Interaction (statistics)¶
A model relation in which the association or effect of one predictor on an outcome changes with the level of another predictor.
Core Idea¶
Statistical interaction is scale- and parameterization-dependent, it need not imply biological or causal interaction, main-effect coefficients become conditional and absence on an additive scale does not imply absence on multiplicative scales. A product term or stratified response surface lets the slope for one variable vary with another; contrasting predicted outcomes tests departure from the chosen additive or multiplicative no-interaction model. 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.
Scope of Application¶
Interaction (statistics) belongs to statistical modeling and is useful where the analyst can specify the typed statistical modeling carrier, including objects, relations, parameters, conventions, evidence, boundaries, and comparison targets, then evaluate the outcome and two or more predictors, model family and link scale, main effects and interaction term or nonparametric surface, reference levels and coding, conditional marginal effect and contrast, null of additivity on the chosen scale, estimation uncertainty and visualization, causal assumptions when effect modification is claimed, hierarchy or marginality and distinction from confounding mediation and simple correlation are explicit.
Clarity¶
The abstraction clarifies a crowded vocabulary by making the outcome and two or more predictors, model family and link scale, main effects and interaction term or nonparametric surface, reference levels and coding, conditional marginal effect and contrast, null of additivity on the chosen scale, estimation uncertainty and visualization, causal assumptions when effect modification is claimed, hierarchy or marginality and distinction from confounding mediation and simple correlation are explicit the center of the account.
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 Interaction (statistics). Interaction (statistics) 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 statistical modeling carrier, including objects, relations, parameters, conventions, evidence, boundaries, and comparison targets. Reject examples whose alleged carrier belongs to a different problem. 2. Lock the constitutive rule. Express the outcome and two or more predictors, model family and link scale, main effects and interaction term or nonparametric surface, reference levels and coding, conditional marginal effect and contrast, null of additivity on the chosen scale, estimation uncertainty and visualization, causal assumptions when effect modification is claimed, hierarchy or marginality and distinction from confounding mediation and simple correlation are explicit independently of one notation or implementation.
Knowledge Transfer¶
Knowledge transfers strongly among subfields of statistical modeling because they reuse the typed statistical modeling carrier, including objects, relations, parameters, conventions, evidence, boundaries, and comparison targets, A product term or stratified response surface lets the slope for one variable vary with another; contrasting predicted outcomes tests departure from the chosen additive or multiplicative no-interaction model., and type the carrier, state every parameter and convention in the definition, test that the outcome and two or more predictors, model family and link scale, main effects and interaction term or nonparametric surface, reference levels and coding, conditional marginal effect and contrast, null of additivity on the chosen scale, estimation uncertainty and visualization, causal assumptions when effect modification is claimed, hierarchy or marginality and distinction from confounding mediation and simple correlation are explicit, compare the nearest accepted identity, and report counterexamples, uncertainty, and limiting cases.
Relationships to Other Abstractions¶
Current abstraction Interaction (statistics) Domain-specific
Parents (1) — more general patterns this builds on
-
Interaction (statistics) is a kind of Relation Prime
The proposed strict upward parent is
prime:relation.
Hierarchy path (1) — routes to 1 parentless root
- Interaction (statistics) → Relation
Neighborhood in Abstraction Space¶
Interaction (statistics) sits in a crowded region of the domain-specific corpus (20th percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.
Family — Regression, Genetics & Interaction Models (10 abstractions)
Nearest neighbors
- Deviance (statistics) — 0.92
- Principle of marginality — 0.92
- Standard error — 0.91
- Hierarchical generalized linear model — 0.91
- Studentization — 0.91
Computed from structural-signature embeddings · 2026-09-08