One Tail Tests

Yida

New Member
#1
Is there anyway to output one tail test pvalues for running OLS regressions?

Also, using ivreg2 and outreg2, I get a row of results that are "r2_a". I assume this is adjusted r-squared?
 

Dason

Ambassador to the humans
#2
What do you mean by a one tail test for OLS regression? A one tail test of individual parameters? Just testing if they're positive or something like that?
 

Yida

New Member
#3
For a two tail test, a pvalue could be say .5
In a one tail test, the pvalue would be .25

For a two tail test with a pvalue of -.8
The one tail pvalue would be .6 (which is 1-.8/2)
 

Dason

Ambassador to the humans
#4
For a two tail test, a pvalue could be say .5
In a one tail test, the pvalue would be .25

For a two tail test with a pvalue of -.8
The one tail pvalue would be .6 (which is 1-.8/2)
Except both of those depend on the direction of the test. In your first example the one-tailed p-value could be .75 if the direction of the effect is the opposite of what was hypothesized.
 

bukharin

RoboStataRaptor
#5
Also, using ivreg2 and outreg2, I get a row of results that are "r2_a". I assume this is adjusted r-squared?
Yes. Assuming you're using the -regress- command, this is documented in -help regress- (see the "Saved results" section at the end of the help file)