  | 
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
st: Confidence intervals differ between Predict, Adjust, and T-test, which to use?
| 
From | 
 
I M <[email protected]> | 
| 
To | 
 
<[email protected]> | 
| 
Subject | 
 
st: Confidence intervals differ between Predict, Adjust, and T-test, which to use? | 
| 
Date | 
 
Tue, 20 Jan 2009 09:10:56 -0800 | 
Hello,
 
After I run a quantile regression estimate, I am unsure of how to summarize the results.  By using the predict, adjust, and t-test command, I arrive at the same mean, but different 95% CI.  Can anyone tell me why the 95% CI are different, and which one to use if I want to summarize the predicted mean?
 
---begin---
sysuse auto, clear
qreg price weight length foreign
adjust, by(foreign) ci
predict adjustedprice
by foreign, sort: ci adjustedprice
ttest adjustedprice, by(foreign)
---end---
adjust command:
Domestic | 5301.89 [4803.46 - 5800.31]
Foreign | 5879.43 [5108.19 - 6650.67]
 
predict command:
Domestic | 5301.886 [4745.151 - 5858.621]
Foreign | 5879.43 [5342.07 - 6416.79]
 
t-test command:
Domestic | 5301.886 [4745.151 - 5858.621]
Foreign | 5879.43 [5342.07 - 6416.79]
 
Note how the calculated mean is the same for each, but the 95% CI is different in each case.  Why are they different, and which one do I use if I want to summarize the predicted mean?
 
Thanks!
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/