se.ave.mean2.dep: Computes the standard error for the average of 2-group mean...

In dgbonett/vcmeta: Varying Coefficient Meta-Analysis

Computes the standard error for the average of 2-group mean differences from two parallel measurement response variables in the same sample

Description

In a study that reports a 2-group mean difference for two response variables that satisfy the conditions of parallel measurments, this function can be used to compute the standard error of the average of the two mean differences using the two estimated means, estimated standard deviations, estimated within-group correlation between the two response variables, and the two sample sizes. The average mean difference and standard error output from this function can then be used as input in the meta.ave.gen, meta.lc.gen, and meta.lm.gen functions in a meta-analysis where some studies have used one of the two parallel response variables and other studies have used the other parallel response variable. Equality of variances is not assumed.

Usage

se.ave.mean2.dep(m1A, m2A, sd1A, sd2A, m1B, m2B, sd1B, sd2B, rAB, n1, n2)

Arguments

m1A	estimated mean for variable A in group 1
m2A	estimated mean for variable A in group 2
sd1A	estimated standard deviation for variable A in group 1
sd2A	estimated standard deviation for variable A in group 2
m1B	estimated mean for variable B in group 1
m2B	estimated mean for variable B in group 2
sd1B	estimated standard deviation for variable B in group 1
sd2B	estimated standard deviation for variable B in group 2
rAB	estimated within-group correlation between variables A and B
n1	sample size for group 1
n2	sample size for group 2

Value

Returns a one-row matrix:

• Estimate - estimated average mean difference

• SE - standard error

• VAR(A) - variance of mean difference for variable A

• VAR(B) - variance of mean difference for variable B

• COV(A,B) - covariance of mean differences for variables A and B

Examples

se.ave.mean2.dep(21.9, 16.1, 3.82, 3.21, 24.8, 17.1, 3.57, 3.64, .785, 40, 40)

# Should return:
#                          Estimate        SE    VAR(A)    VAR(B)  COV(A,B)
# Average mean difference:     6.75 0.7526878 0.6224125 0.6498625 0.4969403

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