I am using a 7-point Likert scale: strongly disagree to strongly agree. However, if the respondent doesn't know the answer, or if it is not applicable, I am hoping that instead of choosing the midpoint score (4) they select ‘N/A’ or ‘Don’t know’. The rationale is to avoid skewing the data, and/or exclude the answer from the data. Is this a valid approach? The downside is that this adds more visual complexity to the survey, which may impact time and/or response rate.
Dear Ezra Conrad Carlson ,
I would recommend separating don't know / not applicable from a 4 (neutral). This is done in the National Student Survey, for example. My advice would then be to leave out the don't know / N/A frequencies in the analysis, thus making the graphs easier to comprehend. I would also recommend a 7 point Likert scale only using words at the ends of the scale, just like yours.
Please check it up:
https://pprg.stanford.edu/wp-content/uploads/Choosing-the-Number-of-Categories-in-Agree-Disagree-Scales-Revilla-M.-Saris-W.-Krosnick-J..pdf
Good Luck.
The way we learned it is ok. As a matter of fact it's important. You can be sure that you will be getting unreliable or unanswered questions if the option of N/A is not available.
As for the visual a 5 anchored scale may be appropriate.
Furthermore, you may want to place the N/A as the last option rather than in the middle...
- Badges
- Science topic
- Similar topics
- Statistics
- Survey