Need help determining what test to use

#1
Hello,

I would like to ask for help determining what test to use. In my assignment I need to determine if different candy causes cavities in different parts of the mouth.

So I have a variable for candies (1 = candy X, 2 = candy Y, ... 13 = candy Z) and 32 variables for every tooth a human has, where 1 = has a cavity, 2 = does not have a cavity.

Am I right to use Chi-Square test in this instance? Thank you in advance for help.
 

hlsmith

Omega Contributor
#2
Pedagogically you could answer this question with a basic proposed test. However, caries on one tooth could effect another. The observations are not independent, so most basic approaches would be inappropriate.


If Candy#1 causes bacteria to grow around tooth 24 and subsequent decay, well you could also get a cary on tooth 23 because of this (e.g., contact cary). Correct?
 
#3
It is probably true, that the tooth can get decay from a nearby tooth, but this is just an assignment used to teach statistics (I am currently studying). So basically it could be anything instead of candy and teeth, for example vegetables and cancer in organs. I should have been more clear on this part. So if we ignore the tooth and candy part, should I be using chi square?
 
Last edited: