Meta-analysis pooling: weights, mean effect, Q, I squared, tau squared

Equations 11.34 to 11.46 of Weisburd et al. (2022): inverse-variance weights, the weighted mean effect with standard error and z, the Q homogeneity statistic (definitional and computational forms agree), Higgins' I squared, the DerSimonian-Laird tau squared, random-effects weights, and the Qwithin / Qbetween partition. Mirrors ca11e34 to ca11e46.

Usage

morie_meta_pool(ys, ses, z_cv = 1.96, groups = NULL)

Arguments

ys: Effect sizes.
ses: Their standard errors.
z_cv: Critical z for the confidence interval.
groups: Optional integer group labels for the moderator partition.

Value

List with weights, mean, se, z, ci, q, df, i2, tau2, weights_random, and q_within / q_between when groups are given.

Examples

V <- c(1, 2, 3, 4, 5, 6, 7, 8)
morie_meta_pool(V, V)
#> $weights
#> [1] 1.00000000 0.25000000 0.11111111 0.06250000 0.04000000 0.02777778 0.02040816
#> [8] 0.01562500
#> 
#> $mean
#> [1] 1.779375
#> 
#> $se
#> [1] 0.809134
#> 
#> $z
#> [1] 2.199111
#> 
#> $ci
#> [1] 0.1934726 3.3652781
#> 
#> $q
#> [1] 3.163912
#> 
#> $df
#> [1] 7
#> 
#> $i2
#> [1] 0
#> 
#> $tau2
#> [1] 0
#> 
#> $weights_random
#> [1] 1.00000000 0.25000000 0.11111111 0.06250000 0.04000000 0.02777778 0.02040816
#> [8] 0.01562500
#>