research analysis confusion. Please help.

Ask here about academic and research issues, like designing studies, recruiting participants, choosing statistics, submitting for publication, etc.
Post Reply
Jalizabeth
Posts: 22
Joined: Thu Nov 29, 2012 9:13 pm

research analysis confusion. Please help.

Post by Jalizabeth » Mon Mar 24, 2014 2:49 pm

Hi all,

I have managed to get myself very confused.
I'm trying to work out what analysis to use on a correlational design with one IV and 3 DV's.
I thought I had it worked out with a MANOVA but I just read that you can only use a MANOVA
If the IV is categorical and mine is continuous (amoubt of time spent).
my DVs are 3 continuous measures of wellbeing.

Going crazy trying to work this out as nothing seems to fit.
I would really appreciate any help or direction anyone could provide.

Many thanks x

User avatar
enid
Posts: 739
Joined: Wed May 30, 2012 5:03 pm

Re: research analysis confusion. PLEASE HELP!

Post by enid » Mon Mar 24, 2014 2:54 pm

how many levels does the iv have? and how many ivs do you have - is it just 1?

Jalizabeth
Posts: 22
Joined: Thu Nov 29, 2012 9:13 pm

Re: research analysis confusion. PLEASE HELP!

Post by Jalizabeth » Mon Mar 24, 2014 3:51 pm

enid wrote:how many levels does the iv have? and how many ivs do you have - is it just 1?
Thanks for your reply, I have one continuous IV (amount of time spent in hours) so just one level.

User avatar
sarahg
Posts: 614
Joined: Sat Dec 10, 2011 1:46 am

Re: research analysis confusion. PLEASE HELP!

Post by sarahg » Mon Mar 24, 2014 4:22 pm

Take a look at Multivariate regression :)

User avatar
Gilly
Moderator
Posts: 1861
Joined: Sun Sep 06, 2009 9:30 pm
Location: Doodling on Paint somewhere...

Re: research analysis confusion. PLEASE HELP!

Post by Gilly » Mon Mar 24, 2014 4:33 pm

Hiya

What is your research question - because that will point you towards the type of statistical analysis.

Are you trying to look at whether or not there are differences between different groups, which would point you more towards T-Tests, ANOVAs etc... or are you looking at how do each of the variables (your wellbeing measures) affect the 'amont of time spent' which would point you more towards correlational and regression tests - I'm imaginging a multivariate regression if you want to see how your 3 tests impact on your IV.

EDIT: Darn you Sarah, beating me to it :D
You're not calling for help, are you?! ;)

"The problem with quotes on the Internet is that it is hard to verify their authenticity" - Abraham Lincoln.

Jalizabeth
Posts: 22
Joined: Thu Nov 29, 2012 9:13 pm

Re: research analysis confusion. PLEASE HELP!

Post by Jalizabeth » Mon Mar 24, 2014 4:52 pm

Thank you Gilly and Sarah, I have not done a multivariate analysis before.
my stats text books do not seem to go into them or they call it a GLM I think?
So I would have never come across it! Yes my research question is looking
At the effect of engaged time use on 3 wellbeing measures. I'm looking to see if time use
Is related to/predicts scores on the wellbeing measures. Effectively I'm going to see that
Increased engaged time use increases scores on the wellbeing measures (severity ofsymptoms, subjective wellbeing and social support). I would also be interested to see if there are any interactions between the DV's too.

Thank you again I really appreciate the replies.

User avatar
miriam
Site Admin
Posts: 7835
Joined: Sat Mar 24, 2007 11:20 pm
Location: Bucks
Contact:

Re: research analysis confusion. PLEASE HELP!

Post by miriam » Mon Mar 24, 2014 5:47 pm

Sounds like I'd start with very straightforward correlations...
Miriam

See my blog at http://clinpsyeye.wordpress.com

User avatar
sarahg
Posts: 614
Joined: Sat Dec 10, 2011 1:46 am

Re: research analysis confusion. PLEASE HELP!

Post by sarahg » Mon Mar 24, 2014 10:33 pm

Gilly wrote:
EDIT: Darn you Sarah, beating me to it :D
Always one step ahead ;)

Put simply - GLM's or, generalized linear model's (Note: This is different from GENERAL linear model),typically refers to a set of statistical analysis whereby the outcome (DV) is not normal distributed. So for example binomial logistic regression - whereby the outcome variable is binary, or poisson regression, whereby the outcome variable is modelled as a discrete variable.

From what I can gather your DV's appear to be normally distributed, thus you would not need to use a GLM.

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest