Decoy Effect¶

Add a third option that is worse than one existing choice on every dimension but not strictly worse than the other, and preferences shift toward the option that dominates the decoy — even though the decoy itself is never chosen.

Core Idea¶

The decoy effect (also called the asymmetric-dominance effect or attraction effect) is the empirical finding that adding a third option to a choice set — where that option is asymmetrically dominated, meaning it is worse than one of the original options on every relevant dimension but not strictly worse than the other — shifts preferences among the two original options toward the one that dominates the decoy, in violation of the independence-of-irrelevant-alternatives axiom (IIA) that rational choice theory requires. Huber, Payne, and Puto (1982) documented it systematically across consumer-goods categories; the pattern has since replicated across product categories, restaurant menus, financial products, political-candidate presentations, and mate-choice scenarios. The mechanism is comparative: the dominated option is never chosen, but its presence establishes a local comparison standard that makes its dominating neighbor look superior by direct contrast along the dimensions where the domination is clearest. Human preferences, on this account, are not pre-formed and read off a stable utility function at the moment of choice; they are constructed during the choice process in light of the available comparison relations, with dominance serving as a salient organizing feature. The decoy's asymmetric position creates an asymmetric information structure — one pair of options admits a clear "winner" on all dimensions, the other pair does not — and this asymmetry is exploited by the comparison process so that the locally dominant option gains share even at the expense of an option it has not changed. The practical implication, exploited in pricing and menu design, is that the composition of a choice set is a substantive variable: which dominated options are present changes which of the remaining options is chosen, making the set itself a design lever distinct from the options it contains.

Structural Signature¶

Sig role-phrases:

the binary baseline set — two original options trading off on at least two attributes, neither dominating the other
the multi-attribute chooser — an agent who evaluates options on several dimensions, registers dominance, and is susceptible to contrast (the substrate; no comparison, no effect)
the asymmetrically-dominated decoy — a third option worse than one original on every relevant dimension but not strictly worse than the other, never itself chosen
the local comparison standard — the decoy installing a clear winner on one pair, an asymmetric information structure absent for the other pair
the contrast-driven flattery — the comparison process making the dominating neighbor look superior along the dimensions where the domination is clearest
the share-shift — preference moving toward the option that dominates the decoy, even though that option has not changed
the regularity/IIA violation — the dominating option's share going up when the set grows, contradicting independence of irrelevant alternatives and revealing preference as constructed, not read off a stable utility function
the two-set protocol — measure the binary split, add the dominated option, measure again; the magnitude of the move is the effect, attributable to set geometry not to the options

What It Is Not¶

Not an option meant to be chosen. The decoy is asymmetrically dominated and is essentially never selected; its function is to install a local comparison standard that flatters its dominating neighbor by contrast. A decoy that "sells well" has misunderstood the mechanism — share-shift, not decoy sales, is the whole point.
Not a change in the options' merits. The dominating option's attributes do not change; what moves its share is the arrangement of the set. Attributing the share-shift to the option getting better misses that a never-chosen alternative, by its mere presence, redistributes choice among options that are exactly as they were.
Not just "adding a third option" or more choices. The added option must be asymmetrically dominated — worse than one original on every relevant dimension but not strictly worse than the other — and the originals must trade off on at least two attributes. A generic third alternative, or options that do not trade off, provide no asymmetry for the effect to exploit; it is a specific dominance geometry, not set size.
Not evidence of stable underlying preferences. The effect is the clean violation of independence of irrelevant alternatives: it shows preference is constructed during choice in light of the comparisons the set makes available, not read off a fixed utility function. Reading the post-decoy choice as the chooser's "true" preference inverts what the effect demonstrates.
Not the compromise or similarity effect. Those are sibling context-dependent-preference effects that install different comparison relations — a middle option gaining share between extremes (compromise), a newcomer stealing disproportionately from its near-neighbor (similarity), an unavailable-but-tempting option (phantom). The decoy effect is specifically the asymmetric-dominance relation; the discriminating question is which relation the added option installs.
Not a property of any optimization or aggregation system. Adding an asymmetrically dominated alternative does not change a linear program's optimum, shift a chemical equilibrium, or move a market that aggregates without individual-level comparison. The effect requires a multi-attribute chooser that registers dominance and is susceptible to contrast; invoking it for a formal or physical system is a category error.

Scope of Application¶

The decoy effect lives in one domain — the choice behavior of multi-attribute, comparison-making choosers susceptible to contrast — and the contexts below are application settings of that single substrate, not structurally distinct systems. Its component parents (anchoring, framing/reference-point, preference construction) carry the cross-context lesson; it has no purchase on non-choosing systems (a dominated alternative does not move a linear program or a chemical equilibrium), which stay out of this map.

Behavioral-economics theory — the canonical regularity/IIA violation that motivates context-sensitive choice models (range-frequency, prospect-theory-adjacent accounts).
Marketing and pricing — the deliberately priced decoy SKU (the Economist subscription bundle, the medium popcorn) lifts a target option's share without the decoy itself selling.
Menu and choice-architecture design — restaurant tiers, software pricing, and policy-option slates exploit (or accidentally produce) the share-shift by placing dominated alternatives beside the intended choice.
Mate-choice and social judgment — a target paired with a less-attractive similar decoy is rated more attractive than when paired with a dissimilar comparison.
Animal choice (contested) — a small, mixed literature reports decoy-like shifts in hummingbird, honeybee, and frog choice tasks; if it holds, it widens the substrate to comparison-capable animals, but the evidence is weaker and the claim stays hedged.

Clarity¶

Naming the decoy effect forces apart two things rational-choice theory treats as one: the option set and the preferences defined over it. Under the independence-of-irrelevant-alternatives axiom the relative standing of two options is supposed to be a property of those options alone, invariant to what else sits beside them; the decoy effect is the clean counterexample that makes the axiom's failure unmistakable and so reclassifies set composition from an inert backdrop into a live causal variable. This is the conceptual hinge: once a never-chosen option can move share between the options that remain, "preference" can no longer be read as a value read off a stable utility function at the moment of choice. The concept reframes choice as evaluation of the comparisons the set makes available — preferences constructed in light of the dominance and contrast relations present — and that reframing is what turns the effect from a curiosity into a generative model with testable predictions about which additions move which choices.

The sharper distinction the label installs is between incidental and engineered set structure. A dominated alternative that happens to be on a menu and one a designer placed there for its dominance relation produce the same share-shift, so the practitioner learns to read a choice architecture not as a neutral list of what is available but as a configuration whose comparison structure is itself doing work. The diagnostic question becomes precise and answerable by construction: present the binary set and measure the split, add the asymmetrically dominated option and measure again, and the magnitude of the move is the effect — letting the analyst attribute a share change to the arrangement of the alternatives rather than to any change in the alternatives themselves, and audit a menu for dominance relations that are silently demoting or promoting options the designer never meant to favor.

Manages Complexity¶

Why share moves among options in a real choice set is otherwise an open-ended question — every menu, product line, listing, and ballot could shift demand for its own idiosyncratic reasons, and a context-sensitive theory of choice threatens an explosion of category-specific stories. The decoy effect compresses that by isolating one structural variable, asymmetric dominance, and one repeatable two-set protocol: measure the binary split, add the dominated option, measure again, and read the share-shift directly. The Economist-bundle case, the popcorn-sizing case, restaurant tiers, real-estate listings, candidate slates, and mate-choice photographs then cease to be separate findings and become instances of one comparison relation operating on whatever attributes are present, so the analyst tracks dominance structure rather than product content. It also collapses the regularity-violating share change to a single design lever distinct from the options themselves, and slots the related context-dependent-preference effects (compromise, similarity, phantom-decoy) into one family organized by which comparison relation the added option installs. A designer or modeler thereby reasons from a low-dimensional handle — what dominance relations does this set make available, and toward which neighbor do they tilt — instead of re-deriving demand for each arrangement, reading the qualitative direction of the shift off the geometry of the set.

Abstract Reasoning¶

The decoy effect licenses moves organized around set composition as a causal variable and asymmetric dominance as the lever. Diagnostic: confronted with a share-shift between two options that did not themselves change, the analyst infers that the arrangement of the set, not the merits of the alternatives, moved the choice — and looks for the dominance relation responsible: a third option worse than one original on every relevant dimension but not strictly worse than the other, installing a local comparison standard that flatters its dominating neighbor. The signature read is a regularity violation: the dominating option's share goes up when the set grows, the opposite of what adding an alternative should do, and that upward move is the fingerprint of an asymmetrically dominated decoy doing its work along the dimensions where the domination is clearest. A second diagnostic distinguishes incidental from engineered structure: because a decoy that happens to be on a menu and one a designer placed there produce the same share-shift, the analyst audits a real choice set for dominance relations that may be silently promoting or demoting options no one meant to favor.

Interventionist: the corrective and exploitative moves are predictions tied to a clean two-set protocol. To raise a target option's share, add an alternative asymmetrically dominated by it — predicting the share moves toward the dominating neighbor even though the decoy itself is never chosen; to undo an accidental distortion, remove the dominated alternative — predicting the share reverts. The measurement that grounds these is constructive: present the binary set and record the split, add the asymmetrically dominated option and record again, and the magnitude of the move is the effect, attributable to the set's geometry rather than to any change in the options. The deeper interventionist commitment is that "preference" here is not a fixed quantity to be elicited but something constructed during choice in light of the comparisons the set makes available — so to shift the outcome one engineers the comparison structure (which dominance relations are present, toward which neighbor they tilt) rather than trying to change what the chooser antecedently wants.

Boundary-drawing: the concept fixes which comparison relation it covers and separates it from its family. It is specifically the asymmetric-dominance (attraction) relation — a decoy dominated by one option and tilting share toward it — and is the wrong diagnosis for its siblings, which install different relations: the compromise effect (a middle option gaining share when extremes are added), the similarity effect (a new option stealing disproportionately from its near-neighbor), the phantom decoy (an unavailable-but-tempting option). The discriminating question is which comparison relation the added option installs, and that classification slots any given share-shift into the right cell of the context-dependent-preference family. The boundary also bounds where the effect lives at all: it requires a chooser evaluating options on multiple attributes who registers dominance and is susceptible to contrast, so a set with options that do not trade off on at least two dimensions provides no asymmetry for the decoy to exploit. Predictive: given a contemplated addition to a choice set, the analyst forecasts the qualitative direction of the share-shift from the geometry alone — toward whichever existing option dominates the newcomer — before any choice data are collected, reading the tilt off the dominance structure rather than re-deriving demand for the arrangement.

Knowledge Transfer¶

Within behavioral economics and consumer-choice research the effect transfers as mechanism, intact, with the caveat that its "domains" are application contexts of one substrate — a multi-attribute comparison-making chooser susceptible to contrast — not structurally distinct systems. With that understood, the set-composition-as-causal-variable framing, the asymmetric-dominance lever, and the clean two-set protocol (measure the binary split, add the dominated option, measure again) carry without translation across behavioral-economics theory (the canonical regularity/IIA violation motivating context-sensitive choice models), marketing and pricing (the deliberately priced decoy SKU — the Economist bundle, the medium popcorn), menu and choice-architecture design (restaurant tiers, software pricing, policy-option slates), and mate-choice and social judgment (a target paired with a less-attractive similar decoy rated higher). The vocabulary (asymmetric dominance, attraction effect, regularity violation, constructed preference), the diagnostic (a share-shift toward an unchanged option signals the set's geometry, not the merits; audit for dominance relations silently promoting options), and the interventions (add a decoy dominated by the target to lift its share; remove an accidental decoy to revert) all move freely, because the same comparison-and-contrast machinery is operative in each.

Beyond a comparison-making chooser the effect does not travel — and the boundary is sharp. The decoy effect requires an agent that evaluates options on multiple attributes, registers dominance, and is susceptible to contrast; in systems that lack such cognition the analogue simply does not exist — adding an asymmetrically dominated alternative does not change a linear program's optimum, shift a chemical equilibrium, or move a market that aggregates without individual-level comparison. So invoking the decoy effect for a physical or formal optimization system is a category error, not a stretched analogy. The one genuine widening of the substrate is within comparison-making cognition: a small and contested literature reports decoy-like shifts in hummingbird, honeybee, and frog choice tasks, which — if they hold — would make this a shared abstract mechanism across comparison-capable animals, not a human-only effect (the evidence is mixed and the effects, where present, smaller and less reliable, so the claim must stay hedged). What otherwise carries is not the named effect but the more-general parents it composes: anchoring (the dominated option as comparison standard), framing and reference-point shifting (dominance turned into a salient feature, the comparison point moved), and the broader preference construction / context-dependent preference pattern. The decoy effect is best read as a named recipe for triggering that composite under a specific dominance relation — and as one row in a context-dependent-preference family (the compromise, similarity, and phantom-decoy effects install different comparison relations). When the cross-context lesson is wanted, carry those parents (and the family), while the decoy effect's own cargo (asymmetric dominance specifically, the SKU/menu application set, the share-shift protocol) stays bound to multi-attribute choosers. So: as mechanism it stays inside comparison-making choice (possibly across some animals); its components travel as the real structure; and it has no purchase on non-choosing systems (see Structural Core vs. Domain Accent).

Examples¶

Canonical¶

Huber, Payne, and Puto (Journal of Consumer Research, 1982) gave the effect its seminal demonstration and its name, "asymmetric dominance." Participants chose between two products described on two attributes that traded off — for instance a beer that was cheaper but rated lower in quality versus one pricier but higher in quality — so that neither option dominated the other. When a third option was added that one target dominated (worse on both dimensions, or worse on one and tied on the other) while the competitor did not dominate it, the dominated target's choice share rose relative to the two-option baseline, contradicting the regularity/IIA requirement that adding an alternative cannot raise a rival's share. The decoy itself was almost never chosen; its whole function was to serve as a local comparison standard.

Mapped back: The two trade-off products are the binary baseline set; the added worse-on-all option is the asymmetrically-dominated decoy, which installs the local comparison standard that flatters its neighbor. The rise in the target's share is the share-shift, and its going up when the set grows is the diagnostic regularity/IIA violation. The paired conditions — measure the binary split, add the decoy, measure again — are the two-set protocol that isolates set geometry from the options' merits.

Applied / In Practice¶

The Economist's subscription pricing is the canonical field case, popularized by Dan Ariely in Predictably Irrational (2008). The offer listed web-only at $59, print-only at $125, and print-plus-web at $125 — where print-only is asymmetrically dominated by the bundle (identical price, strictly less content). Presenting all three options to 100 MIT Sloan students, Ariely recorded 16 choosing web-only, 0 choosing print-only, and 84 choosing print-plus-web. Removing the dominated print-only decoy and re-running the choice, 68 chose web-only and only 32 chose the bundle. The decoy sold to nobody, yet its mere presence moved a majority toward the premium bundle — a set-composition lever in live pricing.

Mapped back: Print-only is the asymmetrically-dominated decoy; against it, print-plus-web becomes an unmistakable winner, the local comparison standard producing contrast-driven flattery. The bundle's share moving from 32 to 84 out of 100 is the share-shift, and its rising as the set enlarges is the regularity/IIA violation. The two presentation conditions are the two-set protocol, attributing the swing to the arrangement rather than to any change in the bundle itself.

Structural Tensions¶

T1: Engineered lever versus incidental distortion (the same share-shift, intended and accidental). The entry stresses that a dominated alternative a designer placed for its dominance relation and one that merely happens to sit on a menu produce the identical share-shift. This makes the effect double-edged as a practical object: it is a deliberate pricing and menu-design lever (add a decoy dominated by the target to lift its share) and, in the same stroke, a silent hazard — a real choice set may harbor dominance relations promoting or demoting options no one meant to favor. The tension is that the mechanism does not distinguish intent; it fires on geometry alone. So the same audit that lets a designer weaponize a decoy is the one a designer must run to catch a decoy already sabotaging the set, and neither the effect nor its measurement tells you which case you are in. Diagnostic: Is the dominance relation in this set one someone placed to tilt the choice, or one that arose incidentally and is now moving share no one intended?

T2: Constructed preference versus a "true" preference to distort (which choice is the real one?). The decoy effect is the clean IIA violation showing preference is constructed during choice in light of the comparisons the set makes available, not read off a stable utility function. But this undercuts the very baseline the protocol relies on: if there is no antecedent preference, then the binary split is not more "true" than the post-decoy split — it is just the choice the two-option set constructs. The entry warns that reading the post-decoy choice as the chooser's real preference inverts the lesson; yet symmetrically, treating the pre-decoy choice as the undistorted truth smuggles back the stable utility the effect denies. The tension is that calling the decoy a "manipulation" presupposes a genuine preference being corrupted, while the effect's own thesis is that preference has no set-independent form to corrupt. Diagnostic: Does the analysis treat the binary-set choice as the chooser's real preference (assuming a stable utility the effect denies), or accept that every set — decoy or not — constructs a preference with no privileged baseline?

T3: A sharp category boundary versus a contested lower edge (how far down does a chooser go?). The effect draws a hard line: it requires an agent evaluating options on multiple attributes that registers dominance and is susceptible to contrast, so invoking it for a linear program's optimum, a chemical equilibrium, or a market that aggregates without individual comparison is a category error, not a stretched analogy — no comparison, no effect. Yet the lower edge of "chooser" is genuinely unsettled: a small, mixed literature reports decoy-like shifts in hummingbird, honeybee, and frog tasks, which if real would widen the substrate to comparison-capable animals. The tension is that the boundary is razor-sharp on one side (no purchase whatever on non-choosing systems) and blurry on the other (where exactly does comparison-with-contrast begin among animals?), so the same concept is both strictly bounded and evidentially open at its frontier. Diagnostic: Does the system in question actually perform multi-attribute comparison with susceptibility to contrast, or is a decoy being posited for a process that only optimizes or aggregates — and if it is an animal, is the evidence for contrast-driven choice there or merely assumed?

T4: Asymmetric dominance versus its sibling relations (one family, several levers). The decoy effect is specifically the asymmetric-dominance (attraction) relation, and it sits in a family whose members install different comparison relations: the compromise effect (a middle option gaining share between extremes), the similarity effect (a newcomer stealing disproportionately from its near-neighbor), the phantom decoy (an unavailable-but-tempting option). A given real-world share-shift is an observed outcome, but the relation that produced it is not read off the outcome — the same rise in a target's share could come from dominance, compromise, or similarity structure. The tension is that the decoy effect's precise diagnosis and its precise lever (add a decoy dominated by the target) are only correct once the relation is correctly classified, and misreading a compromise or similarity shift as a decoy shift applies a lever that will not reproduce the effect. Diagnostic: Which comparison relation did the added option install — asymmetric dominance, compromise, similarity, or phantom — and is the intended lever matched to that relation rather than to the bare fact of a share-shift?

T5: Autonomy versus reduction (a named recipe or the parents it composes). "Decoy effect" is a named, canonically studied finding with proprietary cargo — asymmetric dominance specifically, the SKU and menu application set, the Economist bundle and the two-set share-shift protocol — that stays bound to multi-attribute choosers. Yet the portable structure is not proprietary: what carries across contexts is the more-general parents the effect composes — anchoring (the dominated option as comparison standard), framing and reference-point shifting (dominance made salient, the comparison point moved), and the broad preference-construction / context-dependent-preference pattern — plus its place as one row in a family (compromise, similarity, phantom). The tension is between a standalone effect that earns its own experiments and pricing playbook and the recognition that its cross-context lesson belongs to those composed parents, with the decoy effect best read as a named recipe for triggering the composite under one specific dominance relation. Diagnostic: Resolve toward the parents (anchoring, framing/reference-point, preference construction) when carrying the lesson to a new comparison setting; toward the named decoy effect when diagnosing or engineering an asymmetric-dominance share-shift in a specific choice set in situ.

Structural–Framed Character¶

The decoy effect sits toward the structural end — best read as mixed-structural, though a less pure-nature case than isostasy: a genuine, evaluatively neutral comparison-and-contrast mechanism that operates automatically in choosers (possibly some animals), kept domain-specific by its choice-behavior vocabulary and its being a composed recipe rather than a single prime.

On evaluative weight it is neutral: the effect describes how set composition redistributes choice, convicting no one — even the "IIA violation" framing is a descriptive contrast with a normative axiom, not a verdict that the chooser is foolish. Unlike declinism's mild "distortion" charge, the decoy effect renders no epistemic indictment. On institutional_origin it is essentially none: the share-shift is a feature of how a multi-attribute chooser registers dominance and contrast, not an artifact of any survey, agency, or convention — the name is a research label, but the mechanism is natural choice cognition. On human_practice_bound it patterns partway structural: the effect operates automatically in the chooser, not as a constituted practice performed on someone, and — decisively — the entry establishes it is not human-only; the contested hummingbird/honeybee/frog literature would, if it holds, widen the substrate to comparison-capable animals, and the machinery runs observer-free in whatever agent registers the comparison. Yet it is strictly substrate-bound: it has "no purchase on non-choosing systems," so adding a dominated alternative moves no linear program or chemical equilibrium — invoking it there is a category error, not even analogy.

Two things keep it off the structural pole and mark it domain-specific. First, vocab_travels fails: asymmetric dominance, the attraction effect, the regularity/IIA violation, the SKU-and-menu application set, and the share-shift protocol are pinned to the choice-behavior substrate and rename off it. Second — and unusually — it has no single portable structural skeleton to instantiate: the entry establishes it is a named recipe for triggering a composite under one specific dominance relation, assembled from parents that each travel on their own — anchoring (the dominated option as comparison standard), framing/reference-point shifting (dominance made salient, the comparison point moved), and the broad preference-construction / context-dependent-preference pattern — and sitting as one row in a family alongside the compromise, similarity, and phantom effects, which install different comparison relations. So the portable content belongs to those composed parents and the family, carried individually and recognized across comparison-making agents; what is distinctive to the decoy effect — the asymmetric-dominance relation specifically, the pricing/menu application set, the two-set share-shift protocol — stays bound to multi-attribute choosers. Its character: structural in being a real, evaluatively neutral, no-institutional-origin comparison mechanism recognized across choosers and possibly animals — but expressed in choice-behavior vocabulary and constituted as a composed recipe of anchoring/framing/preference-construction rather than a single prime, leaving it mixed-structural, with zero reach into non-choosing systems.

Structural Core vs. Domain Accent¶

This section decides why the decoy effect is a domain-specific abstraction and not a prime, and it carries the case for its domain-specificity — there is no separate section for that. Like declinism, the decoy effect's portable content is not one skeleton but a composite of parents, and it does not reach non-choosing substrates at all.

What is skeletal (could lift toward a cross-domain prime). Strip the pricing and menu vocabulary and what survives is not a single relational structure but a composite of thin ones: a comparison standard installed by one element that reshapes evaluation of another (a context point that anchors), a salient relation (here, dominance) made into the organizing feature that tilts the judgment (a reframing / reference-point move), and the broad fact that the evaluation is constructed from the available comparisons rather than read off a fixed scale (preference construction). The genuinely skeletal move — that an added element which is never itself selected redistributes evaluation among unchanged elements by altering the local comparison structure — is what the effect shares with those parents, which is exactly why it instantiates anchoring (the dominated option as comparison standard) and framing/reference-point shifting (dominance made salient, the comparison point moved), plus the broader preference-construction / context-dependent-preference pattern. This is the portable core it composes from, not what makes the decoy effect distinctive — and even that composite runs only within an agent that compares on multiple attributes and is susceptible to contrast.

What is domain-bound. What is distinctive to the decoy effect in particular stays bound to multi-attribute choosers. The asymmetric-dominance relation specifically (worse than one original on every dimension but not strictly worse than the other); the requirement of a binary baseline set trading off on at least two attributes; the two-set share-shift protocol (measure the split, add the dominated option, measure again) that attributes the move to set geometry; the regularity/IIA violation framing against rational-choice theory; and the worked application set — the priced decoy SKU, the Economist bundle, the medium popcorn, restaurant tiers, mate-choice photographs — are the operative vocabulary, instruments, and cases the phenomenon actually studies. The decisive test: remove the multi-attribute comparison-making chooser and there is no effect at all. Adding an asymmetrically dominated alternative does not change a linear program's optimum, shift a chemical equilibrium, or move a market that aggregates without individual comparison — so invoking the decoy effect there is a category error, not even a stretched analogy. Unlike most domain-specific entries, it does not reach other substrates loosely; off the choosing substrate there is simply no phenomenon.

Why this does not clear the prime bar. A prime's vocabulary travels and its transfer is recognition of the same mechanism, not analogy. The decoy effect's transfer is bimodal in an atypical way. Within choice cognition it travels intact as mechanism across contexts — behavioral-economics theory, marketing and pricing, menu design, social judgment, and (contested) some animals — but these are application settings of one substrate, one comparison-and-contrast machine recognized repeatedly, not a mechanism spanning structurally distinct systems. Beyond a comparison-making chooser it does not travel at all. And when the portable lesson is wanted, it belongs to the parents the effect composes — anchoring, framing/reference-point shifting, and preference construction, each of which travels on its own — together with its place as one row in a context-dependent-preference family (compromise, similarity, phantom decoy) whose siblings install different comparison relations. The cross-context carriers are those composed parents and the family; the decoy effect, as named, is a named recipe for triggering the composite under one specific dominance relation, whose distinctive cargo lifts to no single umbrella and generalizes to no non-choosing substrate — which is precisely why it remains a domain-specific abstraction rather than a prime.

Relationships to Other Abstractions¶

Current abstraction Decoy Effect Domain-specific

Parents (3) — more general patterns this builds on

Decoy Effect is a kind of Context-Dependent Preference Prime

Decoy Effect is the asymmetric-dominance menu species in which an unchosen inferior alternative shifts preference between two unchanged focal options.
Decoy Effect is a decomposition of Comparison Prime

The decoy works only by placing multi-attribute options in a shared frame where pairwise dominance relations can be read and one local winner becomes salient.
Decoy Effect is a decomposition of Preference Prime

The measurable payload is a change in an agent's ordering over an unchanged focal choice set, expressed as the dominating option gaining share after the never-chosen decoy is added.

Hierarchy paths (3) — routes to 2 parentless roots

Decoy Effect → Context-Dependent Preference → Preference

Show alternative paths (2)

Not to Be Confused With¶

Compromise effect. The sibling context-dependent-preference effect where a middle option gains share simply by sitting between two extremes (looking like the safe, moderate pick). The decoy effect turns on asymmetric dominance — a decoy worse than one option on every dimension — not on middle-ness. Tell: does the added option make a third option look like a moderate compromise (compromise effect) or make one existing option look like an all-dimensions winner over a dominated newcomer (decoy effect)? Different comparison relation installed.
Similarity effect. The sibling effect where a new option steals share disproportionately from its nearest neighbor (the similar option), redistributing among close substitutes. The decoy effect instead lifts the neighbor that dominates the decoy. Tell: does the newcomer cannibalize the option most like it (similarity effect) or boost the option that dominates it (decoy effect)? Steal-from versus flatter-by-contrast.
Phantom decoy. The sibling effect where an especially attractive option is shown but is unavailable (out of stock, sold out), pulling preference toward the available option most like it. Unlike the phantom, the decoy effect's third option is fully available and real — it simply is never chosen because it is dominated. Tell: is the influential option present-but-unattainable (phantom) or present, attainable, yet dominated and unchosen (decoy)?
Bait-and-switch. A deceptive sales tactic: advertise an attractive option, then deny or withdraw it to push a costlier one. The decoy effect deceives no one and withdraws nothing — every option, including the decoy, remains genuinely available; the shift is a real comparison effect on unchanged options. Tell: is the advertised option made unavailable to redirect the buyer (bait-and-switch, deception) or left fully available while a dominated peer shifts share honestly (decoy effect)?
Loss leader. A pricing tactic where an item is sold cheaply (even at a loss) to be bought, drawing traffic that buys other things. The decoy is the opposite: priced so it is never bought, its whole function being to shift share to a neighbor by contrast. Tell: is the low-value option meant to sell and draw customers (loss leader) or to go unchosen while flattering another option (decoy)? A decoy that "sells well" has misfired.
The anchoring / framing / preference-construction parents (umbrella). The substrate-neutral primes the effect composes — anchoring (the decoy as comparison standard), framing/reference-point shifting (dominance made salient), and preference construction (choice built from available comparisons). Not confusable peers but the generalization the cross-context lesson rides on. Tell: the parents (and the compromise/similarity/phantom family) carry the lesson to any comparison setting; "decoy effect," treated more fully in a later section, is the named recipe for the asymmetric-dominance relation specifically.

Neighborhood in Abstraction Space¶

Decoy Effect sits in a crowded region of the domain-specific corpus (23^rd percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Choice Paradoxes & Collective Decision-Making (14 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-07-12