Moderators: statman, Analyst Techy, andris, Fierce, GerineL, Smash

Posts: 1
Joined: Thu Aug 07, 2014 11:40 am


Postby milka2412 » Thu Aug 07, 2014 12:41 pm


I am a French student in psychology and I am currently using spss to analyze my research project's results.
I work on the perception of dyslexic and non dyslexic secondary school students. I used spss to analyze the differences to a questionnaire (likert scale) between a group of dyslexic pupils and a group of non dyslexic pupils. I did the same to analyze the differences between girls and boys and between four age groups. I think I manage to do and analyze it quite ok.
Now I would like to check the differences with all the independent variables (dys/non dys, sex and age). To do so, I did a multivariate general linear model and obtained the document attached. Being inexperienced in statistics, I don't really know how to interpret these results since it is quite a complex test.
I also don't understand some results. If I take a look at the tests of between subjects effects, I can see that for the "groupeage", there are some significant differences for 6 questions (1, 18, 20, 28, 29, 44).But when I did a general linear model with just the age group as an independent variable (dependent variable being always the answers to the questionnaire questions), I only get significant differences for 3 questions (1, 20,29). I don't understand why there are differences between the two tests and it does the same using "sex" as the independent variable.

To sum up I have two questions :
- how should I interpret the results of the multivariate general linear model (and especially the interactions of the IVs and their effect on the DVs)?
- Why there are some differences in the results between the multivariate general linear model and the other documents centered only on one VI (age or sex)?

I hope my explanation was clear enough and in order to get it clearer, all the documents I talked about can be downloaded on those links I try to attach them to this message but I wouldn’t work) :

I really thank you for all the help you could provide me

Sincerely yours

Posts: 1477
Joined: Tue Jun 10, 2008 4:50 pm


Postby GerineL » Mon Aug 11, 2014 2:16 pm

In short, the interactions indicate that there are differences in one variable, but only for certain levels of the other variable.

For instance, there are differences in age, but only for females.

If you test the effects one of the time (age and sex separately), and there is in fact an interaction, your results may get influenced by that (for instance, it seems like there is no relation, whereas in fact there is a relation but it is opposite for males and females).

I did not look at all your results. I did notice however that you have an incredible amount of variables. I wonder whether you may be able to do some things differently. For instace, could you use age as a continuous variable? Is it useful to test difference between 2000, 2001, 2002 and 2003? Are these really separate factors?
Maybe something to think about, as you may also have a power problem...

Who is online

Users browsing this forum: No registered users and 3 guests