Data Analysis question?

[danzgymn86]

---

Members do not see this ad.

So, I have 2 levels of my IV and then 2 levels of a moderator variable.
Then I have a lot of DVs...
I have equal numbers in the 2 levels of the IV but vastly different numbers in my moderator (63 to 79, I believe).
What kind of analysis should I be running on this because there are different Ns?
I did a multivariate analysis but I wasn't sure that's right?

 

[deleted176373]

So, I have 2 levels of my IV and then 2 levels of a moderator variable.
Then I have a lot of DVs...
I have equal numbers in the 2 levels of the IV but vastly different numbers in my moderator (63 to 79, I believe).
What kind of analysis should I be running on this because there are different Ns?
I did a multivariate analysis but I wasn't sure that's right?

You are looking to run a two way MANOVA, which under SPSS is a multivariate analysis under the GLM, if I remember correctly. When you say you have different numbers in your moderators, are you referring to different numbers of subjects, I believe that a two way MANOVA should be robust in this example (this I would have to double check to be 100% sure.)
.

Mark

 

[JockNerd]

First off, just because everyone mixes this up and it's important....

Are you 100% sure you mean *moderator* and not *mediator*?

 

[thewesternsky]

JockNerd, you took the words right out of my mouth. I'm T.Aing a basic research methods class right now, and I just had to give a third of the class "0" on a question because they described moderators where they were supposed to describe mediators.

Danz, if you have a moderator, Mark is right. If you have a mediator, though, you'll need to do a slightly different analysis.

 

[JockNerd]

I realized I threw out a question without giving resources.

This web site and the linked sites can help define mod/med
http://www.victoria.ac.nz/psyc/staff/paul-jose/files/helpcentre/help5_moderation_example.php

And Kenny's web site about mediation, if that's what you have.
http://davidakenny.net/cm/mediate.htm

 

[deleted176373]

First off, just because everyone mixes this up and it's important....

Are you 100% sure you mean *moderator* and not *mediator*?

That is an interesting issue. I had not considered that, nice catch JN and TWS. Glad I got it right, lol, I have had 2 stats classes so far, one undergraduate and one graduate one... the graduate one was taught in a very Lewinian/Socratic way, and I somehow learned something, but it's ambiguous what I learned from Singer in that class.

Mark

PS - The lineage of our department is rank with the influences of Lewin, Festinger, Schachter, and Singer. It makes for an interesting heritage and comes complete with its own dogmatic baggage. 😉 Not a criticism, but an observation!

 

[danzgymn86]

Yes, 100% sure it's a moderator 🙂