hedges2007: Estimate standardized mean differences for clustered data

In knickodem/WBdif: Investigate DIF and generate reports

Estimate standardized mean differences for clustered data

Description

Computes mean difference standardized for clustered data by the "total variance", and their standard errors, using the method described by Hedges (2007).

Usage

hedges2007(
  outcome,
  groups,
  pooled = TRUE,
  comp.name = NULL,
  cluster = NULL,
  subset = NULL
)

Arguments

outcome	numeric vector
groups	vector of length(outcome) indicating the two groups to compare
pooled	If TRUE (default), the pooled standard deviation is calculated to standardize the mean difference. If FALSE, the standard deviation of comp.name will be used.
comp.name	If pooled = FALSE, the name of the group in groups whose standard deviation will be used to standardize the mean difference.
cluster	optional vector of length(outcome) indicating the primary sampling unit in a multi-stage / clustered sampling design
subset	optional logical vector of length(outcome) indicating a subset of the data to use in the calculations

Details

The standardized mean difference and its standard error are computed using Equations 15 and 16 of Hedges (2007). When cluster sample sizes are unequal, the average cluster size is used to obtain a close approximation of the standardized mean difference estimate and standard error (p. 351).

Value

A named vector containing the standardized mean difference and its standard error.

References

Hedges, L. V. (2007). Effect sizes in cluster-randomized designs. Journal of Educational and Behavioral Statistics, 32(4), 341–370. https://doi.org/10.3102/1076998606298043.

knickodem/WBdif documentation built on Feb. 3, 2024, 2:20 a.m.