SPSS would remove all missing cases for the entire subject and this reduces the n-value at all the points in the analysis rather than just at the point where the data is missing. Using linear interpolation and means tends to use the mean of the entire data set at a point and not mindful of differences in groups. Help on this.
Here's another option:
Use VARSTOCASES to restructure the data from WIDE to LONG (so that there is one row per observation, and multiple rows per ID).
Then use MIXED to perform the analysis.
The Command Syntax Reference Manual provides the following example of how to perform repeated measures ANOVA via MIXED. You should be able to modify it quite easily to include your between-Ss factor and the interaction term.
Model 9: Repeated Measures Analysis
Suppose that SUBJ is the individual’s identification and Y is the response of an individual observed over several STAGEs. The covariance structure is compound symmetry.
MIXED Y BY STAGE
/FIXED = STAGE
/REPEATED = STAGE | SUBJECT(SUBJ) COVTYPE(CS).
HTH.
Hi,
For your question: Repeated Measures ANOVA analysis is the suitable one.
Kindly see the following link:
http://www.theanalysisfactor.com/repeated-measures-approaches/
http://www.stat.cmu.edu/~hseltman/309/Book/chapter14.pdf
http://talklab.psy.gla.ac.uk/simgen/BarrEtAllAppx.pdf
http://theanalysisinstitute.com/workshops/Repeated-Measures/
Just refreshing on these answers. Bruce Weaver's answer really hit it. I have used this approach and it keeps the values for each subject irrespective of missing data at some time in the longitudinal data collection. Thanks all!!
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA