Effective simple-random-sample size of a weighted design

deff = 1 + cv^2(w) = n * sum(w^2) / (sum w)^2 and n_eff = n / deff = (sum w)^2 / sum(w^2).

Usage

Neffsrs(w)

Arguments

w

Positive selection weights, one per unit.

Value

List with neff, deff, cv2, n, sumw, sumw2.

References

Kish, L. (1965), Survey Sampling, Wiley, Sect. 11.7; standard published form.

Examples

V <- c(1, 2, 3, 4, 5, 6, 7, 8)
Neffsrs(V)
#> $neff
#> [1] 6.352941
#> 
#> $deff
#> [1] 1.259259
#> 
#> $cv2
#> [1] 0.2592593
#> 
#> $n
#> [1] 8
#> 
#> $sumw
#> [1] 36
#> 
#> $sumw2
#> [1] 204
#> 
#> $method
#> [1] "Kish effective sample size, deff = 1 + cv^2(w)"
#> 
#> attr(,"class")
#> [1] "morie_rich_result" "list"