Best Statistical test for data


New Member
Hi All,

Just wondering if someone could help with an issue with my data.

I am wanting to analyse the relationship between continuous variable (
cholesterol) and BMI categorised in four categories (under, normal, over, obese).

What test would be best to look at this ? Would multi-way ANOVA be usefuL? Sorry my biostats knowledge is quite bad.

Help would be much appreciated. Many thanks.



TS Contributor
A 1-way ANOVA followed by a post-hoc test would be appropriate. If you want to compare under/over/obese against normal (as control), try Dunnett's test.

Another option would be to use ANOM. However, the null hypothesis is subtly different than the usual. In ANOM, the null is that no mean is different from the group mean. In your case ANOM would test each mean (under/normal/over/obese) against the overall mean. You would have to decide whether that hypothesis is of interest to you.