I have a population in which I am looking at two variables in individuals.

The first is the size of their vertebrae, the other is the presence of pathology in the teeth. The first is in mm (ranging from 9 mm to 15 mm) and ranges over 8 vertebrae, the other is binary (present and not present (and missing)) and ranges over 6 teeth.

Now, what I want to know is if there is a relation with small, or big, vertebrae and the presence of pathology in the teeth. Which statistical test/procedure is suitable?


The answer depends on how many observations you have collected and what your hypothesis is. You should specify this for yourself to be certain about the required analysis.

Supposing that you want to get an idea about the relationship between vertebrae size and teeth pathology, and not about an association between specific teeth/vertebrae (eg. in humans 3rd molar upper right (N# 1) to the vertebra Th1) that may be connected.

Then I would consider to aggregate the information that you have observed; e.g. to take the median vertebrae size per individual (mm) and compare it with overall presence of teeth pathology (Yes/No) or number of teeth with pathology (N) per individual. Here is an overview of statistical tests you could use then.

By the way, I would choose for the median because vertebrae size is likely to be a skewed variable. I may be wrong.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy