倾向值分析--统计方法与应用 (美)郭申阳//马克·W.弗雷泽|译者:郭志刚//巫锡炜
作     者(美)郭申阳//马克·W.弗雷泽|译者:郭志刚//巫锡炜
出 版 社重庆大学
出版时间2012-6-1
ＩＳＢＮ9787562466222

  USING PROPENSITY SCORES IN QUASI-EXPERIMENTAL DESIGNS

William M. Holmes

STATA COMMANDS FOR PROPENSITY USE



            Shenyang Guo and Mark W. Fraser (2010) have written an excellent book on propensity score analysis that provides many Stata commands. While their treatment is more technical than this book, the Stata commands are quite useful. The support site for their book is:  http://ssw.unc.edu/psa/.  The Stata Journal article by Becker and Ichino (2002) provides a good overview of propensity score use and extensive commands with Stata. It can be accessed at http://www.stata-journal.com/sjpdf.html?articlenum=st0026.



IMBALANCE ASSESSMENT

            Imbalance may be assessed with the ANOVA command or ttest.

            

            . anova confounder1 treatment



            .ttest confounder 1 treatment



PROPENSITY ESTIMATION

            Propensities may be estimated with logistic, probit, or linear regression or with discriminant function analysis commands when combined with the predict command.



. logistic confounder1 confounder2 confounder3 i.treatment

. predict propen, xb



. probit treatment confounder1 confounder2 confounder3

. predict propen



. regress treatment confounder1 confounder2 confounder3

. predict propen



. candisc confounder1 confounder2 confounder3, group(treatment)

. predict propen



MATCHING

            The psmatch2 program provides a means for propensity score matching within Stata.  Alternative matching programs may be accessed with the R interface for Stata.



            Exact and Coarsened Exact

            See psmatch2 within Stata



            Nearest Neighbor and Caliper

            See psmatch2 within Stata



            1-Many

            See psmatch2 within Stata



            Optimized, Full, or Genetic

            To do optimized, full, or genetic matching within Stata, the R interface must be used to access the corresponding R programs.



STRATIFYING

            Creating propensity strata can be done with the recode command.



            recode propen 0/.19999=1 .2/.39999=2 .4/.59999=3 .6/.79999=4 .8/.99999=5, gen(propenstrata)

            



REGRESSION AND THE GENERAL LINEAR MODEL

            Regression and General Linear Model estimation may be done with the regress and glm commands.



             . regress outcome1  i.treatment



            . glm outcome1 i.treatment, family(gaussian) link(identity)



TWO-STAGE LEAST SQUARES

            2SLS regression can be done with the ivregress command.



. ivregress 2sls outcome1(treatment = instrument1) , vce(robust)



SAMPLE WEIGHTING

Sample weighting is done within the analytical commands using   pweight. For example with a logit analysis, the command would be:



logit treatment confounder1 confounder2 [pweight=ipa]



WEIGHTED LEAST SQUARES

            Weighted Least Squares may be done within analytical program (see sample weighting above) or with the wls0 command.



.wls0 outcome treatment, wvar(ipa) type(abse) noconst graph



GENERALIZED LINEAR MODEL



            . glm income educ jobexp i.black, family(gaussian) link(identity)



MISSING DATA ANALYSIS

            Analysis of missing data can combine mdesc with mvpatterns or with the misstable command.



            mdesc confounder1 confounder2 confounder3         



            mvpatterns confounder1 confounder2 confounder3



            . misstable summarize



IMPUTATION OF MISSING DATA

            Multiple imputation can be done with the mi impute command. Variables to be imputed must first be registered.  The example below imputes 5 values with a random seed of 123. Estimation of imputed data in Stata may be done with the mi estimate command.

            . mi register imputed confounder1 confounder2

            . mi impute regress confounder1  confounder2 confounder3, add(5) rseed(123)



            . mi estimate: logit treatment confounder1 confounder2 confounder3

            

return to Home                                   

Revised 1/31/2013                                                        
http://www.faculty.umb.edu/william_holmes/propensitystata.htm

Propensity Score Analysis: Statistical Methods and Applications, Second Edition，这本书专门有一章讨论以及提供了一些命令示例，但是还是有点不清晰，我最近也在学习，不太懂

请移步gpscore2命令。加油！我已完成论文许久，而忙于学习机器学习，淡忘已不少知识，还请您多看论坛。

许久不上论坛，谢谢您的推荐

多谢您！

你好，我在论坛上看到你也关注GPS广义倾向得分匹配了，看完SJ还是有俩问题不懂，想请教下你呢：1GPS究竟是怎么找到匹配组的呢，比如研究出口强度为50%下的生产率，是找不出口企业和真实出口强度就为50%的企业一起匹配吗？2他的平衡性检验又是检验了哪两个组是否相等呢？

抱歉回复的晚了些。我未曾系统的学习gps，仅学习了如何实现结果。且半年有余而未曾接触gps。才疏学浅，不敢回复您的问题。不过，help原始命令，可以看到实现代码，代码中将回答您的两个疑问。原作者的stataj文章，也有介绍。还请您自行寻找材料了。抱歉。

谢谢！

你好，我能购买你的gpscore2命令使用教程吗？我最近才学习，Fraction Logit 估计GPS遇到困难了，一周了也没解决掉这个问题。请好心人解惑啊

您好，我最近在广义请相匹配上遇到了问题，能加您QQ或者微信吗？愿意有酬劳的请教您问题。我的QQ和微信都是724696126，期待您的回信。

我是用gpscore2命令出现r(420)错误

我是用gpscore2命令出现r(420)错误

我是用gpscore2命令出现r(420)错误

你可以试试doseresponse2命令

这本书里面还没有详细介绍GPS