Help getting p-values of individual beta coefficients as output from mvregress

12 views (last 30 days)
Prabhjot Dhami
Prabhjot Dhami on 4 Dec 2020
Answered: Shraddha Jain on 22 Dec 2020
I am trying to get the individual p-values of my beta coefficients which are output from mvregress.
This is my current code:
[beta, Sigma, E, CovV, logL] = mvregress(X,Y)
Errors = sqrt(diag(covB));
t_values = beta ./ reshape(Errors, size(beta,1), size(beta,2));
pVal = 1 - tcdf(t_values, length(X) -1);
My concerns are:
  1. Have I calculated the t-values correctly
  2. I have a negative t-value of -3.2337, which should lead to a p-value of less than 0.05. However, the resulting value I get is 0.9984. Do I have to include the absolute values of my t-values as input?
Thank you.

Answers (1)

Shraddha Jain
Shraddha Jain on 22 Dec 2020
Hi Prabhjot,
As per my understanding, t-value is negative when the regression coefficient is negative. A possible workaround for calculating t-values and p-values from multivariate regression coefficients is provided in a similar question. Refer to this link.
Hope this helps!

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!