R/rd-mm_randhie.R

In stevedata: Steve's Toy Data for Teaching About a Variety of Methodological, Social, and Political Topics

#' @importFrom tibble tibble
NULL

#' Data from the RAND Health Insurance Experiment (HIE)
#'
#' These are data from the RAND Health Insurance Experiment (HIE).
#' People who have read \emph{Mastering 'Metrics} should recognize these data. They're
#'  featured prominently in that book and the authors' discussion of
#'  random assignment and experiments.
#'
#' @format The data are a list of two data frames (or "tibbles"). The first is the baseline data.
#'
#' \describe{
#' \item{\code{plantype}}{the plan coverage of the respondent, as a factor}
#' \item{\code{age}}{the age of the respondent}
#' \item{\code{blackhisp}}{whether the respondent is not white}
#' \item{\code{cholest}}{the cholesterol level of the respondent (in \code{mg/dl})}
#' \item{\code{educper}}{the education-level of the respondent}
#' \item{\code{female}}{whether the respondent is a woman}
#' \item{\code{ghindx}}{a general health index}
#' \item{\code{hosp}}{was the respondent hospitalized last year?}
#' \item{\code{income1cpi}}{the family/household income of the respondent, adjusted for inflation}
#' \item{\code{mhi}}{a mental health index}
#' \item{\code{systol}}{the systolic blood pressure level of the respondent (in \code{mm HG})}
#' }
#'
#' The second is the outcome data.
#'
#' \describe{
#' \item{\code{plantype}}{the plan coverage of the respondent, as a factor}
#' \item{\code{ftf}}{the number of face-to-face visits for the respondent}
#' \item{\code{out_inf}}{the total of out-patient expenses for the respondent}
#' \item{\code{totadm}}{the number of hospital admissions for the respondent}
#' \item{\code{tot_inf}}{the total health expenses for the respondent}
#' }
#'
#' @details Data are already cleaned in a way that facilitates an easy
#' replication of Table 1.3 and a partial replication of
#' Table 1.4 in \emph{Mastering 'Metrics}. Check
#' the book's website for more information. I want to note
#' that my treatment of the data leans heavily on Jeff Arnold's
#' treatment of it. Check \url{https://jrnold.github.io/masteringmetrics/} for
#' more information. Future updates to the data may pursue a more exhaustive
#' replication. I will only note these data are a mess and the authors of
#' \emph{Mastering 'Metrics} do not do a great job annotating code.
#'
#' @source RAND Health Insurance Experiment.
#'
"mm_randhie"