perc_diff: Calculate percentile differences from an ordered categorical...

In cimentadaj/perccalc: Estimate Percentiles from an Ordered Categorical Variable

Description

Calculate percentile differences from an ordered categorical variable and a continuous variable.

Usage

perc_diff(
  data_model,
  categorical_var,
  continuous_var,
  weights = NULL,
  percentiles = c(90, 10)
)

perc_diff_df(
  data_model,
  categorical_var,
  continuous_var,
  weights = NULL,
  percentiles = c(90, 10)
)

Arguments

data_model	A data frame with at least the categorical and continuous variables from which to estimate the percentile differences
categorical_var	The bare unquoted name of the categorical variable. This variable SHOULD be an ordered factor. If not, will raise an error.
continuous_var	The bare unquoted name of the continuous variable from which to estimate the percentiles
weights	The bare unquoted name of the optional weight variable. If not specified, then estimation is done without weights
percentiles	A numeric vector of two numbers specifying which percentiles to subtract

Details

perc_diff drops missing observations silently for calculating the linear combination of coefficients.

Value

perc_diff returns a vector with the percentile difference and its associated standard error. perc_diff_df returns the same but as a data frame.

Examples

set.seed(23131)
N <- 1000
K <- 20

toy_data <- data.frame(id = 1:N,
                       score = rnorm(N, sd = 2),
                       type = rep(paste0("inc", 1:20), each = N/K),
                       wt = 1)


# perc_diff(toy_data, type, score)
# type is not an ordered factor!

toy_data$type <- factor(toy_data$type, levels = unique(toy_data$type), ordered = TRUE)

perc_diff(toy_data, type, score, percentiles = c(90, 10))
perc_diff(toy_data, type, score, percentiles = c(50, 10))

perc_diff(toy_data, type, score, weights = wt, percentiles = c(30, 10))
# Results as data frame
perc_diff_df(toy_data, type, score, weights = wt, percentiles = c(30, 10))

cimentadaj/perccalc documentation built on May 28, 2020, 1:06 p.m.