Kish's effective sample size

Computes the effective sample size using Kish's formula: $$ESS = \frac{(\sum w)^2}{\sum w^2}$$

Usage

effective_sample_size(x, weights = NULL)

Arguments

x

A data.frame, weighted_df, survey_taylor, or survey_nonprob. For weighted_df and survey objects, the weight column is auto-detected.

weights

Bare name (NSE). Weight column. Auto-detected for weighted_df and survey objects. Required for plain data.frame.

Value

A named numeric scalar: c(n_eff = <value>). The name "n_eff" is part of the API contract.

See also

Other diagnostics: summarize_weights(), weight_variability()

Examples

df <- data.frame(x = 1:5, w = c(1.2, 0.8, 1.5, 0.9, 1.1))
effective_sample_size(df, weights = w)
#>   n_eff 
#> 4.76378