Feature score from deviation contrasts (random-effects meta per condition)

For each condition's deviation contrast, meta-analyzes the per-variant absolute effect sizes (random-effects, via metafor::rma) and returns the pooled Z-score (\(\hat\mu / \mathrm{se}\)). One score per condition – the "meta" feature score of mash_posterior.ipynb.

Usage

calculateFeatureScores(contrastResult, metaMethod = "REML")

Arguments

contrastResult

A contrast table from mashPosteriorContrast (variants x contrasts) carrying mean_contrast_*_deviation and se_contrast_*_deviation columns.

metaMethod

Between-study variance estimator forwarded to metafor::rma (default "REML").

Value

A data.frame with condition and zScore.

See also

nSignificantScore, scoreFromCs