URGENT!!! Please HELP!!

#1
Thanks in advance for any help you can offer.. I am in need of some advice as to how to go about some results... So firstly i need some Stats advice before going on the the actual SPSS advice! Please see details below:

Basically.... I need to compare the difference between people who view a website and see if their knowlege improves before and after viewing the site. This is the first part details are:

1) Knowledge tests
a) 14 questions true or false
b) measured on the same sample group (i.e same participant answers both questionnaires)
c) measured before and after
d) and this is done on biology and non-biology students (22 each)

The next bit are agree and disagree questions asking about the sites design and usability features to then modify it, the details for this part are:

2) Design and usability tests
a) 26 questions after looking at the site for the first time
b) 13 questions after looking at the site after modifications
c) measured on the same sample group (i.e same participant answers both questionnaires)
d) and this is done on a mixed group of 20 students (i.e. irrelevant what they study)

So from what i gather first step is to test if data is normal and i think the test to use is Shapiro-Wilk and Kolmogorov-Smirnov

The next step is to then do an appropriate test. But i am not sure which is appropriate..... I thought for 1) it would be two sample t-test no?

I know your all real busy, but any help would be highly appreciated, thanks a lot.... If anyone can advise already, please suggest best way to do tests if possible (in spss)? Thanks peepz....
 
#2
Shapiro-Wilk is a test of normality. In SPSS, go into Analyze, Descriptive Statistics, Explore to get normality tests (it's in the "plots" button for some reason). If it comes up not significant, you can assume the data are normal. You also need to check for homogeneity of variance, but you can do that as part of your analysis if you do an ANOVA.

I'm not sure of your design -- are you wanting to see if there are differences between biology students and non-biology students? Or do you just want to see if knowledge improves in general? In either case you can run a t-test, although if it's the former I would suggest an ANCOVA. (Go to Analyze, General Linear Model, Univariate.)

Let me know if you want more details on ANCOVA or the next stage of your test.
 
#3
Thanks, I figured it out I was going to start a thread when i finished to make it easier for other people with the similar problems that may also be confused by the same things... In the end i found that it was pretty simple, just would have been easier if there were more guides about! Thanks anyways :p