Invert Measure Values for Reverse Scoring — margot_invert

Inverts measure values using either z-score negation or ordinal scale inversion. This function is used internally by margot_causal_forest when flip_outcomes is specified.

Usage

margot_invert_measure(x, method = c("zscore", "ordinal"), scale_bounds = NULL)

Arguments

x: Numeric vector of values to invert
method: Character string specifying inversion method: - "zscore": Simple negation for already standardized data (default) - "ordinal": Invert on ordinal scale using bounds
scale_bounds: Numeric vector of length 2 specifying [min, max] bounds for ordinal scale inversion. Required when method = "ordinal". If NULL and method = "ordinal", bounds are inferred from data range.

Value

Numeric vector of inverted values

Details

Z-score method: Simply negates the values: x_flipped = -x This assumes the input data is already standardized (z-scores). For raw data that needs standardization, consider using scale() first.

Ordinal method: Uses the formula: x_flipped = (max + min) - x This preserves distances between values while reversing their order.

Examples

if (FALSE) { # \dontrun{
# z-score inversion (for already standardized data)
z_scores <- scale(c(1, 2, 3, 4, 5))[, 1] # standardize first
margot_invert_measure(z_scores, method = "zscore")

# ordinal scale inversion with known bounds
likert <- c(1, 2, 3, 4, 5, 3, 2)
margot_invert_measure(likert, method = "ordinal", scale_bounds = c(1, 5))

# ordinal scale inversion with inferred bounds
margot_invert_measure(likert, method = "ordinal")
} # }