Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
| From | Ian Li <ianz.magic@hotmail.com> |
| To | Statalist <statalist@hsphsun2.harvard.edu> |
| Subject | st: Oaxaca Decomposition with Deviation Transform Coding |
| Date | Mon, 2 Apr 2012 03:01:27 +0000 |
Dear Statalist, I would like to perform a Blinder-Oaxaca decomposition that accounts for the choice of the reference category for categorical variables (see Jones and Kelley (1984) "Decomposing Differences Between Groups: A Cautionary Note on Measuring Discrimination", Sociological Methods and Research). I use the -oaxaca- command written by Ben Jann, and has also read the Stata Journal article written by him on the use of this command (Jann (2008) "The Blinder-Oaxaca Decomposition for Linear Regression Models", The Stata Journal vol. 8, no. 4, pp. 453-479). In estimating the model using the 'categorical' command to perform deviation transform coding, however, I get the exact same result to that when the 'categorical' option is not specified. This makes me suspicious of the output result. However, I can't see any mistakes with my specification of the command. I have copied the output below, and would be grateful to anyone who can point out where I have gone wrong. Regards Ian do "C:\Users\Ian\AppData\Local\Temp\STD00000000.tmp" . oaxaca lnhrsal natphysci infotech engineer archi agrienv nursing medicine educ soccult creatoth go8 tech iru if empnat1 == > 1, by(female) weight(0.5) Blinder-Oaxaca decomposition Number of obs = 569325 1: female = 0 2: female = 1 ------------------------------------------------------------------------------- lnhrsal | Coef. Std. Err. z P>|z| [95% Conf. Interval] --------------+---------------------------------------------------------------- Differential | Prediction_1 | 3.065791 .0013305 2304.22 0.000 3.063184 3.068399 Prediction_2 | 2.970103 .0010395 2857.33 0.000 2.968066 2.972141 Difference | .095688 .0016884 56.67 0.000 .0923787 .0989972 --------------+---------------------------------------------------------------- Decomposition | Explained | .0041661 .0008113 5.13 0.000 .0025759 .0057563 Unexplained | .0915218 .0018572 49.28 0.000 .0878817 .095162 ------------------------------------------------------------------------------- . end of do-file . do "C:\Users\Ian\AppData\Local\Temp\STD00000000.tmp" . oaxaca lnhrsal natphysci infotech engineer archi agrienv nursing medicine educ soccult creatoth go8 tech iru if empnat1 == > 1, by(female) weight(0.5) categorical (natphysci infotech engineer archi agrienv nursing medicine educ soccult creatoth mgt > comm, go8 tech iru otheruni) Blinder-Oaxaca decomposition Number of obs = 569325 1: female = 0 2: female = 1 ------------------------------------------------------------------------------- lnhrsal | Coef. Std. Err. z P>|z| [95% Conf. Interval] --------------+---------------------------------------------------------------- Differential | Prediction_1 | 3.065791 .0013305 2304.22 0.000 3.063184 3.068399 Prediction_2 | 2.970103 .0010395 2857.33 0.000 2.968066 2.972141 Difference | .095688 .0016884 56.67 0.000 .0923787 .0989972 --------------+---------------------------------------------------------------- Decomposition | Explained | .0041661 .0008113 5.13 0.000 .0025759 .0057563 Unexplained | .0915218 .0018572 49.28 0.000 .0878817 .095162 ------------------------------------------------------------------------------- . end of do-file * * 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/