More scale points on survey than reporting out

Hi @ClaireG,

First of all, it’s commonly accepted in the literature that non binary scales are better with symmetric choices around a neutral point (“neither satisfied nor insatisfied”).

The “do not apply” should not be counted in the scale as these responses should be excluded from your data analysis when calculating a % of satisfaction.

Results should already be accompanied by the number of respondents to the question (and this number can be different for each questions of your survey because of DNA exclusion)

If you report results in a Qualtrics dashboard, you can exclude choices for numeric questions in the data mapping.

Hoping this helps,

Élie

@ElieD - That does help, thank you. It’s good to have a reference point with the literature on survey design, I’m getting back into this after 4 years away. Part of what I’m mentally stuck on is what level of detail could be lost for the department with only reporting the % happy. I found an earlier report where the department had received a report with all the response options. Some questions had a much bigger very satisfied than others. 

I’ve used Qualtrics advanced reporting before, and I’ve also figured out collapsing response options. Part of the report our office prepares includes an additional column of data, which is a numerical goal point for each item. Then the survey data that’s below get’s highlighted (example less than 80% aren’t satisfied or very satisfied). For this semester to get through populating the reports I’ve prepared SPSS code for collapsing scales. Which includes a system missing for the “did not use” matrix.

@ClaireG Beside Symetric Scale and Excluding “Did not use”. One big question to ask here is: “What’s the purpose of spending money collecting those response?”. Is the department have a KPI for %satisfied and that’s all they want to monitor and they don’t give a thing about others as long as they having high number. Or the purpose is to find improving chance for the course to reduce drop-off rate, then you really have to dig deep.

@ClaireG Thanks for asking this question! And thanks to @dxconnamnguyen and @ElieD for your contributions! 

Another thought - The best practice is to have all 5 standard options (Very Dissatisfied, Dissatisfied, Neutral, Satisfied, Very Satisfied.) It could be good to do a top/box bottom box calculation comparing the satisfied group vs. very satisfied. For satisfaction, I would also recommend using at least one overall satisfaction before branching into the matrix's questions.

Hope this helps!