I have a survey for which I ask participants to self-report their motivation through a series of questions. Now the problem is that I have selected 1500 participants for which I have the contact information and I am asking them to participate in the survey.

Not all of them agreed to participate, which was expected. The problem is that responding to the email itself and completing the survey is most likely done by highly motivated individuals. This would skew my results in favor of the more motivated sample when compared to the less motivated sample.

Is there a way to control for this or a way to reduce this noise?

  • $\begingroup$ You could pay people to take the survey? If the collected data doesn't have the correct information, I don't think any algorithm will help you. $\endgroup$
    – kilojoules
    Commented Jun 19, 2018 at 22:47


Browse other questions tagged or ask your own question.