of your total target group (as made in step 2)
- Determine the number at which you start selecting the sample, using the following link: random.org . The result is ‘Y’
- Y is the first number in your sample. Mark the name of the person corresponding with this number on your target group list. The next number is Y+X. Mark this name again. The next again Y+X+X. etc. Continue listing these numbers and the names on the target group list until you have reached the whole sample size (as determined in step 3) .
- Check if the sample is a fair representation of the total target group. This can be done by comparing the sample list with the total target group list. Compare, for example the percentage of male and female, the educational level, or the income level/poverty classification.
- Now make a reserve list in case certain people cannot be reached during the interviewing period: follow points of step 4 again.
Step 5: Making the sample list
Make a well-structured document containing all the people that need to be interviewed, to be used for identification at a later stage and for comparing data from the same people.
|Participant code||Participant name||Programme code|
The participant code will be a unique number, that will be entered in the automatic system. For example 001, 002, 003 etc
The programme code may be used for identifying in which of the organisation’s programme/project a participant participates.
Step 6: Composing a control group
Do the survey among a group of people that is in principle similar to your target group, but that does not (or not yet) participate in your activities. For instance people who will be included in your target group in the future or people who have not been selected because of your criteria (income/poverty?), but who were very close to selection. It is very important that the control group is as much as possible comparable with the target group except for the fact that they are not participating in your projects.
In principle, follow steps 1-5 for this control group.