meta.sub.cor: Confidence interval for a subgroup difference in average...

In dgbonett/vcmeta: Varying Coefficient Meta-Analysis

Confidence interval for a subgroup difference in average Pearson or partial correlations

Description

Computes the estimate, standard error, and confidence interval for a difference in average Pearson or partial correlations for two mutually exclusive subgroups of studies. Each subgroup can have one or more studies. All of the correlations must be either Pearson correlations or partial correlations.

Usage

meta.sub.cor(alpha, n, cor, s, group)

Arguments

alpha	alpha level for 1-alpha confidence
n	vector of sample sizes
cor	vector of estimated correlations
s	number of control variables (set to 0 for Pearson)
group	vector of group indicators: 1 for set A 2 for set B 0 to ignore

Value

Returns a matrix with three rows:

• Row 1 - estimate for Set A

• Row 2 - estimate for Set B

• Row 3 - estimate for difference, Set A - Set B

The columns are:

• Estimate - estimated average correlation or difference

• SE - standard error

• LL - lower limit of the confidence interval

• UL - upper limit of the confidence interval

References

\insertRef

Bonett2008avcmeta

Examples

n <- c(55, 190, 65, 35)
cor <- c(.40, .65, .60, .45)
group <- c(1, 1, 2, 0)
meta.sub.cor(.05, n, cor, 0, group)

# Should return:
#                Estimate         SE         LL        UL
# Set A:            0.525 0.06195298  0.3932082 0.6356531
# Set B:            0.600 0.08128008  0.4171458 0.7361686
# Set A - Set B:   -0.075 0.10219894 -0.2645019 0.1387283

dgbonett/vcmeta documentation built on July 12, 2024, 3:12 p.m.