RESEARCH DOSSIER
Consumer Attention, Information Search, and Price Salience
Abstract. Consumer attention, information search, and price salience describe how buyers encounter, select, acquire, integrate, and act upon market information under limited cognitive resources and costly search. This fundamental research monograph synthesizes classic theories of information economics and adaptive choice, direct observation of eye movements and retail search, randomized field evidence on tax-inclusive price tags, experiments on partitioned prices and shrouded fees, and recent digital-commerce studies through 2026. It distinguishes exposure, attention, comprehension, attribute weighting, and final purchase; evaluates competing accounts of salience, rational inattention, and search stopping; and separates causal effects of display interventions from observational ranking effects and market equilibrium responses. Experimental and methodological discussions cover eye tracking, clickstream data, incentivized choice, structural search models, field randomization, welfare identification, privacy, heterogeneous consumers, mobile search aids, and human–artificial decision support. The synthesis identifies open problems concerning mechanism discrimination, longitudinal learning, portability across interfaces, and informed consumer autonomy. This is a critical narrative synthesis of identifiable published scientific literature, not an original experiment, registered systematic review, or meta-analysis.
Keywords: consumer attention; information search; consumer search costs; price salience; tax salience; partitioned pricing; hidden fees; eye tracking; clickstream; rational inattention; consideration sets; consumer decision-making; price transparency; digital shopping; human–artificial search
The Research Object and Its Explanatory Boundaries
Consumer attention, information search, and price salience describe a connected sequence of operations in market choice: consumers encounter options, allocate scarce perceptual and cognitive resources, acquire some information rather than other information, interpret the cost components that become available, and decide when further investigation is worth its expense. Attention denotes differential allocation of processing to attributes, alternatives, or cues. Search denotes actions that obtain, reveal, compare, or verify information before commitment. Price salience denotes the probability and weight with which a monetary cost enters evaluation at a particular decision point. These constructs can be measured separately even though they interact. A shopper may see a label without remembering it, remember a price without using it to compare options, or compare prices while overlooking a charge added later. The research object is their causal and functional integration in consumer behavior Bettman et al. (1998) Reutskaja et al. (2011).
The inquiry centers on prechoice search, screening of alternatives, the visibility of prices and fees, and the acquisition of multiattribute information. Numerical price encoding, reference-price formation, willingness to pay, and left-digit effects form a neighboring topic, CEC-04. Choice-set expansion and assortment overload belong principally to CEC-05; ranking algorithms and recommender governance require their own dossier, CEC-23. These boundaries preserve the significance of interactions. A reference price can affect whether a displayed amount attracts attention, and an algorithm may order which amounts a shopper encounters, yet the present explanatory task is to identify what is inspected and how the inspection path changes choice. The relevant outcome measures include gaze, clicks, time, recalled information, perceived total cost, actual choice, field demand, and subsequent decision quality. No one measure is an adequate universal substitute for the others Russo & Leclerc (1994) Chandon et al. (2009).
This monograph is a critical narrative synthesis grounded in identifiable original studies, economic models, behavioral experiments, and current marketplace research through 2026. Its method compares theories with the concrete contrasts that each study identifies: randomized changes in disclosure can test an immediate response to visibility, whereas observational logs reveal patterns of navigation within an existing interface. Eye movements inform the sequence of sampling but require interpretation before being treated as preferences. The synthesis does not report a new experiment, a comprehensive preregistered literature search, or a meta-analysis. Its central question is how consumers decide what to inspect, how market interfaces alter that decision, and under which conditions the resulting attention changes welfare rather than simply observed transactions Caplin & Dean (2015) DellaVigna (2009).
From Information Economics to Process-Oriented Consumer Psychology
George Stigler's 1961 article The Economics of Information established an influential economic account of costly price discovery. In that tradition, information can be purchased with time, effort, travel, or direct expenditure, and searching becomes worthwhile when its expected benefit exceeds its incremental cost. The central move was to treat imperfect knowledge about prices as an element of market organization. It helps explain why apparently identical goods can carry different observed prices and why consumers do not compare every seller. Yet price dispersion alone does not identify a shopper's perceptual process. Similar observed stopping points can arise from rational information costs, attentional capture, prior experience, or strategic withholding of information. A psychologically complete analysis therefore needs observations between exposure and purchase Stigler (1961).
Martin Weitzman's 1979 treatment of optimal search for the best alternative formalized a consequential sequential problem: information about different opportunities is costly and may be sampled in an order determined by reservation values. Its elegant stopping rule works under specified distributions, independence conditions, and payoffs. In practical consumer markets, a searcher may learn while searching, revise the menu, mistake which attribute predicts quality, or incur costs that differ from those specified in the model. The importance of the classic rule is methodological. It supplies a benchmark against which to investigate actual inspection sequences and violations that are large enough to matter. To say a consumer searched less than a theoretical ideal requires an explicit account of beliefs, available opportunities, search cost, and what counts as the attainable optimum Weitzman (1979).
Consumer psychology progressively made the sequence of information acquisition observable. Johnson and Payne's 1985 computational comparison demonstrated that choice procedures can economize on effort while sacrificing varying amounts of normative accuracy, with task contingencies altering that tradeoff. Payne, Bettman, and Johnson developed the adaptive decision-maker perspective, and Bettman, Luce, and Payne synthesized constructive consumer choice in 1998: limited processing resources and task conditions lead consumers to assemble choice procedures as they decide. This tradition replaces an assumption of universal exhaustive calculation with experimentally distinguishable strategies. The evolution from normative search economics to traceable process models is a cumulative research development: economic models specify an ideal information problem, while process measures reveal how people actually solve a version of it Johnson & Payne (1985) Payne et al. (1993) Bettman et al. (1998).
Conceptual Architecture of Selective Consumer Attention
Attentional selection can occur among products, among attributes of the same product, among sellers, and among stages of a transaction. The objects of selection are consequently nested. A shopper first notices a subset of search results, next compares a limited subset of listings, then reads selected details from one listing, and eventually inspects a price summary. An observed preference for a highly ranked product may reflect its evaluation after comparison or its greater probability of entering the consideration set at all. These stages require separate outcomes. The choice probability conditional on having examined an item cannot be inferred from the unconditional share of purchases when the inclusion mechanism is itself affected by visibility Häubl & Trifts (2000) Rock et al. (2024).
Selection has both bottom-up and goal-dependent components. Text size, contrast, spatial position, and a conspicuous monetary symbol can make an item visually prominent. Prior knowledge, product involvement, budget constraints, and a buyer's purpose can direct attention to less visually conspicuous attributes. The same shopper may emphasize total price when planning a routine replenishment, durability when making a long-lived purchase, or uncertainty about a vendor when purchasing from an unfamiliar marketplace. Eye-tracking experiments on shelf displays document that visibility and evaluation are related but distinguishable; increased visual attention to a shelf location need not translate uniformly into stronger brand evaluation. Experimental design must isolate the cue from the quality and relevance of what the cue reveals Chandon et al. (2009) Russo & Leclerc (1994).
Price salience is narrower than global attention. It may be created by displaying an all-in price, by moving a tax into the posted label, by emphasizing fees, or by arranging comparisons so that a price difference becomes conspicuous. Salience may increase the likelihood of noticing the monetary attribute, change how long it is examined, and alter its decision weight. These pathways can diverge. A consumer may attend closely to a price but weight quality more heavily, while another may infer cost from a familiar store without fixating on the tag. Bordalo, Gennaioli, and Shleifer's account of salient thinking and Kőszegi and Szeidl's focusing model articulate distinct theories of context-dependent attribute weighting; the behavioral phenomena require experiments that discriminate among these mechanisms Bordalo et al. (2013) Kőszegi & Szeidl (2013).
Search Costs as Psychological and Economic Constraints
An operational search cost is the marginal burden of obtaining additional usable information. It includes monetary expenditure, elapsed time, required physical action, cognitive effort, uncertainty about the value of the next source, and the risk of abandoning the current purchase context. Digital interfaces frequently lower some components while raising others. A price-comparison page can reduce travel time and allow cross-store sorting, yet a listing containing many nonstandard attributes can increase mental integration demands. On a phone, typing a query, navigating back, and comparing two screens may impose greater switching effort than the same task on a wide monitor. Treating all of these as one unmeasured psychological cost hides meaningful differences in intervention design Lynch & Ariely (2000) Zheng et al. (2026).
Search depth can be measured as alternatives visited, pages opened, attributes revealed, click count, or total duration, but those variables represent different costs. One click may reveal an entire comparison table; five clicks might provide five snippets of redundant information. Time spent can mean deliberate verification, confusion, distraction, or complicated navigation. A credible study should specify the underlying information opportunities, the value of the signal, and how acquisition changes the decision maker's posterior beliefs. Caplin and Dean's revealed-preference approach derives restrictions under which apparently inaccurate choices are compatible with optimal costly information acquisition. Such tests warn against automatically labeling sparse search as irrational, particularly when search cost and subjective information quality are unobserved Caplin & Dean (2015).
Information costs are heterogeneous. Experienced category buyers may recognize brands, substitute durable memory for current inspection, and rely on useful defaults; novices may search widely without knowing which evidence predicts performance. Buyers under time pressure may terminate search earlier while retaining reasonably effective comparisons within the visited set. Reutskaja and collaborators found evidence for optimizing within inspected items and for a stopping pattern combining features of optimal search and satisficing under experimental time constraints. The result should not be generalized to every shopping situation, because the laboratory's payoffs, visible alternatives, time limits, and visual arrangement help determine the observed strategy. Heterogeneity is therefore part of the theory's testable content rather than an unexplained residue Reutskaja et al. (2011) Russo & Leclerc (1994).
Information Acquisition, Screening Rules, and Adaptive Choice Strategies
Consumers can use compensatory search, collecting information across attributes and allowing one advantage to offset a disadvantage, or noncompensatory screening, rejecting alternatives that fail a threshold. Elimination-by-aspects directs attention to diagnostic attributes and iteratively removes candidates. Lexicographic choice compares a primary attribute first and turns to secondary attributes only when necessary; conjunctive screening rejects an option that misses any indispensable criterion. These rules produce sharply different search paths even when they eventually select the same product. The central inferential advantage of process tracing is that it can test whether a shopper alternated by product, compared one attribute across products, or repeatedly returned to a small candidate set Johnson & Payne (1985) Bettman et al. (1998).
Strategy choice is sensitive to task structure. More attributes increase the burden of full weighted comparison; stronger constraints on decision time make a brief exclusion rule attractive; highly diagnostic attributes can make limited screening efficient. A buyer who plans to avoid products exceeding a fixed budget has good reason to inspect price early, whereas a buyer seeking a safety-critical feature may read a certification before price. Process adaptations can be understood as responses to information structure without asserting that the shopper computes an optimal strategy in advance. The effort–accuracy framework identifies the tension between accuracy of a chosen alternative and the resources used to arrive at it, but an actual purchasing environment adds trust, uncertain seller behavior, and unequal consequences of errors Payne et al. (1993) Bettman et al. (1998).
Information ordering changes the set of alternatives that receive serious evaluation. An early price filter may prevent costly inspection of unaffordable options, while an early promotional cue may attract attention before the total price is revealed. Search procedures thus shape constructed preferences: a product's attractiveness is assessed using the attributes available at the time evaluation occurs. Ariely's work on user control over information flow examines implications for judgment, knowledge, memory, and confidence; Häubl and Trifts examine interactive decision aids that structure online product comparisons. Together with adaptive-choice theory, these studies suggest that how information arrives is a scientifically distinct variable from the total quantity eventually displayed Ariely (2000) Häubl & Trifts (2000).
Eye Movements, Visual Search, and the Temporal Structure of Choice
Eye tracking offers a temporal record of where viewers direct foveal vision during a display. It makes visible an intermediate stage that purchase data alone leave hidden. Russo and Leclerc's 1994 laboratory simulation of supermarket shelves described an orientation phase, a comparatively long evaluation phase involving direct comparisons among a few alternatives, and a verification phase focused on checking the tentative choice. This account shows that consumer search has internal organization rather than a homogeneous total duration. It also demonstrates why an investigator must align eye-movement measures with defined periods in the choice task. A fixation can signify inspection, checking, or being visually drawn to a region; interpretation depends on the stage and task Russo & Leclerc (1994).
Chandon, Hutchinson, Bradlow, and Young manipulated product facings and shelf positions in eye-tracking research on in-store displays. They reported that the number of facings influenced brand evaluation through visual attention, while differences between shelf positions showed that visibility alone was not sufficient to predict evaluation. Such findings separate gaining a glance from persuading or informing a buyer. Physical position and layout can affect entry into consideration without guaranteeing a subsequent favorable assessment. A strong causal design measures attention and final behavior together, rather than treating gaze itself as a purchase outcome. Mediational claims additionally need assumptions about unmeasured causes of both gaze and evaluation that go beyond the randomization of display placement Chandon et al. (2009).
Reutskaja, Nagel, Camerer, and Rangel studied eye movements when participants chose among different numbers of options under time pressure. The study's published findings included systematic focal-region search biases, initial exploration not ordered by option value, and termination rules combining normative and satisficing features. In contrast to a story of universal indiscriminate scanning, decision makers often chose effectively within the set actually examined. Eye movements therefore support analysis of exposure, acquisition and stopping, but their evidential limits are important: covert attention can shift without a fixation; gaze can be driven by screen geometry; and repeated viewing may follow uncertainty rather than liking. Combining random visual rearrangements with monetary incentives and choice traces improves mechanism identification Reutskaja et al. (2011).
Monetary Information as a Selective Decision Input
At the point of consumer choice, a monetary amount has several possible representations. A displayed base price is a perceptual input; a recalled regular price is stored information; a tax or delivery charge may remain an anticipated future payment; and a final total is the actual monetary requirement for completing the transaction. Whether consumers use each representation depends on which components are available, attended, integrated, and compared. This four-stage analytical separation explains why accurately posting a component somewhere on a page does not establish that it influenced the decision. Economic models typically treat the relevant budget price as a single quantity, while psychology asks which parts entered the consumer's constructed evaluation. The difference becomes measurable by experimentally moving a constant cost between the initial and final stages of a choice interface Chetty et al. (2009) Morwitz et al. (1998).
A seller can make price salient through visual prominence, time of presentation, repeated reminders, juxtaposition, or a consolidated all-in figure. Such treatments often change more than one psychological feature. Enlarging a label also changes its visual detectability; placing the final cost next to product specifications can facilitate integration; revealing mandatory charges early can affect trust or perceived fairness. A study that manipulates a total-price banner and observes lower purchase intentions has identified the effect of the banner package, not necessarily a pure cognitive weight assigned to price. This distinction matters because the effective policy may be a disclosure format even when the microscopic attentional mechanism cannot be isolated. Original studies of tax-inclusiveness and partitioned pricing establish the empirical importance of presentation within particular contexts Chetty et al. (2009) Morwitz et al. (1998).
Salience also depends on comparative context. A one-dollar difference may dominate evaluation when two products are otherwise similar, while a substantial difference may receive less weight in a comparison dominated by quality or product size. Bordalo, Gennaioli, and Shleifer formalize how an attribute can become especially salient when it stands out relative to values of the same attribute across available alternatives. Kőszegi and Szeidl analyze focusing and choice when attention is drawn disproportionately to attributes with conspicuous ranges. Such models should be evaluated as explanations of a joint distribution of choices, attribute ranges, and sensitivity to experimental context, rather than simply attaching the term salience to any failure of full-price comparison Bordalo et al. (2013) Kőszegi & Szeidl (2013).
Tax-Inclusive Price Tags and the Field Experiment
Chetty, Looney, and Kroft investigated whether consumers respond differently to a sales tax when it appears within the posted price rather than being added at checkout. Their supermarket experiment attached tax-inclusive labels to approximately 750 taxable products for three weeks, leaving the original pretax prices in place and adding information about the complete amount. Scanner data were analyzed against products outside the treated categories and nearby comparison stores. The authors report approximately an 8 percent reduction in demand for the treated items after the label intervention. The monetary tax liability had not changed; the presentation of that liability had. This is especially valuable causal evidence because the intervention affected the information available during an actual transaction rather than only a hypothetical survey answer Chetty et al. (2009).
The interpretation depends on the design. Treatment was implemented in selected taxable product categories rather than through individual random assignment at every purchase; comparisons across stores and categories support the authors' identification strategy while leaving questions about contemporaneous differences and the visibility of the new tags. Displaying extra labels can itself prompt curiosity or make the shopping environment look unusual. The reported effect is a short-run demand response to a specific tax amount, category range, retail context, and implementation period. The result is therefore evidence that price presentation can materially influence real consumer behavior; it is not an estimate of how all taxes, all shoppers, or all digital payment interfaces respond to every display treatment. The paper also discusses the usefulness of alternative mechanisms and inferential qualifications Chetty et al. (2009).
The same research contrasts taxes embedded in alcohol prices with taxes added at the register, using state-level changes and consumption data. That second strategy extends the question beyond the temporary shelf-tag intervention but has the identification limitations of observational variation across jurisdictions, products, and time. Changes in consumption may reflect market adaptation, differences in enforcement, shifting social norms, and correlated policies unless appropriately addressed. The combination of a local field intervention and broader administrative variation strengthens the general proposition that timing and visibility matter while still requiring distinct causal assumptions for the two designs. In welfare analysis the amount of behaviorally perceived tax may differ from the statutory liability, complicating conventional predictions about incidence and efficiency losses Chetty et al. (2009) DellaVigna (2009).
Partitioned Prices, Mandatory Fees, and Consumer Cost Integration
A partitioned price expresses a total transaction cost as a base amount plus one or more additional charges. In their 1998 experimental investigations, Morwitz, Greenleaf, and Johnson examined consumers' processing of total costs and reactions to a base price combined with a surcharge. They reported lower recall of the complete cost and greater demand under particular partitioned-price presentations, with qualifications involving surcharge format and brand affect. The psychological issue is a failure to integrate components into the decision metric or an uneven weighting of those components. Partitioning and shrouding must be distinguished: partitioned charges may be clearly presented, while a shrouded charge is inconvenient to discover or first disclosed after earlier choice commitments. The operational difference matters for both explanation and disclosure design Morwitz et al. (1998).
Brown, Hossain, and Morgan examined shipping-fee disclosure using field and natural experiments in online auctions. Their reported results show that sellers could benefit from high shipping charges when the charges were hidden, whereas the consequences of disclosure depended on the fee level. The authors explicitly found that the reported revenue pattern was not explained by the number of bidders alone. This evidence is particularly informative because auction revenues and seller choices are observed in a transaction environment. At the same time, the experiment is about auction markets, bidding processes, and rules for fee exposure; it does not supply a context-free estimate of all checkout-fee effects in every retail category. Researchers should separately measure first-page visibility, later disclosure, total price, and bidder expectations Brown et al. (2010).
The market consequences of concealed supplementary charges also have a strategic dimension. Gabaix and Laibson model competition when some buyers anticipate add-on costs and others respond myopically to the primary price, yielding conditions under which shrouding can persist even under competition. This theoretical contribution makes a distinctive prediction about seller incentives and consumer sophistication. Its core assumptions need comparison with data: buyers can learn, alternatives can be less substitutable than assumed, and fee-search technology can change the payoffs to shrouding. Empirical evidence on shipping-charge disclosure is consistent with the importance of information suppression but cannot by itself validate every strategic equilibrium assumption. The integration of market theory and behavior should therefore remain conditional on observed practices and institutional details Gabaix & Laibson (2006) Brown et al. (2010).
Attention, Salient Thinking, and Rational Inattention Models
Models of rational inattention begin with the idea that information processing is itself costly. Sims's 2003 formulation provides a formal way to represent limits on information-processing capacity and their consequences for decisions in economic environments. Caplin and Dean subsequently show how costly information acquisition can be studied through revealed-preference restrictions. These approaches are useful because they make limited inspection potentially consistent with deliberate cost minimization. They also impose empirical burdens: probabilities, opportunity sets, and beliefs must be identified sufficiently to test the implications. Sparse gaze, incomplete recall, or an omitted price comparison is not necessarily an error simply because more information could have been collected. The value of further attention depends on what the consumer already knows and expects to learn Sims (2003) Caplin & Dean (2015).
Salient-thinking theories emphasize changes in the subjective weight of attributes caused by their comparison context. In the Bordalo–Gennaioli–Shleifer model, a dimension becomes salient partly through contrasts with other available options; hence introducing an alternative can change the perceived attractiveness of an unchanged product. Focusing models similarly emphasize disproportionate weighting of dimensions with conspicuous differences. These propositions can be contrasted with information-cost accounts: a shopper who has already seen all prices may still weight a distinctive price range differently without acquiring any additional information. The theories are distinguishable by manipulations that hold the information set fixed while changing comparison ranges, and by interventions that hold comparison context fixed while changing the cost of attending to an attribute Bordalo et al. (2013) Kőszegi & Szeidl (2013).
Other mechanisms include reference dependence, learning, limited working memory, and mistaken expectations about taxes or fees. The same behavioral fact can therefore have multiple explanatory pathways. A greater response to tax-inclusive pricing may result from attention to a previously overlooked amount, simplification of mental arithmetic, correction of mistaken beliefs, or changes in trust about the retailer. To adjudicate among these, experiments can add comprehension checks, information tests, cognitive-load conditions, and repeated exposure. Such process controls must be designed cautiously because asking respondents to state the tax immediately before choice can itself make the tax salient. Mechanism testing is a separate causal task after the reduced-form treatment effect on demand has been established Chetty et al. (2009) Dickson & Sawyer (1990) Caplin & Dean (2015).
Price Knowledge, Recall, and the Awareness Gap
Price knowledge is often inferred from the ability to recall or recognize a displayed amount. Yet ability to recall, attention while selecting, and the influence of the amount on purchase choice can diverge. Dickson and Sawyer interviewed and observed shoppers in a supermarket setting, reporting that many had not checked the selected item's price and that more than half could not accurately name its price immediately after placing it in a cart. They also found that many who bought items on special did not know the price had been reduced. The findings illustrate that an opportunity to read a label is not the same as retaining its numerical content. The study's observational nature does not establish that failure of recall caused each choice Dickson & Sawyer (1990).
Remembered prices can be less exact than categorical impressions such as cheap, familiar, unexpectedly high, or affordable relative to a self-imposed budget. The present research object concerns the acquisition and use of price-related information, whereas the systematic study of internal numeric anchors and price reference values belongs to CEC-04. In research design, memory error matters because a respondent may use a displayed price during selection and subsequently fail a test of exact recall. Conversely, a person may reproduce a number on request without having integrated it with the full transaction cost or alternatives. Researchers need time-sensitive measures that distinguish exposure, attention, comprehension, integration, and delayed recollection rather than treating one of them as interchangeable with another Dickson & Sawyer (1990) Russo & Leclerc (1994).
An added disclosure may improve price recognition while leaving transactions unaffected, especially when consumers already expect the charge or when product quality dominates comparison. Alternatively, purchasing may change without large measurable improvements in immediate explicit recall if the disclosure alters the perception of overall affordability. Studies that test both knowledge and demand can describe the relation between the two, but causal mediation needs a design that identifies the effect of acquiring and remembering information separately. A useful operational framework combines displayed information, eye or click exposure, comprehension checks, and objectively recorded behavior. This layered design makes price salience a tractable hypothesis about mechanisms, not merely a synonym for a successful nudge Chetty et al. (2009) Chandon et al. (2009) Morwitz et al. (1998).
Consumer Welfare and the Limits of Behavioral Interpretation
Consumer welfare cannot be inferred from an increase or decrease in sales alone. An earlier total-price disclosure may lead to fewer purchases because it exposes unaffordable costs, because it creates confusion, or because it makes consumers excessively sensitive to an attribute that should have little marginal value. A higher conversion rate can reflect a genuinely easier comparison procedure, an unanticipated fee, or aggressive promotion. Field demand is therefore an important outcome but only one component of evaluation. Better studies specify welfare-relevant reference outcomes such as informed expenditure, postpurchase regret, product fit, budget constraints, or meaningful consumer surplus under a stated behavioral model. Chetty and colleagues show why apparent misoptimization over taxes can alter policy incidence and efficiency calculations Chetty et al. (2009).
Energy-efficiency disclosures offer one way to test whether better information produces more nearly informed consumer decisions, but their welfare interpretation is also demanding. Allcott and Taubinsky examined lightbulb choices through field and experimental evidence in a policy-evaluation framework that explicitly distinguishes behavioral mistakes, heterogeneity, and policy consequences. Results from a lightbulb market cannot be directly transported to every price-transparency intervention: energy costs accumulate over time and require forecasts, while a checkout fee is immediate and contractually specified. The methodological contribution is the insistence that behavioral interventions be evaluated against their costs and benefits for actual consumers instead of relying on a psychological label such as bias Allcott & Taubinsky (2015).
Price transparency also involves distributional consequences. A disclosed fee can benefit price-comparison shoppers while changing the competitive strategies of sellers; consumers who face language barriers, visual constraints, or unfamiliar payment systems may be affected differently from experienced buyers. To understand the relevant welfare effect, an evaluator should estimate heterogeneity in both information exposure and final transaction cost, assess substitution toward other products, and investigate longer-term learning. The direct effect of a display intervention is not automatically the same as its market-wide equilibrium effect after sellers revise price schedules. These distinctions keep behavioral findings scientifically useful for consumer protection without turning an observed short-run response into an unqualified normative verdict Brown et al. (2010) Gabaix & Laibson (2006) DellaVigna (2009).
The Architecture of Physical Retail Search
Physical stores present information through placement, packaging, shelf depth, visual contrast, and the path required to approach a product. These features alter opportunities for inspection before the shopper compares prices or attributes. A product at eye level is often easier to notice than a product near the floor, yet visibility is only one determinant of eventual evaluation. Chandon and colleagues examined how shelf facings and positions affected visual attention and brand evaluation, thereby separating the probability of seeing a product from the evaluative consequences of that exposure. The study establishes why a merchandising measure such as shelf share is an imperfect proxy for the consumer's information state. The same display can attract a glance without producing knowledge of the total price or a decision to buy Chandon et al. (2009).
Russo and Leclerc's process tracing adds a temporal interpretation. A supermarket shopper can initially orient toward the whole display, compare several remaining candidates, and then return to one option to verify a selection. These phases matter when investigating a price label. A label placed at the entry stage may influence which items remain in consideration, while an all-in price encountered only during late verification may influence abandonment after effort has already been invested. An intervention should therefore be specified relative to the purchase sequence, not merely categorized as visible or hidden. Comparing visual search before and after a new label requires stable product layouts, comparable tasks, and careful differentiation between novel display effects and enduring informational benefits Russo & Leclerc (1994) Chetty et al. (2009).
Offline field measurements include scanner-based purchases, intercept interviews, eye tracking, store-path sensors, and observational coding of shelf interactions. Each observes a different spatial resolution and changes the researcher's ability to infer search. A shopper entering the aisle has not necessarily inspected every product there; a hand reaching for an item may reflect familiarity rather than comparative evaluation. Modern reviews of pre-purchase data emphasize that movement, visits, and click traces expand the evidence beyond transactions while preserving important blind spots. A robust physical-retail study links observable opportunity to exposure and final transaction, ideally with randomized display variations. Its transfer to digital commerce should be tested because scrolling, screens, and navigation impose distinctive costs Honka et al. (2024) Dickson & Sawyer (1990).
Online Comparison Tools and the Cost of Acquiring Attributes
Digital search can reduce the effort needed to identify vendors and compare standardized product attributes. The reduction in cost is selective. A side-by-side table makes some differences prominent, while details omitted from the comparison remain expensive to acquire. Häubl and Trifts studied interactive decision aids in online shopping environments and demonstrated that tools can shape decision-making behavior through the way they filter and compare product information. The causal object is the tool as actually implemented, including its interface, available attributes, and guidance. A tool that organizes price and delivery cost together changes attention allocation and arithmetic burden; a tool that shows price alone may shift search toward an incomplete monetary attribute. Such distinctions become central when evaluating apparent gains in consumer efficiency Häubl & Trifts (2000).
Lynch and Ariely's study of online wine retail investigated how search costs change competition over price, quality, and distribution. It shows why easier digital search does not imply a single predictable effect on all sellers' prices. Making price information more accessible can intensify comparisons on price, while information about product quality can change perceived substitutability and give buyers reasons to prefer differentiated products. The mechanism therefore depends on which attribute becomes cheaper to compare and which alternatives are plausibly comparable. Online information provision should be assessed at the attribute level rather than treating all reduced search costs as interchangeable. This approach also clarifies why a retailer can simultaneously lower price-comparison friction and highlight desirable nonprice product features Lynch & Ariely (2000).
Interactive controls, filter presets, and sorting orders create opportunity sets rather than merely displaying neutral facts. A shopper selecting 'price low to high' is revealing an intention to inspect price, but the available filter can also shape that intention by making one criterion convenient. A product removed by a filter may never enter the considered set; a comparison table can reduce the number of pages examined while increasing the usable information extracted per minute. Researchers should measure information content obtained and decision quality in addition to number of clicks. The broader adaptive-choice literature explains why constrained shoppers select workable procedures, whereas search models identify the value of obtaining further information under specified costs and expectations. These perspectives jointly structure digital interface experiments Bettman et al. (1998) Häubl & Trifts (2000) Caplin & Dean (2015).
Clickstream Evidence and the Formation of Consideration Sets
Clickstreams record product-page visits, query terms, navigation sequences, and sometimes scrolling or dwell time. They are valuable because many consumers inspect several items before buying one, and some never purchase. Honka, Seiler, and Ursu's 2024 review synthesizes how pre-purchase observations help estimate preferences, understand how retailers change search, and analyze market outcomes when knowledge is limited. The review draws a critical distinction between search data and purchase data: buying identifies an outcome conditional on exposure and evaluation, whereas browsing can reveal the range of alternatives that was seriously considered. Yet a page visit is a proxy for acquiring information, not a direct observation of comprehension or attribute weighting Honka et al. (2024).
Estimating preference from purchases alone can exaggerate the appeal of prominently displayed products because products rarely encountered have little opportunity to be chosen. Modeling the consideration set changes the estimation problem. Researchers can combine a product's position, likelihood of being visited, and probability of purchase after inspection, while acknowledging that all three may depend on unobserved consumer interests. Honka and colleagues describe search data as a way to improve preference estimation and to investigate how marketing interventions affect the choice funnel. These advantages require assumptions about why a shopper visits particular products, how options are revealed, and which costs are incurred. A structural model that explains the observed click distribution can still be wrong about the underlying utility function if it imposes an incorrect search rule Honka et al. (2024) Weitzman (1979).
Search logs also contain selection and missing-data problems. Platforms record activity only within their observable environment and often cannot see comparisons on competing sites, conversations with friends, saved screenshots, or offline visits. Time between clicks can include distractions unrelated to the purchase. A search result ignored by a user may have been unseen, seen and rejected, or remembered from earlier experience. These explanations produce different counterfactual effects of moving that result higher on a page. Controlled reordering experiments and user-level panels can help distinguish them, but the ethical and privacy costs of intensive tracking require careful minimization. Successful research integrates behavioral traces with transparent measurement design and avoids treating every observed click as an intentional signal of interest Honka et al. (2024) Rock et al. (2024).
Ranking Position, Sponsored Results, and Selection Bias
Search and recommendation pages make some products disproportionately easy to encounter. Screen position, viewport limits, visual badges, and sponsored placements can affect the probability of inspection before product evaluation begins. Rock, Strauss, O'Reilly, and Mazzucato investigated Amazon search behavior with observational analysis of more than two thousand queries and identified a pronounced importance of position and attention allocation. Such evidence is relevant to the mechanisms through which platform design shapes exposure. Its observational structure matters: products are placed in ranks partly because of relevance, seller characteristics, pricing, or platform decisions. Without randomized ranking or strong exogenous variation, positional differences cannot be treated as the isolated causal effect of location alone Rock et al. (2024).
An item may have high sales because it is genuinely attractive to the consumer population, because it receives more exposure, or because buyers infer quality from high placement. A position intervention can also change beliefs about what the platform recommends, creating an interpretive effect beyond visual availability. These pathways require distinct tests. If ranks are randomized while product descriptions remain constant, differences in inspection can estimate a position effect within the tested interface. To determine whether information was actually acquired, researchers may additionally collect recall or comprehension measures, while avoiding questions that themselves change the search process. Platform ranking belongs to the neighboring algorithm dossier when its optimization, governance, and personalization are the primary research object; here its significance is that search architecture mediates attention Rock et al. (2024) Honka et al. (2024).
Sponsored listings and visual hierarchy also raise the question of whether consumers can distinguish advertising, recommendation, and organic search. A clear advertising label is one information cue among many; its mere presence does not demonstrate that everyone saw it or correctly understood its implications. Research should separate the treatment of labeling from the treatment of rank, promotional copy, and product price. In multi-attribute environments, a high rank might draw an initial glance, while the subsequent price comparison determines whether the listing survives screening. A credible field study reports relevant position strata, platform context, changes in the composition of exposure, and behavioral endpoints beyond simple clicks. Mechanistic claims require more than the association between rank and purchase Honka et al. (2024) DellaVigna (2009).
Advertising Disclosure and Experimentally Observed Search
Advertising supplies information before consumers begin a full product comparison. Its content can disclose a price, a category, a quality claim, or a visually memorable feature. Morozov and Tuchman's 2024 study constructed a simulated online bookstore and randomly varied both advertising exposure and the attribute content presented to visitors. The preregistered, incentive-compatible experiment found that advertising a book's low price increased demand for it relative to ads without that price information. Revealing its genre had more polarized effects: some consumers searched or purchased after learning the genre, while others rejected the book without further search. This demonstrates that information disclosure can redirect the search path by changing what consumers believe about a product Morozov & Tuchman (2024).
A noteworthy finding is that the experiments did not show a general increase in the total amount of search from revealing product attributes in ads. Instead, the content changed where attention and interest were directed. This distinction makes broad claims such as 'more informative advertising causes more search' inadequate. A consumer can become better informed by eliminating an option early and conducting fewer subsequent inquiries; alternatively, a cue can focus exploration without increasing aggregate browsing. The study illustrates why a search outcome should be decomposed by advertised product, competing products, attribute content, and ultimate purchase. The simulated bookstore's controlled structure improves identification but does not automatically reproduce the incentives, familiarity, and time pressure of all real e-commerce settings Morozov & Tuchman (2024) Bettman et al. (1998).
Price disclosure in advertising differs from total-price transparency at checkout. An ad may truthfully reveal a low headline price while postponing mandatory delivery fees or other costs. The early cue can increase consideration, after which later information changes the perceived value. Testing the whole consumer pathway requires recording the price initially seen, subsequent information acquisition, fee visibility, and behavior after complete disclosure. Designs can cross-randomize ad content and checkout format to estimate whether early price salience interacts with later cost integration. Evidence from the bookstore experiment and from auction fee interventions motivates such tests but does not itself supply the result of an unperformed combined experiment Morozov & Tuchman (2024) Brown et al. (2010) Chetty et al. (2009).
Mobile Search Aids and New Field Evidence
Mobile commerce places price and product information into a narrow viewport and often separates attributes across scrolling panels, filters, and modal windows. This changes the costs of moving between alternatives and the ease of retaining comparisons in working memory. In a 2026 Information Systems Research field experiment, Zheng, Tong, Fang, Gopal, Li, and Jiang studied the display of popular search genres as a search aid on a large mobile shopping platform. Their study reported shifts in search behavior and purchase outcomes under randomized exposure to the aid. The relevant claim concerns the tested platform, interface, and shopping tasks, not a universal increase in consumer rationality from adding recommendations. The contribution is experimental evidence that relatively small upstream interface cues can change subsequent search pathways Zheng et al. (2026).
Search aids can reduce the need to formulate a query while increasing exposure to products selected by an interface. A decrease in the number of submitted search terms is therefore compatible with more extensive product exploration or a higher final purchase rate. This creates a measurement problem: the study of search must distinguish query effort from result inspection, information acquisition, and final selection. A simple metric such as search count cannot adequately characterize all stages. The 2026 experiment is valuable because it treats a search aid as a manipulable component of a real digital environment rather than assuming that the presence of more information always produces more active query generation. The interpretation still depends on population, platform, ranking methods, and how search behavior was measured Zheng et al. (2026) Honka et al. (2024).
Contemporary evidence invites a cumulative model of consumer search across offline shelves, desktop comparison tools, and mobile platforms. The shared processes include opportunity for exposure, selective sampling, cost-sensitive continuation, belief revision, and stopping. Their technical implementation differs across environments. Physical position is not identical to ranking order; a gaze fixation is not identical to a click; a mobile suggestion is not equivalent to a voluntarily typed query. Comparative studies should preserve these distinctions while testing whether the same theoretical predictions survive changed interfaces. Such work can identify when search tools improve informed choice and when they merely redistribute attention across sellers. Findings from field experiments and pre-purchase data reviews show why the observable search process deserves direct study alongside market demand Reutskaja et al. (2011) Honka et al. (2024) Zheng et al. (2026).
Operational Definitions and an Evidence Matrix
A reproducible study must define which stage of the search process is observed. Exposure is an opportunity to encounter information; attention is selective allocation of processing; acquisition is the obtaining of information; comprehension is an accurate interpretation of its content; integration is the incorporation of that information into option evaluation; and choice is the final recorded action. These stages often correlate, but they are not interchangeable. A click can register acquisition while leaving comprehension unknown. A gaze fixation can establish a location of overt vision without showing which semantic proposition was extracted. A field purchase can establish behavior without revealing which alternatives were evaluated. This layered vocabulary permits competing accounts to be stated as different causal claims rather than as different names for the same observed effect Honka et al. (2024) Russo & Leclerc (1994) Chetty et al. (2009).
An evidence matrix for price salience should cross-classify the manipulated information with the observed response. The independent variable may be the position of a label, the inclusion of tax, the sequence of fee disclosure, the availability of a filter, or the rank of search results. The outcome may be gaze, product visits, quantity purchased, recall of the total cost, or choice satisfaction. Studies with identical nominal treatments but distinct outcomes answer different questions. It is also necessary to record the sample, product category, price scale, decision incentives, observation period, and availability of alternative sellers. Chetty and colleagues' field scanner data and Morwitz and colleagues' experimental cost recall are complementary precisely because the endpoint and design differ Chetty et al. (2009) Morwitz et al. (1998).
Interpretation requires distinguishing direct observation from inferred states. The price that appeared on a screen is observed information availability; a participant's report of having noticed it is a fallible measure; attention as a cognitive process may be modeled from eye movements, time, comprehension, and choice together. Research that combines these data can constrain hypotheses more tightly than a single dependent variable, but multimethod agreement is not a guarantee of causality. Measurement itself can change the process: asking consumers repeatedly what price they see is likely to direct their attention toward price. Study protocols should therefore specify whether process measurements were passive, delayed, or presented during choice and evaluate their likely demand characteristics Reutskaja et al. (2011) Dickson & Sawyer (1990) Honka et al. (2024).
Experimental Designs and the Identification of Causal Effects
The cleanest experimental question holds a product's actual price and features fixed while varying the visibility of one information component. Random assignment of a tax-inclusive label, fee disclosure, or comparison tool can then identify an average treatment effect for the population, period, and implementation under study. A factorial design can distinguish the salience of the total price from the salience of the tax rate itself, provided the text and arrangement are otherwise comparable. When experimental control is restricted by the retail setting, quasi-experimental strategies rely on explicit counterfactual assumptions. The supermarket tax-label study uses comparison products and stores; its inference is more credible when researchers scrutinize parallel trends, seasonal changes, and unintended shelf attention Chetty et al. (2009).
Choice experiments benefit from consequential incentives, but a monetary payoff alone does not make a task equivalent to everyday shopping. Real transactions involve budgets, delivery, trust, returns, brand familiarity, and uncertain product quality. An online mock store can precisely vary display content while sacrificing some of those features. Morozov and Tuchman's bookstore experiment illustrates the value of random exposure and information content; Brown and colleagues' field auctions illustrate market stakes together with more complicated bidding dynamics. These designs answer overlapping but different questions. Scientific synthesis should extract the estimand actually identified by each study, then ask what additional assumptions are required for a claim about a different market or a sustained policy intervention Morozov & Tuchman (2024) Brown et al. (2010).
Causal mediation requires special care. Randomizing a salient label establishes the total effect of that label on behavior under valid assignment and implementation. It does not automatically establish that the effect operated through eye fixations, corrected mental arithmetic, increased distrust, or stronger aversion to spending. Those mediators may respond to prior preferences and to unobserved events after treatment. Better designs can manipulate a component of information availability separately, test recall or attention at appropriate moments, and specify a causal model of intermediate variables. Even then, measurement-induced attention can complicate interpretation. The central lesson is to avoid replacing a measured price-disclosure effect with a more elaborate psychological story unless the necessary intermediate evidence exists Chetty et al. (2009) Reutskaja et al. (2011) Caplin & Dean (2015).
Structural Search Models, Behavioral Traces, and Identification
Structural economic models represent how consumers sample alternatives, form expectations about uninspected options, and stop searching. Weitzman's sequential search framework supplies an influential optimal-search benchmark. Extensions may allow heterogeneous costs, learning, nonstandard stopping rules, or ranking-dependent inspection. A structural model can use the order of page visits and observed choices to estimate parameters that purchase data alone weakly identify. Yet such inference depends on assumed information sets and preference stability. A lower probability of clicking a product may reflect low utility, high effort to reach it, or a ranking algorithm that reduces exposure. Model comparison should examine whether multiple parameter configurations reproduce the same observed purchase and search data Weitzman (1979) Honka et al. (2024).
Process models of adaptive choice provide a different level of explanation. The consumer may compare prices across products, reject candidates below a quality threshold, switch among attributes, or use a familiar brand heuristic. The effort–accuracy approach evaluates the computational demands and consequences of these alternative procedures. Evidence from eye tracking can favor some process classes when the recorded inspection path reflects their predicted transitions. However, fitting eye movements to a strategy without testing out-of-sample predictions risks circular explanation. Combining path traces, manipulated attribute distributions, and a priori predictions improves discrimination. The goal is a model that predicts what inspection changes when a price label moves, a fee becomes visible, or time is constrained Johnson & Payne (1985) Payne et al. (1993) Reutskaja et al. (2011).
Revealed-preference rational-inattention tests address whether stochastic information choices satisfy implications of optimal costly attention. Caplin and Dean's framework is especially relevant to distinguishing attention cost from a presumed defect in preferences. Its domain depends on how the analyst defines the state, signal, feasible information structures, and outcome payoffs. Consumers operate with imperfect and sometimes erroneous subjective beliefs; testing rationality under an externally imposed belief model need not identify the actual person's mental process. The strongest empirical strategy triangulates the different approaches rather than declaring one universal search theory. Economic optimality models provide comparative restrictions, process theories predict inspection sequences, and field experiments test causal changes from information interventions Caplin & Dean (2015) Sims (2003) Honka et al. (2024).
Heterogeneity, Expertise, Time Pressure, and Material Constraints
Consumers differ in their experience, product knowledge, numerical confidence, available time, reading conditions, and consequences of making a mistaken purchase. These differences can alter how much search is necessary to make an informed choice. A repeat buyer may rationally inspect fewer details because past experience supplies relevant information; a first-time buyer may explore many products to discover which features matter. High price salience can be helpful for a strict budget yet less relevant when product compatibility determines whether a purchase is useful. Heterogeneity should therefore be tested in treatment interactions rather than dismissed as sampling noise. Even an average effect estimated precisely in a large experiment may conceal responses of opposite direction in distinct consumer groups Bettman et al. (1998) Lynch & Ariely (2000).
Time pressure makes the stopping problem visible. In eye-tracking experiments on choice among multiple options, Reutskaja and colleagues found that participants' search was shaped by limited time and focal viewing tendencies, with evidence that choices were often effective among options actually inspected. This finding complicates a simplistic description of shortened search as poor decision-making. Under a credible deadline, information about one more item may have a lower expected value than completing a purchase. The proper comparison is between achieved choice quality and feasible alternatives under the same constraint. Researchers should experimentally vary the deadline when practical, assess speed–accuracy tradeoffs, and distinguish a truly consequential time limit from instructions that merely encourage hurried responses Reutskaja et al. (2011) Johnson & Payne (1985).
Material constraints alter both the marginal utility of money and the cost of searching. A mandatory fee can be decisive for a shopper with little budget slack while receiving little attention from a more affluent shopper familiar with the charge. Search platforms can also impose uneven burdens through language complexity, device access, connectivity, or visual design. These are testable moderators of the influence of salient price information. Their interpretation should be grounded in measured resources and contextual variables rather than assumptions based on demographic categories. A consumer-centered evaluation estimates whether disclosure reduces errors in total-cost understanding across the populations most affected by the charge, and investigates whether added complexity introduces new barriers to comprehension Chetty et al. (2009) Morwitz et al. (1998) Honka et al. (2024).
Replication, External Validity, and Cross-Market Transportability
Replication requires reproducing the relevant psychological intervention, not merely the visual appearance of an earlier display. A tax-inclusive label in a supermarket, a delivery total in a mobile cart, and a fee disclosure in an auction may all increase visibility of monetary information, but they differ in legal obligations, timing, product substitution, and consumer expectations. The similarity of their observed effects would support a general mechanism only after these differences are examined. Original work on tax inclusion, partitioned prices, and auction shipping charges provides several convergent reasons to expect presentation sensitivity, alongside distinct designs and endpoints. The existence of converging findings does not license averaging their effects without a defensible synthesis model Chetty et al. (2009) Morwitz et al. (1998) Brown et al. (2010).
Publication and analytic-selection biases are possible whenever researchers can vary outcome definitions, subgroups, time windows, exclusions, and model specifications. Studies that measure clicks, gaze, recall, conversion, and total expenditures contain many correlated endpoints; a selective report of the most favorable endpoint can exaggerate treatment reliability. Preregistration, disclosed primary outcomes, accessible analysis code, and independent replications increase interpretability. Field experiments should report attrition, failed treatment exposure, balance, contamination, and operational changes. For quasi-experimental designs, alternative comparison groups and sensitivity tests are particularly important. A critical narrative synthesis should make these design differences explicit and avoid claiming formal publication-bias estimates without a sufficiently defined dataset and statistical procedure Morozov & Tuchman (2024) Chetty et al. (2009).
Transportability also concerns the behavior of firms and interfaces after a policy becomes widespread. An isolated experiment can estimate responses to a fixed design, while a systemwide requirement for fee disclosure may cause sellers to restructure base prices, delivery charges, or promotional strategies. In the long run, consumers can learn, sellers can innovate around disclosure rules, and the composition of observed buyers can change. Theory about shrouded attributes helps formulate these counterfactual possibilities, and empirical studies can test them using longer panels or policy discontinuities. The distinction between immediate psychological response and market equilibrium is essential for using salience research in consumer regulation or competition analysis Gabaix & Laibson (2006) Brown et al. (2010) DellaVigna (2009).
Transparency, Consumer Autonomy, and Research Ethics
Information design affects the conditions under which consumers exercise meaningful choice. Transparent displays make relevant terms available at the stage when comparison is possible; accessible navigation and comprehensible totals reduce the effort required to verify a transaction. Yet a disclosure that is technically present but visually obscure may have limited practical value. Ethical evaluation should therefore examine actual exposure and comprehension, not only the existence of a text string in the interface. Experiments can compare total-price comprehension, accuracy of cost estimation, and decisions after reflection. These outcomes measure different aspects of consumer autonomy and can reveal when a design appears transparent without enabling informed comparison. The normative judgment rests on specified criteria, while the empirical research identifies their measurable consequences Morwitz et al. (1998) Chetty et al. (2009).
Persuasive presentation can benefit buyers by highlighting reliable price and quality information, but it can also channel attention toward attributes selected primarily for the seller's interest. The research object remains the mechanism of allocation and discovery. A test of whether a prominent discount banner misleads consumers must establish the underlying price information, shoppers' beliefs, the full transaction cost, and their responses under alternative presentations. Seller profit, purchase volume, and number of clicks are not by themselves sufficient ethical indicators. Evidence that fee visibility affects revenues, for example, demonstrates the importance of timing and disclosure without resolving every question about fairness or consumer satisfaction. Ethics and behavioral causality should inform each other while retaining clear definitions Brown et al. (2010) Chandon et al. (2009).
Observational consumer search research presents privacy risks because detailed click paths and movement trajectories can reveal interests, routines, and sensitive circumstances. The scientific value of granular information should be weighed against data minimization, retention, consent, and safeguards against reidentification. Researchers may need aggregate analysis, controlled access, or experimental environments that avoid collecting irrelevant personal details. Such practices have methodological benefits as well: preregistered limited variables reduce scope for opportunistic analysis and make interpretation more reproducible. A transparent study reports what was observed, which events were outside the observation window, and how the absence of full cross-platform tracking affects inference. The ethical architecture of evidence collection is consequently part of research quality Honka et al. (2024).
Human–Artificial Search Assistance in the Artificial Era
Artificial systems increasingly participate in the information environment through search interfaces, comparison aids, recommendation tools, and conversational assistance. For the present research object, the important empirical questions concern the distribution of information acquisition between a human consumer and an artificial system, the visibility of monetary terms, and the human's ability to inspect and correct a recommendation. Cognitive offloading occurs when an external tool performs an operation the consumer would otherwise undertake, such as sorting prices or aggregating mandatory fees. Whether this improves choice depends on the accuracy of the system's information, the consumer's goals, and the costs of verification. Established search economics supplies a framework for describing these operations without assuming that tool use automatically increases welfare Häubl & Trifts (2000) Honka et al. (2024).
A controlled evaluation of an artificial shopping assistant can randomize access to an aid while holding the menu and transaction conditions constant. Outcomes can include completeness of total-cost identification, search duration, coverage of relevant alternatives, willingness to verify displayed evidence, and eventual purchase fit. A factorial design might compare assistants that display itemized fees, all-in prices, or only headline prices. Such a study would test whether computational simplification changes attention to previously neglected attributes, and whether delegation reduces user inspection. The design must distinguish model-generated assertions from verified marketplace facts. The theoretical possibility is motivated by work on comparison aids and modern information search; it remains a research hypothesis until tested under the specified conditions Häubl & Trifts (2000) Caplin & Dean (2015).
The broader Artificial Era places this research within human–artificial relations, where a nonbiological information-processing system can mediate human decisions. The scientific claim is operational: a particular system can select, summarize, rank, or display relevant information and thereby alter the human's observed search path. Attribution of psychological subjectivity to the artificial system is a separate philosophical and empirical question. For consumer psychology the near-term task is to measure reliability, omission of fees, provenance of quoted prices, error correction, and the accessibility of original evidence. An independent scientific study should preserve individual accountability and compare human-only, assisted, and partly delegated search under observable constraints. These are tractable questions linked directly to the established literature on search aids and limited information Honka et al. (2024) Häubl & Trifts (2000) Sims (2003).
Major Scientific Disputes and Alternative Explanations
A recurring dispute is whether observed price salience effects indicate constrained attention, mistaken beliefs, or a changed evaluation of otherwise known information. Those mechanisms can yield the same final purchase pattern. A shopper who ignores the posted tax may simply fail to notice it, underestimate its size, postpone integration until checkout, or deliberately avoid the burden of calculating it because the expected decision benefit is small. Salience theories, rational-inattention models, and bounded-information explanations are not interchangeable. Researchers should design manipulations that vary display prominence, comprehension cost, and contrast across options independently where feasible. They should then examine differences in information recall, search sequence, and actual transactions before attributing the effect to a single mental pathway Chetty et al. (2009) Sims (2003) Bordalo et al. (2013).
A second dispute concerns the meaning of search intensity. More search may produce better comparison, but it can also indicate confusion or unfamiliarity; less search may reflect effective expertise, constrained opportunity, or manipulation of exposure. Sequential search benchmarks describe an optimal stopping rule for particular environments, whereas eye tracking documents behavior in specific visual tasks. Apparent deviations from a benchmark can arise because the model omits subjective beliefs, a changing product menu, or cognitive costs. Process theories give an alternative description through adaptive decision strategies. The empirical problem is to identify the conditions under which predicted search paths diverge and to measure actual decision performance on those same tasks Weitzman (1979) Reutskaja et al. (2011) Payne et al. (1993).
A third dispute concerns the social value of attention-shaping interfaces. Price aggregation can support comparison while a platform may rank products partly according to commercial incentives. Disclosing mandatory charges earlier can improve comprehension while sellers may adjust their price structures. A change that increases apparent conversion can therefore have ambiguous consumer-welfare consequences, and a change that decreases purchases may reflect successful avoidance of unexpectedly costly transactions. Scientific analysis requires the actual counterfactual price and information environment, an appropriate welfare endpoint, and a design capable of separating consumer behavior from seller adjustment. Modern pre-purchase research makes these questions more observable, but it does not eliminate the need for causal identification Honka et al. (2024) Rock et al. (2024) Gabaix & Laibson (2006).
Research Priorities and Testable Questions
The first priority is to distinguish exposure, acquisition, integration, and valuation in the same experimental environment. Researchers can present identical product sets with randomized fee positions, collect passive eye or navigation traces, and test comprehension without repeatedly interrupting the natural choice. The study would identify whether changed purchasing arises principally through the probability of noticing a fee, through more accurate computation of the total, or through differential weighting after comprehension. Such work should preregister the primary contrasts and include control conditions that account for added visual material. A robust mechanism claim would require convergence among behavioral traces, the information actually understood, and consequential choices Chetty et al. (2009) Morwitz et al. (1998) Russo & Leclerc (1994).
The second priority is to test heterogeneous search constraints. A common experiment can vary decision time, search cost, price visibility, and prior familiarity, while holding the product information objectively comparable. It should measure whether the same display reduces errors in one group while increasing cognitive burden in another. Related research can compare mobile devices, desktop layouts, and physical shelves within a consistent monetary task. This cross-environment design would show which features of selective information acquisition are generalizable and which depend on screen geometry or established shopping routines. Analyses should separately report average treatment effects and preplanned interactions with measured expertise and resource limitations Reutskaja et al. (2011) Lynch & Ariely (2000) Zheng et al. (2026).
The third priority is longitudinal and market-level verification. Short-term gains from information disclosure may fade with learning or change when sellers alter fees and product positioning. Repeated panels and policy experiments can follow consumer understanding, search costs, switching, and total expenditure over time. Related artificial-assistance research can test the accuracy of aggregated prices, users' independent verification behavior, and the effects of explanation or source links. Equilibrium models of shrouded attributes supply contrasting predictions about seller adjustments; field experiments and administrative records can test these under documented assumptions. Taken together, these priorities move research from demonstrations of attention-sensitive choice toward causal knowledge of when information architecture improves durable consumer outcomes Brown et al. (2010) Gabaix & Laibson (2006) Honka et al. (2024).
Conclusion
Consumer attention, information search, and price salience constitute an integrated research field concerned with the sequence through which market information becomes effective in choice. Economic search theory explains why exhaustive inspection can be prohibitively costly. Adaptive-choice and process-tracing research shows how consumers sample information, screen alternatives, and stop. Field experiments establish that the placement and completeness of price information can affect real demand even when the underlying monetary cost remains constant. Digital research demonstrates that search tools, advertising content, and interface organization redirect the paths through which buyers encounter products. These findings are strongest when interpreted with the experimental conditions, information sets, and relevant outcome measures made explicit Stigler (1961) Bettman et al. (1998) Chetty et al. (2009) Morozov & Tuchman (2024).
The empirical synthesis supports several distinct propositions. Information availability does not guarantee attention; attention does not guarantee comprehension; comprehension does not guarantee integration into an economically relevant comparison; and observed transactions alone do not reveal which of these stages produced a decision. Search depth is partly determined by cost and expected informational value, while contextual salience can change attribute weighting even when the visible information remains constant. The main theoretical approaches explain overlapping phenomena but yield discriminating predictions about exposure, comparison context, and stopping. Reliable interpretation therefore requires multiple measures, causal contrasts, and careful separation of outcome effects from proposed mental mechanisms Sims (2003) Caplin & Dean (2015) Bordalo et al. (2013) Honka et al. (2024).
The enduring scientific contribution of this literature is a precise account of consumer choice as an information-dependent process situated within physical and digital markets. The practical implication is to evaluate information architecture at the time when a consumer can still compare alternatives and respond meaningfully to complete costs. Future studies should test the generality of disclosure effects across market categories, the long-term adaptations of sellers and buyers, and the benefits and risks of delegated digital search. Within the Artificial Era, the same explanatory discipline applies to information supplied by artificial assistants: verification of cost, clarity of provenance, and observable effects on human judgment are central research targets. The resulting framework supports cumulative inquiry into market choices under limited attention Chetty et al. (2009) Honka et al. (2024) Zheng et al. (2026).
References
- Allcott, H., & Taubinsky, D. (2015). Evaluating behaviorally motivated policy: Experimental evidence from the lightbulb market. American Economic Review, 105(8), 2501–2538. https://doi.org/10.1257/aer.20131564
- Ariely, D. (2000). Controlling the information flow: Effects on consumers' decision making and preferences. Journal of Consumer Research, 27(2), 233–248. https://doi.org/10.1086/314322
- Bettman, J. R., Luce, M. F., & Payne, J. W. (1998). Constructive consumer choice processes. Journal of Consumer Research, 25(3), 187–217. https://doi.org/10.1086/209535
- Bordalo, P., Gennaioli, N., & Shleifer, A. (2013). Salience and consumer choice. Journal of Political Economy, 121(5), 803–843. https://doi.org/10.1086/673885
- Brown, J., Hossain, T., & Morgan, J. (2010). Shrouded attributes and information suppression: Evidence from the field. The Quarterly Journal of Economics, 125(2), 859–876. https://doi.org/10.1162/qjec.2010.125.2.859
- Caplin, A., & Dean, M. (2015). Revealed preference, rational inattention, and costly information acquisition. American Economic Review, 105(7), 2183–2203. https://doi.org/10.1257/aer.20140117
- Chandon, P., Hutchinson, J. W., Bradlow, E. T., & Young, S. H. (2009). Does in-store marketing work? Effects of the number and position of shelf facings on brand attention and evaluation at the point of purchase. Journal of Marketing, 73(6), 1–17. https://doi.org/10.1509/jmkg.73.6.1
- Chetty, R., Looney, A., & Kroft, K. (2009). Salience and taxation: Theory and evidence. American Economic Review, 99(4), 1145–1177. https://doi.org/10.1257/aer.99.4.1145
- DellaVigna, S. (2009). Psychology and economics: Evidence from the field. Journal of Economic Literature, 47(2), 315–372. https://doi.org/10.1257/jel.47.2.315
- Dickson, P. R., & Sawyer, A. G. (1990). The price knowledge and search of supermarket shoppers. Journal of Marketing, 54(3), 42–53. https://doi.org/10.1177/002224299005400304
- Gabaix, X., & Laibson, D. (2006). Shrouded attributes, consumer myopia, and information suppression in competitive markets. The Quarterly Journal of Economics, 121(2), 505–540. https://doi.org/10.1162/qjec.2006.121.2.505
- Häubl, G., & Trifts, V. (2000). Consumer decision making in online shopping environments: The effects of interactive decision aids. Marketing Science, 19(1), 4–21. https://doi.org/10.1287/mksc.19.1.4.15178
- Honka, E., Seiler, S., & Ursu, R. (2024). Consumer search: What can we learn from pre-purchase data? Journal of Retailing, 100(1), 114–129. https://doi.org/10.1016/j.jretai.2024.02.003
- Johnson, E. J., & Payne, J. W. (1985). Effort and accuracy in choice. Management Science, 31(4), 395–414. https://doi.org/10.1287/mnsc.31.4.395
- Kőszegi, B., & Szeidl, A. (2013). A model of focusing in economic choice. The Quarterly Journal of Economics, 128(1), 53–104. https://doi.org/10.1093/qje/qjs049
- Lynch, J. G., Jr., & Ariely, D. (2000). Wine online: Search costs affect competition on price, quality, and distribution. Marketing Science, 19(1), 83–103. https://doi.org/10.1287/mksc.19.1.83.15183
- Morozov, I., & Tuchman, A. (2024). Where does advertising content lead you? We created a bookstore to find out. Marketing Science, 43(5), 986–1001. https://doi.org/10.1287/mksc.2023.0138
- Morwitz, V. G., Greenleaf, E. A., & Johnson, E. J. (1998). Divide and prosper: Consumers' reactions to partitioned prices. Journal of Marketing Research, 35(4), 453–463. https://doi.org/10.1177/002224379803500404
- Payne, J. W., Bettman, J. R., & Johnson, E. J. (1993). The adaptive decision maker. Cambridge University Press. https://doi.org/10.1017/CBO9781139173933
- Reutskaja, E., Nagel, R., Camerer, C. F., & Rangel, A. (2011). Search dynamics in consumer choice under time pressure: An eye-tracking study. American Economic Review, 101(2), 900–926. https://doi.org/10.1257/aer.101.2.900
- Rock, R., Strauss, I., O'Reilly, T., & Mazzucato, M. (2024). Behind the clicks: Can Amazon allocate user attention as it pleases? Information Economics and Policy, 69, Article 101115. https://doi.org/10.1016/j.infoecopol.2024.101115
- Russo, J. E., & Leclerc, F. (1994). An eye-fixation analysis of choice processes for consumer nondurables. Journal of Consumer Research, 21(2), 274–290. https://doi.org/10.1086/209397
- Sims, C. A. (2003). Implications of rational inattention. Journal of Monetary Economics, 50(3), 665–690. https://doi.org/10.1016/S0304-3932(03)00029-1
- Stigler, G. J. (1961). The economics of information. Journal of Political Economy, 69(3), 213–225. https://doi.org/10.1086/258464
- Weitzman, M. L. (1979). Optimal search for the best alternative. Econometrica, 47(3), 641–654. https://doi.org/10.2307/1910412
- Zheng, S., Tong, S., Fang, S., Gopal, A., Li, X., & Jiang, Q. (2026). Priming in search: A large-scale field experiment on the impact of popular search genres in mobile shopping. Information Systems Research. Advance online publication, April 30, 2026. https://doi.org/10.1287/isre.2024.1041