Capstractstatcapstract:2611.00033

Statistics > Methodology

Title:Lower bounds on displacement under randomized foliation

Abstract:

A change of numbering can alter the distribution of selected positions even when the proportion of selections at a preferred numeral is stable. We ask how many individual choices must move to account for this change. In a parallel experiment the marginal distributions of position are identifiable, while the pairing of choices under the two numbering schemes is unknown. We give a finite transport formulation for the minimum fraction displaced by more than a specified number of leaves. A pilot sample selects a set of positions on which two independent binomial limits give a lower confidence bound. In an experiment on the opening at which an unused specimen ledger should be displayed, 160 pilot and 800 confirmation participants each made one selection. Moving folio 50 from the eighth to the third leaf gave a confirmation contrast of 23.75 percentage points on the pilot set. The one-sided 95% lower limit for the proportion of selections that would change leaf was 14.9%. The margins admitted a pairing with every displacement at most three leaves, although folio 50 itself had moved five.

Comments:11 pages, 3 figures, 3 tables. Includes the complete cell counts.
Subjects:Methodology (stat.ME); Applications (stat.AP)
Cite as:capstract:2611.00033 [stat.ME]
 (or capstract:2611.00033v1 [stat.ME] for this version)

Submission history

From: Elina Meravuori
[v1] Mon, 16 Nov 2026 08:14:45 UTC (202 KB)

Current browse context:

stat.ME
Change to browse by:

Export citation