Weight your survey — weight

Generate statistical weights for your survey responses to better represent some population of interest.

Usage

weight_to(
  data,
  target = "us_genpop_acs18",
  auto.remove = TRUE,
  initial.weights = NULL,
  trim.weights = c(0.01, 0.99),
  verbose = TRUE
)

Arguments

data: input data frame of survey responses.
target: either name of target population (run list_targets() for options) or a list of dataframes corresponding to different population distributions to weight survey to.
auto.remove: remove any weighting variables that can't be found in survey.
initial.weights: initial set of weights as starting point for raking algorithm.
trim.weights: percentiles to trim your weights (default are 0.01 and .99, or 1% and 99%); can specify either an upper percentile or both a lower and upper percentile.
verbose: output helpful progress and warning messages (recommended).

Value

output contains (1) estimated weights and (2) summary of how weights will affect your data

Details

The default SurveyMonkey to weight surveys is via raking. For an overview of raking, see this guide.

Examples


if (FALSE) { ## not run
data("smda23")
tgt <- get_target("us_genpop_acs19")
wtd <- weight_to(smda23, target = tgt)

hist(wtd$weights)
print(wtd$weight.summary)
}