reporting chi-square in 2x3 contingency table

css

New Member
#1
Hi everyone,
I am conducting an experiment and I have two groups of participants (controls/ treated; 50 each). For each participant, I performed 40 measures with 3 possible outcomes (A, B, C) and recorded the accumulated frequencies (for example: control1: 10A, 20B, 10C; Control 2: 20A, 10B, 10C...). Now, I want to analyze whether the group is associated with these frequencies and I think the proper way to do so would be a contingency table (2x3) followed by chi-square test. I did so, but now I want to report this result and one doubt arises: It is obvious that the degrees of freedom =2 but I am unsure about which is the value of N.... Is it 100 (the number of participants) or is it 4000 (100 participants x 40 measurements)? The second option seems more reasonable to me but I would appreciate a confirmation or any necessary correction

Thanks in advance.
 

Karabiner

TS Contributor
#2
Your dependent variable is interval scaled (each participant has 3 interval scaled measures), therefore a test for categorical dependent variables (like Chi square) is inadequate.

With kind regards

Karabiner
 

css

New Member
#3
Hi Karabiner,
thanks a lot for your reply. I am happy to be corrected when I still have the opportunity to do things properly. However, if I may, I would like to ask for a little bit more help here. I assume that I should just calculate means of A, B, C and then compare them by a MANOVA... would that be the most correct approach?.

I did choose chi-square because on a previous occasion I had what I considered a similar situation. In that occasion, I analyzed data from an exam involving 30 arithmetic calculations and recorded the frequency of correct responses, errors, and omissions. When I wanted to establish possible differences between genders, I just compared the total number of correct responses, errors, and omissions with a contingency table... I guess I did wrong also in this other occasion, didn't I?

Thanks in advance for your help
 
Last edited: