Analyse a survey?

lotw

New Member
#1
Hi,

It's the first time I want to analyze a survey but I can't get to the right methods/literature... a lot of what I find contradicts itself. I have the following data (also attached an image to illustrate).

This is not the real story but just for illustration: survey.PNG
Its a survey that was done for 300 people who got a certain treatment. Each survey had 15 questions and the 300 persons where spread over 4 different locations. Now the difficult thing I find, is that:
- Sometimes the survey was done 3 times: just before treatment, in the middle, and right at the end. But sometimes this repetition was not done, or just 2 times.
- The count for each location where unequal, so one location could have 50 people doing the survey, and the other location 150...

The 2 type of answers I would like to get are very simple:
1) when someone score high in question 2, is it also high on question 10 (this is an example but I would like to find correlations between the answers of 2 questions)
2) are there differences in answers between locations

My first instinct to answer these questions would be:
1) for question (1) I would think of something like a interclass correlation (that is what I found)
2) for question (2) I would leave out the repetition, and only take surveys, taken on a specific time, like only the survey that where done in the end

So could someone help me? Any help, also literature that could help me is very welcome :)
 

staassis

Active Member
#2
Try Repeated Measures ANOVA. Should have a chance of working as long as each time point is represented with a sufficient number of people.