6

I am having an issue that I can't find an answer to. In my research, I have a mixed qualitative/quantitative survey (roughly 30 participants).

One of the participants answered 3 out of 5 qualitative questions. Of those three answers, the first is very unclear (I can't understand what the person meant), the second does not answer the question but instead addresses another issue, and the third is ambiguous. My issue is that this third answer could easily be understood as an answer to another question, but would be poorly worded as an answer to the question it was put in.

I am debating whether to consider or to exclude this answer, since based on the participant's other answers I am not sure if the answer is to be interpreted at all. However, the answer also contradicts my main hypothesis. I am therefore worried that excluding it might project the appearance that I am arbitrarily excluding data that does not fit my hypothesis.

How can I ethically treat this answer?

Glen_b
  • 257,508
  • 32
  • 553
  • 939
MikaelF
  • 163
  • 4
  • 4
    *If* you exclude it, do it transparently, explain why, and discuss how your findings change as a result. – henning Jul 01 '18 at 07:02

1 Answers1

4

The way you describe it, this participant gave not one, but multiple problematic answers. In addition, the one answer that is your focus here also needs interpretation as to which way it leans.

It is quite common to run survey questions past multiple independent scorers are exclude those that are judged to be unclear by, e.g., a majority. If you can do this (potentially not only for this one participant, but also for the others), this would be a very defensible way forward.

Alternatively, you could report your results with and without the problematic participant. It sounds like you should treat the situation without him as the "default", as in

Results indicated that foo. If the problematic participant X's answers were included, they changed to bar. (See above on a description of the problems with X's answers.)

Stephan Kolassa
  • 95,027
  • 13
  • 197
  • 357
  • 1
    Yes, presenting both outcomes is the only correct thing to do. As a reviewer I always get suspicios when something is excluded, even if it is explained why. – koalo Jul 20 '18 at 07:44