Skip to contents

Natural indirect effect (Pearl 2001)

Source: R/nie.R
Nie.Rd

DUPLICATE: the contrast is already implemented as Nieff in unclr.R; this is an alias, not a second copy.

Usage

Nie(y11, y10)

Arguments

y11

Unit-level Y(1, M(1)).

y10

Unit-level Y(1, M(0)), the cross-world outcome.

Value

List with estimate, se, mean_y11, mean_y10, n.

Details

Formula: NIE = E[Y(1, M(1))] - E[Y(1, M(0))], taken within unit so the standard error is the paired one.

References

Pearl, J. (2001). Direct and indirect effects. Proceedings of the Seventeenth Conference on Uncertainty in Artificial Intelligence, 411-420. Morgan Kaufmann.

Examples

V <- c(1, 2, 3, 4, 5, 6, 7, 8)
Nie(V, V)
#> $estimate
#> [1] 0
#> 
#> $se
#> [1] 0
#> 
#> $mean_y11
#> [1] 4.5
#> 
#> $mean_y10
#> [1] 4.5
#> 
#> $n
#> [1] 8
#> 
#> attr(,"class")
#> [1] "morie_rich_result" "list"