My father is a doctor and he asked me to find out if there is a correlation in two sets of data he collected, one concerning the age of his patients and one concerning the value of a medical exam
I computed the Pearson correlation index (the one in the excell function for example) and it was 0.07, so no correlation apparently.
Is this it (so no correlation) or is there another type of correlation I can try?
Two additional popular correlation methods are Kendall's tau and Spearman's rho. Both of these methods consider the ranks of values (age and the medical exam score in your case) instead of the values themselves (as Pearson's r does).
Kendall's tau is computed by considering each pair of observations and returning the proportion of pairs where the ranks are concordant (in your case, the larger age is paired with the larger medical test value) minus the proportion of pairs where the ranks are discordant (in your case, the larger age is paired with the smaller medical test value). Spearman's rho computes the Pearson's correlation of the ranks of the two variables.
Both Kendall's tau and Spearman's rho return value 1 if the ranks of the two variables are exactly the same (aka the largest age is paired with the largest medical test value, the second largest age is paired with the second largest medical test value, and so on). Similarly, both tests return value -1 if the ranks of the two variables are exactly reversed. This is not true of Pearson's r, which only returns -1 or 1 if the two variables have a linear relationship.
A number of questions on this site dig into how to select between correlation methods in different scenarios or how to interpret differences in the values. You might check out, for instance:
