iddplacebo: Run placebo studies on untreated units and estimate placebo...

In carlbona/idd: Incidence difference-in-differences methods with automatic control selection

Description

iddplacebo estimates placebo studies on untreated units and produces pseudo p-values based on the empirical distribution of post/pre-RMSE ratios.

Usage

iddplacebo(eventvar, popvar, treatvar, postvar, timevar, idvar, data,
  iter = 10)

Arguments

eventvar	Name of the event count variable.
popvar	The person-time (or another exposure) variable.
treatvar	The treatment group indicator dummy (0 for controls, 1 for treated units).
postvar	The post-period indicator dummy (0 for all time points in the pre-intervention period, 1 for all time points in the post-period)
timevar	The time variable (can be coded on any scale).
idvar	The panel id variable.
data	A long-form panel dataset containing the supplied variables.
iter	The number of subsampling iterations. Defaults to 10, but 500-1000 are usually recommended.

Value

Returns a list containing the following elements: $Resdata: a data frame containing the results. $ECDF: the empirical cumulative distribution function. $Effects: a long-form data frame containing the estimated placebo effects. $Treat.ratio: a data frame containing the post/pre-RMSE ratio for the treated unit. $supp_stats: a data frame containing supplementary statistics (p-values etc).

Examples

## Not run: 
data(simpanel)
placebo.out <- iddplacebo(eventvar="y",
             popvar="pop",
             idvar="age",
             timevar="time",
             postvar="post",
             treatvar="treat",
             data=simpanel,
             iter=50)

## End(Not run)

carlbona/idd documentation built on June 12, 2018, 2:23 p.m.