Expand a categorical focal exposure into one directional design per contrast

DMSA pools signed per-probe effects, so a focal factor with k > 2 levels cannot be tested as one directional term. This builds the k - 1 single-contrast designs, each of which is directional, and states the multiplicity correction that applies across them.

Usage

dmsa_contrasts(
  data,
  design,
  reference = NULL,
  correct = c("holm", "BH", "bonferroni", "none")
)

Arguments

data

The analysis data frame.

design

A dmsa_design whose focal_test names a factor (or character, or a numeric column with few unique values) with k > 2 levels.

reference

Level to contrast against. Defaults to the first level, which for a factor is R's own reference.

correct

Multiplicity correction across the contrasts: "holm" (default), "BH", "bonferroni", or "none". Recorded so the choice is visible rather than assumed.

Value

An object of class dmsa_contrasts: a list of dmsa_design objects in $designs, the data frame with the contrast columns added in $data, and the correction in $correct. Feed each design to dmsa_fit() and pass the resulting p-values to dmsa_contrast_adjust().

Examples

set.seed(1)
d <- data.frame(style = factor(sample(c("secure", "anxious", "avoidant"), 40,
                                      TRUE)),
                age = rnorm(40), cID = rep(1:20, each = 2))
des <- dmsa_design(focal = "style", fixed = "age", exchangeable = "cID")
dmsa_contrasts(d, des)
#> Warning: level(s) with fewer than 20 observations: anxious, avoidant, secure. A contrast on a thin cell is unstable and its permutation null is coarse.
#> DMSA categorical focal exposure: style
#>   levels:    anxious, avoidant, secure
#>   reference: anxious   (n = 15)
#>   2 directional contrast(s), corrected by holm:
#>     1. style__avoidant_vs_anxiousavoidant   (n = 10 vs 15)
#>     2. style__secure_vs_anxioussecure   (n = 15 vs 15)
#>   Each contrast keeps a sign. A single omnibus test over 2 df would not.