# Chi square test: non parametric

33 views (last 30 days)
Cside on 6 Oct 2020
Commented: Cside on 9 Oct 2020
Hi, I have 3 groups of categorical data (all logicals) and would like to run the chi square test to get the p-value. However, these 3 groups are of unequal sample sizes. Which chi square function would be recommended for me to use - one that does not require me to key in the degree of freedom and returns me the p-value and chi square value?
Thank you so much!

Jeff Miller on 6 Oct 2020
It sounds like this is what you want:
% Create some fake data to use as an example,
% with 3 groups of different sizes;
score1 = randi(2,44,1) - 1; % 1 = true, 0 = false
score2 = randi(2,66,1) - 1;
score3 = randi(2,77,1) - 1;
% Make some vectors to code the groups
cod1 = 0*score1 + 1;
cod2 = 0*score2 + 2;
cod3 = 0*score3 + 3;
codes = [cod1; cod2; cod3];
scores = [score1; score2; score3];
[t, chisq, p] = crosstab(codes,scores)
##### 1 CommentShowHide None
Cside on 9 Oct 2020
great! this was exactly what I was looking for, thank you Jeff :)