Title


mrmedianobs -- Weighted median of instrumental variable estimates

Syntax

mrmedianobs depvar [varlist1] (varname_endog = varlist_ivs) [if] [in] [, options]

options Description ---------------------------------------------------------------------------------------------- all report percentile and bias corrected confidence intervals level(#) set confidence level; default is level(95) obsboot obtain standard error by bootstrapping at observation level penweighted penalized weighted estimator reps(#) number of bootstrap replications to obtain standard error seed(#) seed for random number generator for bootstrapping to obtain standard error weighted weighted estimator

Description

mrmedianobs performs unweighted, weighted, and penalized IV estimator on observation level data (Bowden et al., 2016).

depvar outcome variable varlist1 covariates to adjust for varname_endog exposure/treatment received/endogenous variable varlist_ivs instrumental variables

Options

all report percentile and bias corrected bootstrap confidence interval limits (only applies to observation level bootstrapping with obsboot).

level(#); see [R] estimation options.

obsboot obtain bootstrap standard error by bootstrapping at the observation level.

reps(#) specifies the number of bootstrap replications for obtaining the standard error. The default is 50 replications.

seed(#) specifies the initial value of the random-number seed. The default is the current random-number seed. Specifying seed(#) is the same as typing set seed # before issuing the command; see set_seed.

weighted use weights.

penweighted use penalized weights.



Examples

Simulated test dataset.

Setup . use https://raw.github.com/remlapmot/mrrobust/master/mrmedianobs_testdata, clear

Unweighted median estimator . mrmedianobs y (x = z1-z20)

Weighted median estimator . mrmedianobs y (x = z1-z20), weighted

Penalized weighted median estimator . mrmedianobs y (x = z1-z20), penweighted

Unweighted median estimator with percentile CI limits of observation level bootstrapping . mrmedianobs y (x = z1-z20), obsboot all



Stored results

mrmedian stores the following in e():

Scalars e(k) number of instruments e(N) number of observations e(reps) number of (bootstrap) replications

Macros e(cmd) mrmedianobs e(cmdline) command as typed

Matrices e(b) coefficient vector e(V) variance-covariance matrix of the estimates

References

Bowden J, Davey Smith G, Haycock PC, Burgess S. Consistent estimation in Mendelian randomization with some invalid instruments using a weighted median estimator. Genetic Epidemiology, 2016, 40, 4, 304-314. DOI

Author

INCLUDE help mrrobust-author