Data Collection Design Issues

Hey People! I want to share with you. As I work with clients, I am noticing that people are not designing their data collection forms properly for a paired -t test. As a result, I am seeing a lot of cases where the pre and post phases are not properly linked. Another words, there is no way to ensure that id number 1 , for instance, is the same person in both pre and post. Why is this a problem? Well, your id variable ensures that you adhere to the assumptions of your statistics test. Some statistics test require only one record per participant. The paired t test requires that you have one record for each period (Pre vs Post). The independent t test assumes there are no participants in both subgroups (Group 1 versus group 2). Others require multiple pairs of data per participant (repeat measures analysis). In any case, you need to collect a field which assigns a dummy identifier (id variable ).

Here is one solution: assign each participant a dummy identifier when you send the survey link to them. Then have a field in your data collection tool where they can enter the dummy id. This is only necessary for pre and post design projects. The additional field will appear in your file and make it simplier to determine if you have pairs for each subject.

For more information please feel free to contact me, Amy at

-Moore to follow Amy

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.