Re: SPSS easy question

Posted by Bruce Weaver on Aug 24, 2011; 11:27am
URL: http://spssx-discussion.165.s1.nabble.com/SPSS-easy-question-tp4729732p4730096.html

Yes, one can partition an overall chi-square into orthogonal components.  I have some notes on this here (starting on p. 6):

    http://www.angelfire.com/wv/bwhomedir/notes/categorical.pdf

And I just found some more here (starting on slide 39):

   http://www.ed.uiuc.edu/courses/EdPsy490AT/lectures/2way_chi-ha-online.pdf

Notice that partitioning is neater if you use the likelihood ratio chi-square rather than Pearson's chi-square (i.e., orthogonal components add up exactly to the total).

HTH.

maartjegoorden wrote

Hello!

Probably this is a very easy question:

I have 2 variables (see table). They both consists of three parts, respectively numbered (1,2,3) .
So 1=depression, 2=panic and 3=generalized fear


I did a chi square test (cross tabs) and it is significant. Now I want to know if the the people who are early in life have a significant effect between depression and panic. Is it okay for me to do another chi square test, where I make a new variable with only depression and panic in it?

Thanks in advance!

Maartje