Notice: On March 31, it was **announced** that Statalist is moving from an email list to a **forum**. The old list will shut down on April 23, and its replacement, **statalist.org** is already up and running.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

From |
"Wald,Kenneth D" <kenwald@ufl.edu> |

To |
"statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu> |

Subject |
st: How to extract matched sample in psmatch2 |

Date |
Sun, 30 Oct 2011 21:02:58 +0000 |

I'm using psmatch2 to generate a comparison sample of one group of survey respondents with another, defined by religious affiliation. I've run a basic analysis but can't figure out how the program gets to the next step--creating the matched sample so I can confirm that the covariates are indeed balanced. For the moment, I don't want to run multivariate analysis or use weighting. I just want examine the values of certain other variables in the two groups, the base and the matched subsample. I've read the documentation and Guo and Fraser but nothing walks me through the program. So I'd be grateful if somebody could recommend something that really explains the Stata routine to do that. Here's the commands I've used and the output. The program generates pscores (_pscore). I did randomize the order of cases prior to estimating the equation as advised. Two other questions if you're inclined. With a dichotomous grouping variable, why did Stata use probit instead of logit which I thought was the default. And, given the distinctive geographical distribution of the base group, I used two location variables--state and region. They are nominal variables and so don't really belong in this kind of model. (I also need the year variable so the sample is balanced on that as well.) Thanks. psmatch2 jewish citysize nielsen employ year race3 region9 sex state agematch faminc educ3, caliper(11.8) Probit regression Number of obs = 91143 LR chi2(11) = 3863.45 Prob > chi2 = 0.0000 Log likelihood = -8533.966 Pseudo R2 = 0.1846 ------------------------------------------------------------------------------ jewish | Coef. Std. Err. z P>|z| [95% Conf. Interval] -------------+---------------------------------------------------------------- citysize | .1399825 .0074226 18.86 0.000 .1254345 .1545305 nielsen | .26099 .023412 11.15 0.000 .2151035 .3068766 employ | -.0502017 .0233954 -2.15 0.032 -.0960559 -.0043475 year | -.0005044 .0023767 -0.21 0.832 -.0051626 .0041538 race3 | 1.132129 .069947 16.19 0.000 .9950352 1.269222 region9 | -.0962822 .0044346 -21.71 0.000 -.1049738 -.0875906 sex | .0162921 .0209977 0.78 0.438 -.0248626 .0574468 state | -.0041037 .000857 -4.79 0.000 -.0057833 -.002424 agematch | .0741419 .0133769 5.54 0.000 .0479236 .1003602 faminc | .0929483 .0111876 8.31 0.000 .071021 .1148755 educ3 | .3533498 .0197301 17.91 0.000 .3146796 .39202 _cons | -4.425644 4.717873 -0.94 0.348 -13.6725 4.821217 ------------------------------------------------------------------------------ There are observations with identical propensity score values. The sort order of the data could affect your results. Make sure that the sort order is random before calling psmatch2. Kenneth D. Wald Distinguished Professor of Political Science University of Florida POB 117325 Gainesville, FL 32611-7325 Office Phone: 352-273-2391 Email: kenwald@ufl.edu Home page: www.clas.ufl.edu/users/kenwald Fall 2011 Semester: Frankel Center for Advanced Judaic Studies 2111 Thayer Bldg. University of Michigan 202 S. Thayer Street Ann Arbor, MI 48104-1608 Department Phone: 734-615-5097 Department Fax: 734-963-2186 Office Phone: 734-615-8541 * * 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/

- Prev by Date:
**st: Grid lines on top of shaded time series graph** - Next by Date:
**Re: st: How to extract matched sample in psmatch2** - Previous by thread:
**st: Grid lines on top of shaded time series graph** - Next by thread:
**Re: st: How to extract matched sample in psmatch2** - Index(es):