Describe a drawn sample

What you actually got, against what the design asked for. Worth a look before analysing: it surfaces strata that came up short, weights that vary more than you expected, and rows the design could never have reached.

Usage

sample_summary(sample)

Arguments

sample

A data frame returned by draw() with weights = TRUE.

Value

A list with a print() method, holding:

design

The design's type, as a string.

n, N

Rows drawn, and rows in the frame.

weight_range

The smallest and largest design weight.

weight_cv

Their coefficient of variation. Large values mean a few rows carry most of the estimate.

unreachable

Frame rows with inclusion probability 0 — the design could never have selected them. NA if the design has no closed-form inclusion probability.

by_group

A data frame of group, drawn, in_frame and rate per stratum or cluster, or NULL for a design with no grouping.

group_col

The column(s) by_group is keyed on.

See also

draw(), deff()

Examples

set.seed(1)
pop <- data.frame(
  id = 1:400,
  site = rep(c("a", "b", "c", "d"), times = c(200, 100, 60, 40))
)
s <- draw(pop, design_stratified("site", n = 40), seed = 1, weights = TRUE)
sample_summary(s)
#> Sample of 40 from 400  (stratified design)
#>   sampling fraction  0.1
#>   design weights     10 to 10   (cv 0)
#> 
#>   by site:
#>     group  drawn  in frame   rate
#>     a         20      200  0.100
#>     b         10      100  0.100
#>     c          6       60  0.100
#>     d          4       40  0.100