by first checking normality and confirming equality of varince you can use t-test with equal or unequal varince (depending upon the result of F-test for equality of varince). if you use spss then it gives both reslts for t-test (equal and unequal variance and i think F-test also)

Hi.

a very basic, you can find that the correlation between:

- Discrete variables were calculated Spearman correlation coefficient.

- For discrete variable and one categorical but ordinal, Kendall's Tau is advisable

- For discrete variable and one nominal categorical or nominal in both possible contingency table with test Chi-square Independence.

- Others may be Phi coefficient or tetracorico

Thank you for the help! :)

One more question: What test should I use if I want to see whether there is a statistically significant difference in the means of a continuous variable between two groups (e.g. the mean blood pressure in groups A and B)? Thanks!

by first checking normality and confirming equality of varince you can use t-test with equal or unequal varince (depending upon the result of F-test for equality of varince). if you use spss then it gives both reslts for t-test (equal and unequal variance and i think F-test also)

I agree with @Raja Fawad, but if the data is not normally.. apply non parametric test as U Mann-Whitney

Thank you guys! You have been a great help! :)

Hello Mr. Mike,

Whuch is best correlation for discrete data