What type of test does matlab use to obtain corr pvalue?

According to the documentation, the correlation pvalues indicate "If pval(a,b) is small (less than 0.05), then the correlation rho(a,b) is significantly different from zero."

But it's never specified which test was used to obtain the pvlaues.

Answer 1

[Manikanta Aditya]

Hi @Luis Jesús Olvera Lazcano,

In MATLAB, the function 'corr' can compute the correlation coefficient and its p-value. The p-value indicates the probability of obtaining a correlation as extreme as the observed value under the null hypothesis that the true correlation is zero. The test used to obtain the p-value for the correlation coefficient depends on the type of correlation you are computing.

So, the type of test used to obtain the p-values in MATLAB depends on the function used and the type of correlation (Pearson, Kendall, Spearman, etc.) being computed.

Here's an example of how to compute the Pearson correlation coefficient and its p-value:

% Example data
x = [1, 2, 3, 4, 5];
y = [2, 4, 6, 8, 10];

% Compute Pearson correlation and p-value
[R, P] = corr(x', y');

% Display results
disp(['Pearson correlation coefficient: ', num2str(R)]);

Pearson correlation coefficient: 1

disp(['P-value: ', num2str(P)]);

P-value: 1.4043e-24

MATLAB uses a t-test to compute the p-value for the Pearson correlation coefficient. For Spearman and Kendall correlation coefficients, MATLAB uses rank-based methods and approximations for hypothesis testing.

Refer the following documentation to know more Correlation coefficients:

• Correlation coefficients - MATLAB corrcoef (mathworks.com)

I hope this helps.

[Luis Jesús Olvera Lazcano]

Thank you so much! That's what I was looking for!