Chi Square (sometimes significant sometimes not)



Hi all,

I was conducting a chi-square + fisher test and ran across a problem:
I am using a chi square to test whether or not there is an association between the different dietary groups (5 groups) and the ferritin categories (low vs normal). This can be found in the attachment AllGroups.
I get no significant pvalue from either the chi square or fishers test.

When I run a chi-square test on only vegans vs nonvegetarians, I get that there is a significant association between the 2 dietary groups and the ferritin groups.

Why doesn't the test containing all dietary groups pick up the significant difference between vegans and nonvegetarians?

Any help in understanding how to report this?

I appreciate any help this community can provide. Thanks!