help with statistics

OpenToSuggestions · 10 May 2009 at 19:39

I am doing some work with a program called SPSS in order to draw some conclusions on some data collected on the weekly unit intake of students. This data covers 5 courses and 5 years.
I am trying to find if the level of drinking has changed over the 5 years and if the attendance varies at all between sexes or even between the courses.

I can use the software to perform t-tests, anova's, histograms etc etc but I simply do not know where to start, as with things like Anova's I don't really know what the outputted data means.

Any pointers on which tests could be used to prove what?

Nix · 10 May 2009 at 19:41

Different tests look for different things. Read up on what each test does and is used for first.

SPSS is a royal git of a program though. I had to get a friend who'd used it lots to help me with it in the end.

dangerstat · 10 May 2009 at 19:43

mglover070588 said:
I am doing some work with a program called SPSS in order to draw some conclusions on some data collected on the lecture attendance of students to lectures. This data covers 5 courses and 5 years.
I am trying to find if the level of drinking has changed over the 5 years and if the attendance varies at all between sexes or even between the courses.

I can use the software to perform t-tests, anova's, histograms etc etc but I simply do not know where to start, as with things like Anova's I don't really know what the outputted data means.

Any pointers on which tests could be used to prove what?

wow that's almost "how do I do statistics" and "how do I use SPSS". You'll likely need a lot of help mate, have you no statistics department at your university?

G|mp · 10 May 2009 at 19:45

1 in 7 Dwarves is grumpy.

OpenToSuggestions · 10 May 2009 at 19:46

dangerstat said:
wow that's almost "how do I do statistics" and "how do I use SPSS". You'll likely need a lot of help mate, have you no statistics department at your university?

It is a small stats module in an engineering degree. We got set this work thurs and it is due in mon :eek:

platinum87 · 10 May 2009 at 19:57

i have used SPSS and its pretty easy..

it should really just say everything you are expected in a hand out

do t-test other correlation, mabye factor analysis and than finish off with some descriptive statistics

i would have thought

Dark88 · 10 May 2009 at 20:00

ANOVA woop!

just start putting in some models..
ATTENDANCE = SEX, or DRINKING = SEX (if that what you data columns are called)
if your P values are below 0.05 then then that variable is significant

then you can start building up more complicated models if you want.

then you can just use descriptive stats to pull out which groups have the highest means for whatever you're looking at.

(btw that's a very quick rushed explanation but will get you some results out)

dangerstat · 10 May 2009 at 20:18

mglover070588 said:
It is a small stats module in an engineering degree. We got set this work thurs and it is due in mon

Well it's basically basic stats. Have you got one of those "Engineering mathematics" textbooks with you (you know the really common one - can't remember who it's by though) that has a chapter on basic stats.

OpenToSuggestions · 10 May 2009 at 21:14

I have some text books but stats is like french to me - It just makes no sense!

Right now I am trying to determine whether the level of drinking over the years recorded has changed. Any pointers on that? I am 99% sure I need to do some sort of T test and some sort of ANOVA, but not sure which ones and why.