Skip to main content

Question 5: Deciding the design

1

We have decided that the respondents should complete the questionnaire by post anonymously. The advantage of this is, the respondent can reply honestly if they did not like the service, without feeling their medical care will be affected. But there are some problems with this; we might find it hard to work out who responded and who didn't. This means we will not know the characteristics of the non responders. We won't be able to send out reminders. What can we do about this?

a)
b)
c)
d)

We should put a number on the return envelope.

If we ask people to just fill in their age, sex and diagnosis and return it, we will still have non-responders (as some people will not do this).

We can not estimate the characteristics of the non-responders unless we can in some way identify the responders in the list that were sent a questionnaire. We can only identify the responders and non-responders if we have put a number or identifier on the return envelope or questionnaire.

If we do not report on the non-responders we don't know if our results are valid. We do not know if we got responses from a representative sample, or if we just got responses from a very selected and bias sample.

Check your answer