# Spearman's rho correlation for 4 groups

#### MagnusMokster

##### New Member
I am trying to correlate responses from a questionnaire with performance data. The respondents are divided into 4 groups. I planned to use Spearman's rho correlation, but all the tables I have seen do not list critical values for N less than 5. Is Spearman's rho still a valid method for correlation between 4 groups, or is there another method I should rather use?

Kind regards
Magnus

#### katxt

##### Active Member
Correlation is between two sets of N numbers. As you have written it you plan to use 4 pairs of numbers in your correlation. Are these the means of the four groups? It really isn't clear. If that is true, are you sure correlation is what you want? Can you give more detail about the data?

#### MagnusMokster

##### New Member
Correlation is between two sets of N numbers. As you have written it you plan to use 4 pairs of numbers in your correlation. Are these the means of the four groups? It really isn't clear. If that is true, are you sure correlation is what you want? Can you give more detail about the data?
Yes, it is the means of the groups. I want to explore if there is a relationship between the responses from my questionnaire (eg mean score from question 1 for each of my 4 groups) and the performance indicator (eg score on a test) for each group.

#### katxt

##### Active Member
What is the significance of the four groups? Do you want to see if they are different?
Does each subject have a question score and a performance score?

#### MagnusMokster

##### New Member
What is the significance of the four groups? Do you want to see if they are different?
Does each subject have a question score and a performance score?
The groups are four different types of profession. I have a total of 208 subjects, and between 30 and 80 subjects in each group. I have question scores for each subject, but only a mean performance score for each group. I want to see if high question score corresponds to high performance score and vice versa

#### katxt

##### Active Member
Be careful if you find the correlation for more than one questions. There is a danger of false positives if you have multiple p values. Look up Bonferroni. kat