[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: I have a t-statistic and am looking for the corresponding p-value

From   Michael Barnes <>
Subject   st: I have a t-statistic and am looking for the corresponding p-value
Date   Wed, 29 Aug 2007 17:59:43 -0700

I have a t-statistic, and the number of observations, is it possible to have 
stata find 1-a/2 (p-value) for me?  Just in case that is unclear let me try 
to clarify...I have an expected value for income Yhat, and I want to know 
the probability that income is below the poverty level PL (a constant).  In 
order to find this I am going to form a confidence interval with PL being my 
lower confidence interval limit.  In other words, Yhat-[t(1-a/2; 
n-2)*s{Yhat}]=lower confidence interval limit, or
Yhat-[t(1-a/2; n-2)*s{Yhat}]=PL.  I know Yhat, s{Yhat}, n-2, and PL, I can 
therefore solve for
t(1-a/2; n-2):  t(1-a/2; n-2)=(Yhat-PL)/s{Yhat}.  Where 1-a/2 is my p-value, 
or probability of being past my lower confidence interval limit (in this 
case the poverty level).  Assume n=20 and I solved for the t-statistic and 
get 2.1...I then have 2.1=t(1-a/2; 18).  Is there any way of telling Stata 
that I want the p-value given a t-statistic of 2.1, and n-2=18?  In this 
case, the p-value would be: 0.025, or 2.5%.

I would appreciate any insights!

Messenger Café — open for fun 24/7. Hot games, cool activities served daily. 
Visit now.

*   For searches and help try:

© Copyright 1996–2015 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index