# p-value of the null hypothesis that the regression coefficient is zero

• Feb 27th 2013, 07:47 PM
kingsolomonsgrave
p-value of the null hypothesis that the regression coefficient is zero
I am asked to do a single and multiple regression on various variables affecting wages using STATA and a data set. Afterwards I'm asked to:

record the coefficient estimates and the p-value for the 2-sided test.

The coefficient estimates are easy because stata gives them to you, but the p-test part...I have no idea how to do this.

I do know what a p-test is, I just don't know how I would set it up. From searching online I found that the hypothesis is (maybe) that a regression coefficient is zero?

Assuming that is what the test is for how do I set up the equation?

(0 - regression coefficient)/(sd/sqrt(n))

If I set it up as above, which sd do I use? would it be the sd of the dependent variable?

thanks for any help!
• Feb 28th 2013, 05:00 PM
chiro
Re: p-value of the null hypothesis that the regression coefficient is zero
Hey kingsolomonsgrave.

You need to get the standard error of the parameter estimator to do this.

Once you do this you then need to know the underlying distribution for the test-statistic: Can you tell me what this is (Hint: Consider a t-distribution)?