Product-of-coefficients mediation estimator

\(ab\), with the Sobel standard error \(\sqrt{a^2 s_b^2 + b^2 s_a^2}\).

Usage

morie_mediation_product(a, b, se_a = NULL, se_b = NULL, alpha = 0.05)

Arguments

a, b

Path coefficients.

se_a, se_b

Optional standard errors, enabling the Sobel interval.

alpha

Two-sided level.

Value

A list with indirect, sobel_se, sobel_ci.

Details

The Sobel interval assumes \(ab\) is normal. It is not: the product of two normals is heavy-tailed and asymmetric, so a symmetric interval misses on one side, and it does so worst at the small-to-moderate effects mediation studies actually report. A percentile bootstrap respects that asymmetry.

References

MacKinnon DP (2008), Ch 3-4. Sobel ME (1982) Sociological Methodology 13:290-312.

Examples

V <- c(1, 2, 3, 4, 5, 6, 7, 8)
morie_mediation_product(V, V)
#> $indirect
#> [1]  1  4  9 16 25 36 49 64
#> 
#> $a
#> [1] 1 2 3 4 5 6 7 8
#> 
#> $b
#> [1] 1 2 3 4 5 6 7 8
#> 
#> $sobel_se
#> NULL
#> 
#> $sobel_ci
#> NULL
#> 
#> $sobel_symmetric
#> [1] FALSE
#> 
#> $method
#> [1] "Product-of-coefficients indirect effect"
#>