Simulate a derived ACS statistic.

Simulate a derived ACS statistic.

Usage

acs_simulate_fn(
  estimates,
  moes,
  fn,
  cov = NULL,
  n_sims = 1000,
  dist = c("normal", "censored_normal"),
  conf = 0.9,
  summary = c("mean", "median", "ci"),
  point = c("mean", "median")
)

Arguments

estimates

Numeric estimates.

moes

Numeric MOEs corresponding to estimates.

fn

Function applied to each simulated row.

cov

Optional covariance matrix on the standard-error scale.

n_sims

Number of Monte Carlo simulations.

dist

Distribution assumption: "normal" or "censored_normal". The censored variant replaces below-zero draws with zero, matching the convention used in Napierala & Denton (2017) for ACS counts.

conf

Confidence level associated with input MOEs.

summary

Summary to return: "mean", "median", or "ci".

point

For summary = "ci", point estimate to report alongside the percentile interval: "mean" (default) or "median".

Value

A data frame summarizing the simulated derived statistic.

Details

For summary = "ci", the returned interval is the central conf-level percentile interval of the simulated derived values. The reported estimate is the chosen point summary of those values.

Examples

set.seed(1)
acs_simulate_fn(c(100, 50), c(10, 5), fn = sum, n_sims = 500)
#>   estimate       se
#> 1  150.002 6.822565
set.seed(1)
acs_simulate_fn(c(100, 50), c(10, 5), fn = sum, n_sims = 500,
                summary = "ci", conf = 0.90)
#>   estimate    lower   upper
#> 1  150.002 138.8195 161.133