Statalist The Stata Listserver


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

st: comparing coefficients accross models


From   Traci Schlesinger <[email protected]>
To   [email protected]
Subject   st: comparing coefficients accross models
Date   Wed, 10 Jan 2007 21:09:16 -0600

hi stata users!

I have two models (and many other similar pairs). The first examines the effects of changes in sentencing policies on black men's admission's to prison for violent crimes. The second examines the effects of changes in sentencing policies on white men's admission's to prison for violent crimes. Here are the models and their results:

. newey2 bmv_rate mt_b ca il ne nj tx moreharsh lfp perblack adol violent, lag(1) t(mergeyear) force


Regression with Newey-West standard errors Number of obs = 61
maximum lag : 1 F( 10, 50) = 53.18
Prob > F = 0.0000

------------------------------------------------------------------------------
| Newey-West
bmv_rate | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
mt_b | 25.63707 7.820457 3.28 0.002 9.929216 41.34492
ca | -1683.497 553.7605 -3.04 0.004 -2795.758 -571.2367
il | -1180.779 335.2558 -3.52 0.001 -1854.16 -507.3985
nj | -1722.034 499.4906 -3.45 0.001 -2725.29 -718.7775
tx | -1444.673 409.6565 -3.53 0.001 -2267.492 -621.8537
moreharsh | -334.9321 134.9855 -2.48 0.016 -606.0585 -63.80571
lfp | 234.4752 99.71753 2.35 0.023 34.18662 434.7637
perblack | -11910.15 3144.488 -3.79 0.000 -18226.04 -5594.259
adol | -4938.191 1418.736 -3.48 0.001 -7787.806 -2088.577
violent_ar~s| .2775444 .0685935 4.05 0.000 .1397702 .4153186
_cons | 3497.981 896.4067 3.90 0.000 1697.496 5298.467
------------------------------------------------------------------------------

. newey2 wmv_rate mt_b ca il ne nj tx moreharsh lfp perblack adol violent, lag(1) t(mergeyear) force


Regression with Newey-West standard errors Number of obs = 61
maximum lag : 1 F( 10, 50) = 22.22
Prob > F = 0.0000

------------------------------------------------------------------------------
| Newey-West
wmv_rate | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
mt_b | 2.133085 1.140938 1.87 0.067 -.1585574 4.424727
ca | -218.6071 78.00666 -2.80 0.007 -375.2881 -61.9261
il | -158.486 48.60826 -3.26 0.002 -256.1185 -60.85343
nj | -226.1264 72.99826 -3.10 0.003 -372.7477 -79.50511
tx | -176.1477 60.56243 -2.91 0.005 -297.7909 -54.50448
moreharsh | -68.57535 26.88301 -2.55 0.014 -122.5715 -14.57924
lfp | 52.18322 12.67309 4.12 0.000 26.72858 77.63787
perblack | -1466.577 462.0384 -3.17 0.003 -2394.608 -538.5452
adol | -755.8335 258.1447 -2.93 0.005 -1274.332 -237.3346
violent_ar~s| -.0159294 .0164146 -0.97 0.336 -.0488991 .0170404
_cons | 489.9222 145.1734 3.37 0.001 198.3329 781.5115
------------------------------------------------------------------------------

From the results of these models, it seems changes in sentencing polices (variable mt_b) increase admission rates more for black men than for white men. I want a way to test that this difference is 'real'. I was told that suest would be the way to go, but none of the examples in the help file quite fit.

i had thought that the chow test example was relevant, but realize i am incorrect. in that example, the same model is run, once on men and once on women. in contrast, my dependent variable (rather than my sample) changes -- from blacks' admission rates to whites'. this is necessary since the i'm looking at state-year level rather than individual level outcomes. at this point, i am not even sure suest can address my question.

does anyone have any suggestions?

thanks,
traci


*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/




© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index