I have two symmetrical questionnaires which have been structured according to a Latin square logic. There are about ten binary features differently distributed in 70 items. This list of items has to be evaluated by the participant by selecting a range of 4 possible responses. Each questionnaire has been filled by a specific sample of randomly selected participants. Each participant was given just one questionnaire ( so participants are different from 1 and 2).
I would like to do some inferential statistics on the two questionnaires and quantify the relationship between the features structuring the anwers. My results look like spreadsheets produced by Googleform : I have all the scores for each item (1,2,3,4) and I can create subgroups according to the ten features.
Hope to have been clear enough, thanks a lot !
Hi. If you are using a standardised scale, then it's very likely that the authors of the scale would have provided the scoring system. Please reach out to the authors in case it's not been provided. Based on the scoring system, you'll have to total the items in the scale and derive the final score.
In case you have developed the scale yourself, then you need to determine the items to be totalled and how to derive the final score.
Once you have calculated the total score for both scales, you can test the association/correlation between the dependent variable (the scales) and the independent variables in your study.
The next step is to determine which test to use and I have a Youtube video on that- https://www.youtube.com/watch?v=F4AKZ9-u_ic
I've also published two articles in which I have carried out Kruskal-Wallis test (a non-parametric test) that might be useful- Article Online learning experiences of social work students in India
Article Wellbeing of Palliative Care Workers During Covid-19 Pandemi...
Hi Rama, thank you for your reply. at a first glance your answer seems to be useful : I'll check your articles!
Hi, after a meeting with a statistician from my lab I've found that a good answer to my question is to apply ordinal regression and then multiple comparisons on the scores obtained. I will rearrange my database and check if this solution bring interpretable results
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages
- Python