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

From |
Jeremy Reynolds <jeremyr@uga.edu> |

To |
statalist@hsphsun2.harvard.edu |

Subject |
st: compare coefficients from two fixed-effects models |

Date |
Fri, 21 Sep 2007 13:18:10 -0400 |

Dear All,

I would like to estimate two fixed-effects regressions using the same sample and independent variables but different dependent variables. Then I would like to compare the coefficients for one or more of the independent variables across the two equations. The issue has arisen in an attempt to implement Jeffrey Edwards' alternative approach to using difference scores as dependent variables with a large panel data set with up to three observations for each respondent (Edwards 1995, Organizational Behavior and Human Decision Processes 64(3) 307-324).

If I were estimating regular OLS models, mvreg would provide a nice way to compare the coefficients across models, but I'm not sure what the best approach is with fixed-effects models.

I have listed three options below, but all of them leave me with questions and there may be better alternatives.

I would appreciate any guidance you may have.

Thanks,

Jeremy

Option 1: use xtreg, fe to estimate the regressions and suest to compare coefficients

Question: Is the suest command appropriate here?

xtreg Y1 X1 X2, i(id) fe

est store eqn1

xtreg Y2 X1 X2, i(id) fe

est store eqn2

suest eqn1 eqn2

test [eqn1=eqn2]: X1

Option 2: use regress with de-meaned variables and clustering/robust standard errors instead of the xtreg command

Question: Should the standard errors from this approach to estimating fixed effects be the same as with xtreg? (They do not seem to be the same)

regress Y1 X1_demeaned X2_demeaned

est store eqn1

regress Y2 X1_demeaned X2_demeaned

est store eqn2

suest eqn1 eqn2, cluster(id)

test [eqn1=eqn2]: X1_demeaned

Option 3: rely on results from a model with the difference score as the dependent variable for tests

Question: The coefficients in the third equation can be obtained by taking the difference between those from the first two equations, but are the significance tests in equation 3 equivalent to the test command?

xtreg Y1 X1 X2, i(id) fe

xtreg Y2 X1 X2, i(id) fe

xtreg (Y2-Y1) = X1 X2, i(id) fe

--

********************

Dr. Jeremy Reynolds

Associate Professor

Undergraduate Coordinator

Department of Sociology

117 Baldwin Hall

University of Georgia

Athens, GA 30602-1611

Phone: (706) 583-8072

Web: http://uga.edu/soc/people/faculty/reynolds_jeremy.php

Fax: (706) 542-4320

*

* 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/

- Prev by Date:
**Random number routines [was: st: Simple Question - Use of rndbin]** - Next by Date:
**st: RE : st: solving for unobserved hetereogeneity with two simultaneous equations** - Previous by thread:
**st: Update to -estwrite- available from SSC** - Next by thread:
**st: ml max: technique** - Index(es):

© Copyright 1996–2017 StataCorp LLC | Terms of use | Privacy | Contact us | What's new | Site index |