estimate_u {psychmeta}

Estimate u ratios from available artifact information

Description

Uses information about standard deviations, reliability estimates, and selection ratios to estimate u ratios. Selection ratios are only used to estimate u when no other information is available, but estimates of u computed from SDs and reliabilities will be averaged to reduce error.

Usage

estimate_u(
  measure_id = NULL,
  sdi = NULL,
  sda = NULL,
  rxxi = NULL,
  rxxa = NULL,
  item_ki = NULL,
  item_ka = NULL,
  n = NULL,
  meani = NULL,
  sr = NULL,
  rxya_est = NULL,
  data = NULL
)

Arguments

measure_id

Vector of measure identifiers.

sdi

Scalar or vector containing restricted standard deviation(s).

sda

Scalar or vector containing unrestricted standard deviation(s).

rxxi

Scalar or vector containing restricted reliability coefficient(s).

rxxa

Scalar or vector containing unrestricted reliability coefficient(s).

item_ki

Scalar or vector containing the number of items used in measures within samples.

item_ka

Scalar or vector indicating the number of items toward which reliability estimates should be adjusted using the Spearman-Brown formula.

n

Vector of sample sizes.

meani

Vector of sample means.

sr

Vector of selection ratios (used only when no other useable u-ratio inputs are available).

rxya_est

Vector of estimated unrestricted correlations between the selection mechanism and the variable of interest (used only when sr is used).

data

Optional data frame containing any or all information for use in other arguments.

Value

A vector of estimated u ratios.

Examples

sdi <- c(1.4, 1.2, 1.3, 1.4)
sda <- 2
rxxi <- c(.6, .7, .75, .8)
rxxa <- c(.9, .95, .8, .9)
item_ki <- c(12, 12, 12, NA)
item_ka <- NULL
n <- c(200, 200, 200, 200)
meani <- c(2, 1, 2, 3)
sr <- c(.5, .6, .7, .4)
rxya_est <- .5

## Estimate u from standard deviations only:
estimate_u(sdi = sdi, sda = sda)

## Estimate u from incumbent standard deviations and the
## mixture standard deviation:
estimate_u(sdi = sdi, sda = "mixture", meani = meani, n = n)

## Estimate u from reliability information:
estimate_u(rxxi = rxxi, rxxa = rxxa)

## Estimate u from both standard deviations and reliabilities:
estimate_u(sdi = sdi, sda = sda, rxxi = rxxi, rxxa = rxxa,
           item_ki = item_ki, item_ka = item_ka, n = n,
           meani = meani, sr = sr, rxya_est = rxya_est)

estimate_u(sdi = sdi, sda = "average", rxxi = rxxi, rxxa = "average",
           item_ki = item_ki, item_ka = item_ka, n = n, meani = meani)

## Estimate u from selection ratios as direct range restriction:
estimate_u(sr = sr)

## Estimate u from selection ratios as indirect range restriction:
estimate_u(sr = sr, rxya_est = rxya_est)