Hey I am doing a repeated measure ANOVA with three different Groups: N1 = 16 , N2 = 17 and N3 = 21.

But when I run the ANOVA it reduces all sample sizes to the lowest n (16). So there are some values from N2 and N3 that are missing and are not count in.

How can I change that in SPSS? Thanks!

Do you mean three groups, or three time points or types of measurements? It sounds like the latter.

There is no way to get the GLM procedure to work with partial data on a case. Any case missing on any variable named in the analysis is deleted (known as listwise deletion). The general linear model approach to repeated measures in that procedure does not provide a solution for incomplete data.

A mixed model approach is available in the MIXED procedure (Analyze>Mixed Models>Linear). This requires restructuring the data so that each observation or time point for each subject is a different case. The Restructure Data Wizard under Data>Restructure in the menus will transform the data as required, using the first option, Restructure selected variables into cases. The Case Studies in the product documentation cover this kind of transformation, as well as running a basic repeated measures model using a mixed model approach (see https://www.ibm.com/docs/en/spss-statistics/27.0.0?topic=models-using-linear-mixed-analyze-repeated-measurements).