post-hoc test with interaction

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

Stefan123
Posts: 2
Joined: Thu Feb 07, 2013 8:17 am

post-hoc test with interaction

Postby Stefan123 » Thu Feb 07, 2013 8:31 am

Hello,
I am working on the analysis of biomass growth of plant (dependent variable) for different clones (independent variable) and locations (independent variable). When I do a univariate test it appears that interaction takes places between the two independent variables.
Due to the interaction I make an analysis for the simple main effect.

Now I want to know or the clones differ in biomass within a site. I know that it is possible to do this analysis with splitting the data, but in that case a lot of degrees of freedom get lost. I know that it is possible in SAS when slicing the data, but don´t know how to do this in SPSS.

Is it possible to investigate the difference between clones within a site while using all the data, to keep all the degrees of freedom?

I use SPSS statistics 17.0

Regards, Joost
statman
Administrator
Posts: 2723
Joined: Tue Jun 12, 2007 12:08 pm
Location: Florida, USA

Re: post-hoc test with interaction

Postby statman » Thu Feb 07, 2013 1:28 pm

What is "slicing the data?" This does sound like Split File but ???
See the note below

NOTE: Please read the Posting Guidelines and always tell us your OS, the SPSS version and information about your study and data!

Statman
Statistical Services
Stefan123
Posts: 2
Joined: Thu Feb 07, 2013 8:17 am

Re: post-hoc test with interaction

Postby Stefan123 » Fri Feb 08, 2013 8:45 am

Slicing the data is something that can be used in SAS.
My boss says that in that way you can keep all the degrees of freedom from all the data when looking at the data of the clones at one site.

I also made the analysis with Split File, but than the problem is that you cannot keep all degrees of freedom from all the data, but only from the small part we analyze.

So I want to make the analysis of the clones within one site but keeping the degrees of freedom from all the data.

For me it sounds strange to make the analysis of a part of the data and keeping the degrees of freedom of all the data but my boss say it is a good way of analyzing.

Is this possible in SPSS?

regards,
Joost

Who is online

Users browsing this forum: No registered users and 2 guests

cron