Table · dataset · 2026

<p>Overall metrics across the four text-containing configurations on the 25-fold OOF evaluation (, 29 participants). This table documents the <i>aggregate-level equivalence</i> of the four text-containing configurations on every overall metric—the empirical basis for the subgroup-conditional hypothesis tested in Table 3; it is not intended to establish a within-family ordering. Confidence intervals (in brackets) were obtained by resampling <i>participants</i> (not individual clips) with replacement, so that the uncertainty estimate reflects between-participant variability rather than the larger but non-independent clip count^a.</p>

Listed in figshare

<p>Overall metrics across the four text-containing configurations on the 25-fold OOF evaluation (, 29 participants).