Test whether the effect of X differs across levels of a modifier

Effect modification and interaction are different claims. Interaction asks what happens when you set both variables; effect modification only asks whether the effect of one varies across strata defined by the other, and the modifier need not be causal – it can be a proxy or a stratifying label. Only the exposure need be unconfounded.

Usage

Fxidf(Y, X, C_mod)

Arguments

Y

Outcome.

X

Exposure.

C_mod

Candidate effect modifier.

Value

List with estimate, se, t, coef, effect_at_0, effect_at_1, n.

Details

Formula: fit Y = g0 + g1 X + g2 V + g3 X V; g3 is the additive modification and g1 + g3 v the stratum-specific effect.

References

VanderWeele, T. J. (2009). Epidemiology 20:863-871.

Examples

Fxidf(Y = c(1, 2, 3, 4, 5, 6, 7, 8), X = c(1, 2, 3, 4, 5, 6, 7, 8),
  C_mod = c(1, 2, 3, 4, 5, 6, 7, 8))
#> $estimate
#> [1] -4.597448e-17
#> 
#> $se
#> [1] 2.598062e-16
#> 
#> $t
#> [1] -0.1769568
#> 
#> $coef
#> [1]  3.201815e-16  5.000000e-01  5.000000e-01 -4.597448e-17
#> 
#> $effect_at_0
#> [1] 0.5
#> 
#> $effect_at_1
#> [1] 0.5
#> 
#> $n
#> [1] 8
#> 
#> $method
#> [1] "Additive effect modification, X by V interaction"
#> 
#> attr(,"class")
#> [1] "morie_rich_result" "list"