Propensity Score matching problem

[Collins Mukhala]

Dear Community Experts,

I am doing propensity score analysis to estimate impact of innovation hub training on digital inclusion on a number of specific areas using nearest neighbour in StataNow (18.5). The outcome variables are Score, with some having zero values. I previously ran all the analysis quite well without getting any error. However when I came to this variable shown in bold the picture, I am getting the error shown and I am not getting the full output or totally none. I have looked in the dataset, there is no such a variable that Stata has created. I am stuck and would like your help how I can surmount this problem. Here are my syntaxes and I attach my dataset.

I would greatly appreciate your help.

Regards,

. psmatch2 Admitted Sex Age Education Hub Program, out(digital_marketing_platform_score)

Probit regression Number of obs = 361

LR chi2(5) = 6.72

Prob > chi2 = 0.2421

Log likelihood = -234.74935 Pseudo R2 = 0.0141

Admitted Coefficient Std. err. z P>z [95% conf. interval]

Sex -.2784766 .1457454 -1.91 0.056 -.5641324 .0071792

Age -.0417891 .1228007 -0.34 0.734 -.282474 .1988958

Education .050571 .0723285 0.70 0.484 -.0911902 .1923323

Hub .1710627 .1382033 1.24 0.216 -.0998108 .4419363

Program .127274 .1163634 1.09 0.274 -.1007941 .355342

_cons .1983421 .3905123 0.51 0.612 -.567048 .9637322

_digital_marketing_platform_score invalid varname

r(198); /*############ PROPENSITY SCORE ANALYSIS ON DIGITAL PLATFORMS FOR MARKETING #####*/ pscore Admitted Sex Age Education Hub Program, pscore(mypscore) blockid(myblock) comsup psmatch2 Admitted Sex Age Education Hub Program, logit /*PSTEST assesses balance in the matched samples*/ pstest Sex Age Education Hub Program, treated(Admitted) weight(_weight) both /*Viewing matched pairs*/ psgraph, treated(Admitted) support(comsup) pscore(mypscore) /*Estimating the Treatment Effects suing psmatch2 command without covariates*/ psmatch2 Admitted, outcome(digital_marketing_platform_score) pscore(mypscore) neighbor(1) /*Estimating the Treatment Effects using psmatch2 command with covariates*/ psmatch2 Admitted Sex Age Education Hub Program, out(digital_marketing_platform_score) psmatch2 Sex Age Education Hub Program, outcome(digital_marketing_platform_score) neighbor(1) psmatch2 Admitted, outcome(digital_marketing_platform_score) pscore(mypscore) neighbor(2) /*##### USING TEFFECTS PSMATCH, NNMATCH AND PSMATCH ATE #####################*/ teffects psmatch (digital_marketing_platform_score) (Admitted Sex Age Education Hub Program, probit),atet teffects nnmatch (Sex Age Education Hub Program) (Admitted), atet control(0) tlevel(1) vce(robust, nn(4)) psmatch2 Admitted Sex Age Education Hub Program, out(digital_marketing_platform_score) logit ate

[Clyde Schechter]

_digital_marketing_platform_score has 33 characters. Stata variable names are limited to 32 characters. Pick a shorter name.

[Collins Mukhala]

Dear Clyde,

Sorry for my late feedback. My appreciation to your help. I rename the variables that were giving such and error to short names and it worked perfectly well.

Regards