SPSS - Using mixed univariate ANOVA (UNIANOVA) or mixed linear regression (MIXED) instead of repeated-measures ANOVA

#1
Which process is better for using instead of Repeated-Measures ANOVA (SPSS syntax: GLM -- GUI command: Analyze --> General Linear Model --> Repeated Measures)? and Why?


(1) Mixed univariate ANOVA (SPSS syntax: "`UNIANOVA`" -- GUI command: Analyze --> General Linear Model --> Univariate)


(2) Mixed linear regression (SPSS syntax: "`MIXED`" -- GUI command: Analyze --> Mixed Models --> Linear)
 
Last edited:
#2
two .... modalities (verbal and visual)
I don't know what is meant with "modalities". I guess that it is just two different measured variables.

To me it seems like the subject is the random effect variable. It is like if you have selected a number of families with 5 brothers and measured the length and weight (like the two modalities). Then the family would be the random effect and the five brothers like the five time repeated measures.

I don't know the spss treat these things.

If I am correct about the two modalities verbal and visual, I would prefer to have them as two columns, and the time variabel in one column. That would be a long format.

First let us se if this makes sense for you.
 

Karabiner

TS Contributor
#3
I am not sure what to think of your strategy to convert into "long" format and perform an univariate
analysis of variance. Where does this account for the correlation of observations within subjects?
Why don't you just leave it in the "wide" format and perform a repeated-measures analysis of variance,
including 2 repeated-measures factors "time" (5 levels) and "format" (2 levels)?

With kind regards

Karabiner
 
Last edited: