Questionnaire - Guidance needed.


I am currently refreshing my knowledge about statistics and I would appreciate if someone could provide me some guidance on how to look at my current data set.

I distributed a questionnaire asking participants:

- whether they are working for a governmental institution, a corporation or a scientific institution
- how many years of experience they have in a particular field
- in which country they are currently located
- and other questions that go in a similar direction such as educational background

However, the more important part of the questionnaire was the part where they had to rate a number of different factors from 1=not important to 5=very important.

Instead of just looking how each factors was rated overall and by a particular group, I also wanted to see if there are some correlations. For example, does the fact that they are working for the government affects the way they are rating the presented factors. Another example would be, does the number of years of experience correlate with how the factors have been rated.

I would be grateful for some basic ideas, I am using SPSS.



Active Member
The relevant analysis sub-menus in SPSS:

Dimension Reduction -----> Factor
Scale -----> Reliability Analysis
Correlate ------> Bivariate
Regression -----> Linear

I assume that you do not have many participants.
Hi Staassis!

Thank you for your quick reply. So far, I have 170 participants who filled out the questionnaire.

I already checked the mean of each factor and also split the file to compare the groups. I now would like to see if there are correlations between groups (i.e. scientists) and the how they rated the presented factors. In that case, I would compare nominal data with ordinal data, but I am not sure how, are there any other options other than chi square?
The factors are the result of a systematic literature review and within the questionnaire, participants are given the option to rate them from not important to very important.


Active Member
Didn't quite understand. Presumably, the participants answered many questions and those should be aggregated into factors using Principal Components Analysis or other flavors of Factor Analysis. Also, it is worth checking whether the responses in each subscale are consistent with each other. This is done using Reliability Analysis.

The factors you defined quite subjectively may not be the best way to summarize the data. A cleaner approach is forming factors via an objective, scientific, systematic procedure. That is why I said:

Dimension Reduction -----> Factor
Scale -----> Reliability Analysis
We conducted a systematic literature review to identify factors that might be important in the context of a certain problem. Therefore, we want to identify the most important factors as well as if the perception of the factors is different among different groups.

Going back to my previously mentioned example, I attached two screenshots which might help to understand what I mean. Thanks again for your grea support!
Last edited:


Active Member
Christian, I do not quite understand why you displayed those tables. If there is miscommunication, then please spend your time on describing the experiment carefully. However, if there are many questions in the survey and they can be split into subcategories, everything I have written before is relevant.

Wikipedia pages can give you more information on the types of factor analysis and reliability analysis.


TS Contributor
You could search for “statistical tests decision tree“, using an internet search engine.

As far as I can see, you want to know whether certain characteristics of your participants are related to their responses on certain ordinal items (which are meant to measure certain “factors“). From your examples, you will at least find the Spearman correlation useful then (e.g. years of experience vs. factor response), and the Kruskal-Wallis H-test (e.g. country of origin vs. factor response). Mind that you won‘t be able to use any test which requires an interval scaled response variable.

With kind regards

Thank you very much Karabiner, this already gives me a better idea of how I can look at my data! Hopefully, you do not mind if I have some follow up questions later on.

Best regards