I don't understand the results of my ANOVA

aelw

New Member
#1
I have run general linear model (ANOVA) in SPSS to test whether habitat type (2 categories) and habitat disturbance (2 categories) have an effect on the size of deer groups. When I use the data for ALL groups (including mixed sex groups, all male groups and all female groups - they're all under the one category of group size with no consideration as to gender), then the effect of habitat disturbance is significant (group sizes are larger in the habitat disturbed by humans). However, if I split up the group size category into male groups, female groups and mixed sex groups, and run three separate ANOVAS to see if female groups are larger in the disturbed or undisturbed habitat, and if male and mixed groups are larger in either habitat, there is no longer a significant difference in group sizes between the habitats for any of them (male, female, mixed). I have no idea how to explain this, except that I decreased the sample size by approximately 2 thirds, and the differences between the sizes of e.g. female groups in the disturbed and undisturbed habitats are so subtle that they were not detected. But I'm not sure this is a good enough reason..

Any help is VERY much appreciated!

I should add that my sample sizes are: 16 mixed groups, 13 male groups and 22 female groups
 

Miner

TS Contributor
#2
There could be a number of reasons, but the most likely is the reduction in Power by the reduction in sample size. If you want to see whether there is a difference by group composition, add it in as a third factor.