Use to justify counterbalancing, not to predict an exact effect in a new trial.
Evidence captured during full-source processing of the supplied file.
The authors estimate that serial position by itself accounted for roughly four percent of preference variation in their multilevel analysis.
12
12
Discussion and limitations
Paraphrase checked against local PDF sheet 12 on 2026-09-05; not presented as a verbatim quotation.
Whiskey Knowledge Databases › Evidence Excerpts
Verified Paraphrase checked against the complete supplied source.
Evidence
The authors estimate that serial position by itself accounted for roughly four percent of preference variation in their multilevel analysis.
Locator and context
Locator: Page 12–12
Source section: Discussion and limitations
Context: Evidence captured during full-source processing of the supplied file.
Interpretation and caution
Use to justify counterbalancing, not to predict an exact effect in a new trial.
Verification
Paraphrase checked against local PDF sheet 12 on 2026-09-05; not presented as a verbatim quotation.
Rechecked Interpretation
Main p.10 explains the approximately four-percent figure as the change in model R² when serial position was omitted; p.12 summarizes it. It is not four percent of each person's preference or a causal share. Group mean ratings are the outcome, and the observational model retains unmeasured confounding. Both tables and figures on pp.9–10 were visually checked September 27, 2026.