Dear colleagues,
In our research, we used a questionnaire with students' answers on scales. We transfered the responses into numbers and calculated medians. Do you consider it appropriate to sum the values in batteries for each student and then run a statistical test?
Thanks for your opinions and literature, if there is any.
Best regards,
Martin
If it is meaningful for whatever research question you have, it could be fine, but more information is necessary. You talk about batteries, scales, and responses. Are you saying that there are multiple batteries, within each of these there several scales, and the scales are based on several items. Of these items, they are all on the same rating scales, say 1-5, and high values mean the same thing for each, so you simply want to use the median rather than some other measure of location (e.g., mean) to calculate aggregate scores for each individual for the scale? And then sum or mean of these medians for measures for the battery estimate. This would not be most people's first choice, but depending on why you are doing this (i.e., your research question) it might be okay.
Hi Daniel Wright ,
Thank you for your comment.
Your summary is correct :) However, median was chosen as the correct measure as the scales naturally produce ordinal data.
However, thank you for your confirmation ;-)
Best wishes,
Martin
- Badges
- Science topic