Lexis ratio¶

A dispersion ratio comparing observed variation in grouped binomial proportions with the variation expected under one common success probability.

Version: v2 · 2026-09-19 · History
Domain-specific #: 5313
Domain group: Formal Sciences
Origin domain: Experimental Design & Statistics
Subdomain: Historical Statistics → Experimental Design & Statistics

Core Idea¶

The statistic has several finite-sample conventions, group sizes affect expected variance, values above or below one indicate over- or underdispersion rather than directly identifying causes and modern homogeneity tests often replace it. Success proportions are computed for repeated groups, their between-group dispersion is standardized by the binomial sampling variance implied by a pooled probability, revealing extra heterogeneity among trial mechanisms. The abstraction is therefore identified by a declared carrier, a transformation or constraint over that carrier, and an invariant that tells an analyst whether the named structure is genuinely present.

How would you explain it like I'm…

More-Than-Luck Checker

If lots of kids each flip a coin ten times, their counts of heads will differ a little, just from luck. The Lexis ratio checks whether the groups differ more than luck alone would make them. If they do, something may be different about the groups, but the number doesn't say what.

The Spread-Versus-Luck Ratio

Imagine several groups each flip coins many times and count heads. If every flip really had the same chance, the groups' results would still differ a little just by luck. The Lexis ratio compares how much the groups actually differ to how much luck alone would make them differ. A value near 1 matches luck. Above 1 means the groups differ more than luck predicts, and below 1 means they differ less. The number tells you that something unusual is going on, but not what is causing it.

Binomial Dispersion Ratio

The Lexis ratio is a statistic for repeated groups of yes/no trials. You compute each group's success proportion and measure how much those proportions vary from group to group. Then you compare that to the variance you'd expect if every group shared one common success probability, estimated by pooling all the data. The ratio of observed variance to expected binomial variance is near one if the groups are consistent with a common probability. Above one signals overdispersion, meaning extra variation, and below one signals underdispersion. It doesn't identify the cause, such as differences between groups or dependence among trials. Group sizes affect the expected variance, and today chi-square homogeneity tests often do the same job.

The Lexis ratio is a dispersion statistic for grouped binary trials. For each group, record the size and successes and compute the sample proportion. Estimate a pooled success probability under the null model that all trials share a common probability. The observed between-group variance of the proportions is then standardized by the binomial sampling variance implied by that pooled estimate. Values near one are consistent with homogeneity. Values above one indicate overdispersion, suggesting heterogeneity among trial mechanisms or dependence among trials. Values below one indicate underdispersion. The ratio diagnoses over- or underdispersion rather than identifying its cause. Several finite-sample corrections exist, unequal group sizes affect the expected variance, and the statistic is closely related to, and often replaced by, chi-square tests of homogeneity.

Scope of Application¶

Lexis ratio belongs to historical statistics and is useful where the analyst can specify the typed historical statistics carrier, including objects, relations, parameters, conventions, evidence, boundaries, and comparison targets, then evaluate the grouped binary trials and group sizes, successes and sample proportions, common-probability null model, pooled success estimate, observed between-group variance, expected binomial variance, ratio and finite-sample correction, interpretation around one, heterogeneity and dependence alternatives and relationship to chi-square homogeneity testing are explicit.

Clarity¶

The abstraction clarifies a crowded vocabulary by making the grouped binary trials and group sizes, successes and sample proportions, common-probability null model, pooled success estimate, observed between-group variance, expected binomial variance, ratio and finite-sample correction, interpretation around one, heterogeneity and dependence alternatives and relationship to chi-square homogeneity testing are explicit the center of the account. A claim should name the carrier, the governing operation or relation, the applicable assumptions, and the recognition test.

Manages Complexity¶

Without the abstraction, an analyst must reason directly over many local details: the carrier roles, admissibility assumptions, competing conventions, derived invariants, boundary cases, and proof or validation obligations specific to Lexis ratio. Lexis ratio compresses them into the roles in the structural signature. That compression permits comparison across instances without erasing the variables that determine validity. It also exposes which details may be varied safely and which are constitutive.

Abstract Reasoning¶

Identify the carrier. State what the elements, states, objects, or observations are: the typed historical statistics carrier, including objects, relations, parameters, conventions, evidence, boundaries, and comparison targets. Reject examples whose alleged carrier belongs to a different problem.
Lock the constitutive rule. Express the grouped binary trials and group sizes, successes and sample proportions, common-probability null model, pooled success estimate, observed between-group variance, expected binomial variance, ratio and finite-sample correction, interpretation around one, heterogeneity and dependence alternatives and relationship to chi-square homogeneity testing are explicit independently of one notation or implementation.

Knowledge Transfer¶

Knowledge transfers strongly among subfields of historical statistics because they reuse the typed historical statistics carrier, including objects, relations, parameters, conventions, evidence, boundaries, and comparison targets, Success proportions are computed for repeated groups, their between-group dispersion is standardized by the binomial sampling variance implied by a pooled probability, revealing extra heterogeneity among trial mechanisms., and type the carrier, state every parameter and convention in the definition, test that the grouped binary trials and group sizes, successes and sample proportions, common-probability null model, pooled success estimate, observed between-group variance, expected binomial variance, ratio and finite-sample correction, interpretation around one, heterogeneity and dependence alternatives and relationship to chi-square homogeneity testing are explicit, compare the nearest accepted identity, and report counterexamples, uncertainty, and limiting cases.

Relationships to Other Abstractions¶

Current abstraction Lexis ratio Domain-specific

Parents (1) — outgoing typed relations

Lexis ratio is a kind of Measurement Prime

It remains its own entry because its identity is fixed by the grouped binary trials and group sizes, successes and sample proportions, common-probability null model, pooled success estimate, observed between-group variance, expected binomial variance, ratio and finite-sample correction, interpretation around one, heterogeneity and dependence alternatives and relationship to chi-square homogeneity testing.

Hierarchy path (1) — routes to 1 parentless root

Lexis ratio → is a kind of Measurement

Neighborhood in Abstraction Space¶

Lexis ratio sits in a crowded region of the domain-specific corpus (26^th percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Dispersion Measures & Multilevel Regression (11 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25