Dear Statalist friends,
I am estimating a Propensity Score Matching (PSM) for a complex survey. Since the command does not let the use of weights as it is mentioned in the help file
I decided to use the
command.
With this I was wondering if I should use the option
for the standard errors since the help file mentions
I was wondering what is the difference between these two variances since when estimated with the population option the T-stat is almost ten times bigger than the sample variance.
If anyone knows which would be the best course of action regarding (i) the use of expand in replacement of the sampling weights and (ii) the use of the population variance, please let me know.
Thank you!
I am estimating a Propensity Score Matching (PSM) for a complex survey. Since the command does not let the use of weights as it is mentioned in the help file
As far as we know it's not really clear in the literature how to accommodate sample weights in the context of matching. If you are aware how to properly account for sampling weights, please let us know.
Code:
expand
With this I was wondering if I should use the option
Code:
population
population When using ai(integer), calculate the population variance rather than the sample variance
If anyone knows which would be the best course of action regarding (i) the use of expand in replacement of the sampling weights and (ii) the use of the population variance, please let me know.
Thank you!