Trademark litigation increasingly extends beyond the moment of purchase to post-sale confusion, in which a bystander, not a purchaser, misattributes the source, sponsorship, or affiliation of a product observed in ordinary use. Courts and litigants routinely rely on consumer surveys modeled on point-of-sale paradigms, such as Eveready and Squirt formats, yet these methods were designed to examine confusion proximal to the moment of purchase (i.e., conditions that differ substantially from a naïve observer’s fleeting encounter with a product in public). This paper argues that transplanting point-of-sale methodology into post-sale contexts, without commensurate adjustment, risks producing survey evidence that lacks ecological validity (i.e., the correspondence between a research procedure and the marketplace event it purports to represent). Drawing on legal theory underlying post-sale confusion and on literatures spanning perception, cognitive psychology, consumer behavior, and survey methodology, this paper develops a correspondence matrix comprising six dimensions: Universe, Stimulus, Cognitive, Temporal, Comparative, and Behavioral Correspondence. Moreover, the framework is used to contrast point-of-sale (“traditional”) and post-sale contexts, offering researchers, litigants, and courts a structured, theory-grounded basis for evaluating confusion surveys conducted for post-sale confusion and for weighing their evidentiary value.
This is a substantially reduced version of the actual white paper. If interested, please contact douglas.olsen@novetics.com for the most recent extended version.
Consumer Confusion and Trademark Infringement
Lanham Act and Trademark Infringement
In 1946 Congress enacted the Lanham Act to protect consumers from deception, preserve the goodwill established by trademark owners, and facilitate fair competition (15 U.S.C. §§ 1051–1127). During the first several decades after enactment, trademark infringement focused almost exclusively on point-of-sale confusion (i.e., at the time of purchase, there is confusion about whether the junior mark’s goods or services originate with, are sponsored by, or are affiliated with the senior mark), consistent with the conventional view that a trademark’s primary role is to identify source at the point of sale.
The Lanham Act uses the phrase “likely to cause confusion, or to cause mistake, or to deceive” (15 U.S.C. § 1114), but it does not specify when confusion must occur or whose confusion is legally relevant. This lack of definition, combined with marketplace changes such as the growth of national advertising and digital retailing, encouraged a broader view that confusion may arise in contexts beyond the purchase transaction itself (Beebe, 2006; Dinwoodie & Janis, 2024). Courts responded by recognizing additional theories of consumer confusion: reverse confusion, initial-interest confusion, and post-sale confusion. These are outlined in Exhibit 1.
Exhibit 1
Types of Consumer Confusion
| Type | Definition | Example |
| Point-of-Sale (Traditional) Confusion | Occurs when consumers are likely to believe, at the time of purchase, that the junior mark’s goods originate with, are sponsored by, or are affiliated with the senior mark (i.e., the actual trademark owner). This is also referred to as “traditional confusion.” | A person purchases a pair of Jay-Ban sunglasses on a digital retail platform, under the belief that they were made by, sponsored by, or affiliated with Ray-Ban due to similar appearance and/or branding. |
| Initial-Interest Confusion | Occurs when a similar trademark captures consumer attention through mistaken source, sponsorship, or affiliation. Even if the consumer is not confused at the time of purchase, the alleged harm is the diversion of consumer attention during the search process. | A consumer on a digital retail platform searching for sunglasses clicks on a pair of Jay-Ban sunglasses, believing they are made by Ray-Ban. The consumer is not confused when making the actual purchase (i.e., they know they are buying Jay-Ban, not Ray-Ban), but the initial confusion diverted attention away from actual Ray-Ban sunglasses. |
| Reverse Confusion | Occurs when consumers mistakenly believe that the senior user’s goods originate with, are sponsored by, or are affiliated with a well-known/established junior user. The alleged harm is that the junior user’s commercial dominance overwhelms the senior user’s trademark identity, diminishing its goodwill and ability to control its reputation. | A small sunglass manufacturer holds the trademark “Cooling Vision.” A much larger company such as Ray-Ban promotes a new product line as “Cooler Vision.” Consumers come to believe that the smaller company copied Ray-Ban. |
| Post-Sale Confusion | Occurs when an observer (not the purchaser) sees the junior mark in public and believes the source of manufacture, affiliation, or sponsorship is the senior mark. The alleged harm may involve a reduced likelihood of purchase of the senior mark or a reduction in prestige. | A passerby sees someone wearing a pair of sunglasses and mistakenly believes they are made by Ray-Ban. Perhaps months later, this person considers purchasing sunglasses, but does not consider Ray-Ban because of something associated with the previous viewing occasion (e.g., perceived poor quality). |
The focus of this paper is on examining post-sale confusion. However, discussing this type of confusion in isolation would likely diminish understanding of some of its distinctive elements, especially as they relate to survey practices that have become relatively entrenched when studying consumer confusion. Given this, efforts will be made to contrast it with point-of-sale confusion (also referred to as “traditional confusion”). Point-of-sale confusion in conventional trademark litigation occurs when a consumer makes a purchase, mistakenly believing that the product or service of the junior mark is manufactured by, sponsored by, or affiliated with the senior mark. This is usually what people think of when they consider trademark infringement-related confusion, and it is arguably the type of confusion most closely tied to the Lanham Act (see, e.g., Zhao, 2023).
Hence, in research examining point-of-sale confusion in a digital retail context, one or more pristine, professionally shot photographs, along with details of the product as they would appear in the marketplace (e.g., on a shopping site such as Amazon), might be appropriate. Due to the (typically) close temporal proximity of these images to the actual purchase experience, the linkage of confusion to material harm is straightforward. Specifically, people confuse the junior mark for the senior mark and shortly thereafter inadvertently purchase the junior mark.
On the other hand, post-sale confusion arises in the marketplace when an individual who is not confused (i.e., they know exactly what they are buying) purchases a product and then wears or uses it in a context where the public may observe it. Onlookers who see the item in use then mistakenly believe it is made by, sponsored by, or affiliated with the senior mark. The starting point for post-sale confusion itself, then, is a naïve observer encountering the product in a real-world context, perhaps on a bus, at a concert, in a social media post, and so forth.
Accordingly, the stimuli appropriate to a post-sale study would differ markedly from those used to study point-of-sale confusion. In a point-of-sale context, a handbag might have close-up detail of stitching, an unblemished appearance, and perhaps a complete array of photographs that capture every possible dimension. A post-sale “in use” context, by contrast, is messy. Consider what a handbag might look like when worn by a person wearing it in a grocery store. An arm and/or clothing may droop over the top of it, and maybe a child walking with the shopper may occlude another portion of it. The bag itself may appear discolored, dirty, or worn from use. The background is hardly pristine and context-free. The level of attention and involvement would be incidental, not intentional.
Post-sale confusion, then, differs from point-of-sale confusion in more than degree. Specifically, the context differs markedly, and the individual experiencing the confusion is not an active consumer but a naïve bystander. Compounding this, as demonstrated in Exhibit 2, the pathway to establishing material harm rests on a series of assumptions with no clear parallel in point-of-sale studies. Yet no established procedure exists for examining parameters associated with a post-sale confusion study, and no prior work has systematically identified the elements a study must contain to achieve ecological validity.
The goals of this paper are threefold: (1) to examine the legal theory underlying post-sale confusion, given that it may affect procedures and measures used to infer elements of any alleged harm; (2) to establish a multidimensional framework for evaluating ecological validity; and (3) to compare and contrast point-of-sale and post-sale confusion using this new framework.
Exhibit 2
Harm Resulting from Post-Sale Confusion
Legal Theory Associated With Post-Sale Confusion
Although this paper focuses primarily on survey methodology rather than a detailed consideration of trademark law, the legal principles underlying post-sale confusion necessitate consideration because they determine the types of empirical evidence required if one wants to establish material harm. In traditional point-of-sale confusion, where confusion and purchase are assumed to occur in close temporal proximity, it makes sense that mere confusion regarding source, affiliation, or sponsorship is closely linked to the possibility of material harm. However, as depicted in Exhibit 2, post-sale confusion involves a series of intermediate psychological and behavioral processes, and understanding the specific theory of harm is therefore essential to designing survey procedures that produce legally relevant evidence.
Sheff (2012) distinguishes between two principal forms of post-sale confusion. The first, bystander confusion, occurs when observers mistakenly believe that a product bearing the junior mark is associated with, sponsored by, or affiliated with the senior trademark owner, and a negative inference is then drawn regarding the senior mark’s product. The second, status confusion, concerns trademarks whose value derives largely from prestige, exclusivity, or symbolic meaning. Under this theory, imitation may erode the signaling value of a prestigious trademark, even when consumers fully understand that the junior mark’s product originates from a different manufacturer. Although conceptually distinct, both theories ultimately require direct evidence, inferences, or assumptions that observer perceptions produce measurable marketplace consequences.
Connecting Post-Sale Confusion to Actual Behavior
Although post-sale confusion has been widely accepted, it has also faced significant scholarly criticism. Where traditional trademark infringement emphasizes confusion affecting consumers’ purchasing decisions, post-sale confusion concerns the perceptions of third-party viewers after products are already in the marketplace. The key issue is a lack of clarity regarding how such confusion can be linked to the marketplace injuries the Lanham Act was designed to prevent (Bone, 2012; Lemley & McKenna, 2010; Tushnet, 2008).
Tushnet (2008) contends that the link to material harm has often rested on intuitive assumptions about how consumers will act, rather than on support from consumer research. Raustiala and Sprigman (2018) similarly argue that observer confusion should be legally significant only when it can be shown to influence marketplace behavior and produce measurable commercial consequences. Further, Zhao (2023) reviews post-sale confusion across different contexts and, like others, concludes that it does not necessarily result in harm to either the consumer or the senior mark. Zhao also observes that “courts have adopted a moral-based rationale that is primarily concerned with punishing the defendant for intending to profit from the plaintiff’s trademark, rather than preventing confusion or a loss of sales” (p. 928).
These criticisms target both theories of post-sale harm. Whether the alleged injury involves diminished quality perceptions or reduced prestige, observer confusion alone cannot establish material harm. Instead, for empirical fulfillment, both theories would seem to require evidence supporting the full causal sequence. First, observers must mistakenly attribute the junior mark’s product to the senior mark. Second, this confusion must alter perceptions of the senior mark, whether by negatively affecting perceived quality or by diminishing prestige and exclusivity. Third, these changed perceptions should affect marketplace actions, measurable through decreased purchase intent, lower willingness to pay, reduced likelihood of recommendation, diminished brand loyalty, or similar measures. Without evidence-based support for this sequence, claims of material harm remain inferential and may require considerable leaps of faith. One source of this empirical support could come from survey evidence.
Survey Use in Consumer Confusion Litigation
Courts evaluating survey evidence under Federal Rule of Evidence 702 and the principles articulated in Daubert v. Merrell Dow Pharmaceuticals, Inc. (1993) appropriately ask whether the appropriate universe has been identified, whether respondents represent that universe, whether control conditions are adequate, whether questionnaire wording is neutral and non-leading, whether demand characteristics and response biases have been minimized, and whether accepted survey administration and analytic procedures satisfy recognized standards of scientific reliability. These considerations have received extensive treatment within both the legal (Diamond, 2011; Diamond & Franklyn, 2014; Jacoby, 2013a, 2013b; McCarthy, 2025) and market research literatures (see, e.g., Burns & Veeck, 2026; Iacobucci & Churchill, 2023; Malhotra, 2022). Additionally, experimental design scholars have long emphasized representative sampling, construct validity, experimental controls, and threats to inference as determinants of empirical credibility (Cook & Campbell, 1979; Shadish et al., 2002).
Underlying all of this is the requirement of ecological validity, which may be defined by the extent to which a survey represents the marketplace situation it is meant to emulate. Consumer judgments are inherently contextual. Product evaluations, brand perceptions, memory, comparison processes, and purchase decisions are influenced by the physical, temporal, informational, and social environments within which consumers encounter marketplace stimuli. Findings obtained under highly controlled laboratory conditions may potentially identify statistically significant behavioral relationships, yet still provide limited evidence regarding consumer behavior when those contextual conditions differ substantially from those encountered in actual marketplaces. Work in consumer research emphasizes that experimental rigor depends not only on statistical control but also on how faithfully the elements of an experiment are operationalized to represent the marketplace phenomenon they are intended to model, avoiding artificial distinctions created solely by the research design (Hoyer et al., 2024; Lang, 2026; MacInnis et al., 2020; Morales et al., 2017; van Heerde et al., 2021; Viglia et al., 2021).
Trademark law has independently embraced an analogous principle stressing the need to adhere to marketplace conditions. In discussing the design of likelihood-of-confusion surveys, McCarthy explains, “the closer the survey methods mirror the situation in which the ordinary person would encounter the trademark, the greater the evidentiary weight of the survey results” (McCarthy, 2025, § 32:163). This observation has become a foundational principle of trademark survey methodology and is echoed by Swann, who writes, “The need to adhere to the real-world marketplace is as critical to the selection and presentation of stimuli as it is to the selection of an appropriate survey format” (Swann, 2022, p. 60). Pangan (2012) similarly observes that “the gravamen of a survey’s reliability is its ability to simulate ‘actual marketplace conditions in which consumers encounter the parties’ products,’ so as to be a reliability indicator of consumer confusion,” and that a survey’s failure to “approximate actual market conditions can provide grounds for inadmissibility” (p. 19), since a court may exclude a survey under Federal Rule of Evidence 702 when it is invalid or unreliable. Insisting on real-world correspondence also guards against plaintiffs using artificial stimuli that might inflate confusion estimates relative to actual marketplace conditions.
The Need for a Multidimensional Ecological Validity Framework
Although ecological validity must be present, without a concrete consideration of how this is operationalized, it may yield somewhat vague and amorphous demands that recognize its importance without clarifying how to achieve it. As an analogy, most would agree that “good health” is important, but without understanding how to measure or attain this, this lofty goal remains tenuous. Within the context of a consumer confusion survey, what elements of a purchase context need to be present to achieve ecological validity? How would one know if an important element were missing? At present, work on consumer confusion in trademark infringement applies ecological validity piecemeal. For example, elements such as the nature of the stimuli presented are not recklessly invoked, but still are done so in an ad hoc manner. There is no systematic, theory-based procedure for ensuring that assessments of ecological validity are thorough and comprehensive. Stated differently, although such efforts may yield stimuli representative of many environmental elements, without a structured approach, key details and features may be inadvertently left out.
The Marketplace Correspondence Framework (MCF)
The Multidimensional Marketplace Correspondence Framework identifies six complementary dimensions through which marketplace correspondence is established: (1) Universe Correspondence, (2) Stimulus Correspondence, (3) Cognitive Correspondence, (4) Temporal Correspondence, (5) Comparative Correspondence, and (6) Behavioral Correspondence. As presented in Exhibit 3, each dimension addresses a distinct relationship between the research procedure and the alleged infringing event, while collectively providing a comprehensive framework for evaluating the ecological validity of consumer confusion surveys. Exhibit 4 provides an overview of the dimensions, as well as some of the implications for point-of-sale versus post-sale confusion investigations.
Although conceptually distinct, these dimensions should not be treated as independent. For example, the temporal elements associated with the time available to view a stimulus may impact cognitive components, such as attentional resources brought to bear or the cognitive processes employed. Deficiencies in one dimension may therefore affect the evidentiary value of others, and meaningful evaluations of marketplace correspondence require considering the framework as an integrated whole rather than a checklist of isolated methodological characteristics.
The framework established here makes three related theoretical contributions. First, it synthesizes work from psychology, consumer behavior, and trademark law to demonstrate convergence on the principle that empirical evidence benefits from a meaningful correspondence between research procedures and the phenomena they seek to explain. Second, it operationalizes ecological validity through the concept of marketplace correspondence, transforming a broad methodological objective into a systematic framework that guides survey design and may be used to critique existing research. Third, there are clear implications with respect to how marketplace correspondence influences evidentiary strength.
Exhibit 3
Components of the Marketplace Correspondence Framework
Integrating Traditional Survey Methodology
The MCF is not an alternative to established principles of survey methodology, but a conceptual structure for understanding how those principles serve the broader objective of marketplace correspondence. Traditional methodological requirements related to appropriate universe selection, representative sampling, neutral questionnaire wording, suitable control conditions, accepted survey administration procedures, and the like, remain indispensable to scientifically credible consumer survey research. The framework neither modifies nor replaces these well-established standards. Rather, it integrates and builds on them, strengthening the conceptual basis for evaluating whether accepted methodological procedures have been applied in a way that appropriately represents the marketplace event at issue. By doing so, it provides a basis for examining all forms of alleged confusion, not just the two forms highlighted in this document.
Exhibit 4
Point-of-Sale vs. Post-Sale Requirements by Type of Correspondence
Component 1: Universe Correspondence
Universe Correspondence is among the key elements courts consider when ruling on the admissibility and weight of a survey. Courts assess whether the correct reference universe has been specified and whether the sample recruited, screened, and collected is consistent with that universe. Failing to specify the correct universe from the outset may be disqualifying, and as discussed below, the appropriate universe varies with the type of consumer confusion alleged. Of note, a survey can specify the correct universe and still fail Universe Correspondence at the point of execution if sufficient sampling frame error is present. A sample would be considered “overinclusive” if it omitted a requisite screening question (i.e., people who were not qualified to participate did). Conversely, it would be “underinclusive” if it imposed an extraneous screening element (i.e., people who should have been able to participate were not able to).
Potential Sampling Frame Error Issues
A sampling frame is the actual list, panel, or group from which a survey draws its respondents, and in practice it rarely lines up perfectly with the reference universe it is meant to represent. Sampling frame error occurs when there is variation between the population defined by the researcher and the nature of the group being sampled (see, e.g., Malhotra, 2022), thereby resulting in the systematic omission of members the survey is supposed to study, and/or the erroneous inclusion of those outside the universe.[1]
For example, suppose a point-of-sale confusion survey is conducted to examine the purchase of a luxury handbag, and assume the universe is correctly specified as individuals who purchase leather handbags over $500 in a digital retail context. An online survey might be appropriate, but it would fall short if the mechanism to recruit participants systematically excluded people who purchased online. Similarly, it would be flawed if it included people who would not purchase handbags in this price range or those who would never purchase such products online. Sampling frame error may arise from several sources, including: uncertainties introduced by the opacity of the specific panels the researcher draws on; screening questions used by the panel company prior to the survey; and screening questions within the survey itself. Sampling frame error can persist even when quota targets are met on a handful of observable variables (Diamond, 2011). Issues like this can arise due to: (1) uncertainties due to opacity of panel information; (2) screening questions used by the panel company; and/or (3) screening questions within the survey itself.
Point-of-Sale Universe Correspondence
In a point-of-sale (traditional) confusion study, the this is fairly well understood:
- the relevant universe is defined by current and potential purchasers of the defendant’s product (this is the junior mark, unless reverse confusion is claimed).
Post-Sale Universe Correspondence
The question of who should comprise the universe for a post-sale study is less settled.
- Post-sale confusion involves an observer who is not a purchaser at all. One view (Barber & Yaquinto, 2022, based on General Motors Co. v. Urban Gorilla LLC) would limit the universe to naïve observers who might ultimately purchase either the plaintiff’s or the defendant’s product.
- Even if one holds the sample to just those who are current and/or potential purchasers of both the junior and senior marks, it is not clear in what proportion each group should be represented. For example, should this be based on market share?
- If the target universe for the junior and senior marks is the same, this issue would be moot, but if different, it may consequential implications for who and how participants are recruited. The main point here is, simply saying “both should be included” may seem liberating at first, but it does not rid one of potential bias that may be introduced by assumptions underpinning the sampling mechanism chosen.
Component 2: Stimulus Correspondence
Stimulus Correspondence asks whether the research environment provides respondents with information that reasonably corresponds to what is available during the marketplace encounter where the alleged infringing event is said to occur. The theoretical foundation for this is well established within ecological approaches to perception, which hold that experiments should sample the environmental conditions to which researchers intend to generalize rather than rely on simplified laboratory displays divorced from their natural contexts (Brunswik, 1955; Gibson, 1979; Neisser, 1976). Contemporary research reinforces this conclusion, with studies of natural vision increasingly showing that perception in everyday environments differs systematically from perception under highly controlled laboratory conditions, and that these differences may materially affect the information available to observers (Felsen & Dan, 2005; Hasson et al., 2010; Nastase et al., 2020; Sonkusare et al., 2019).
Researchers have accordingly called for greater precision in specifying the real-world conditions to which experimental findings are meant to generalize, rather than broad assertions that a study is simply more realistic (Holleman et al., 2020; Morales et al., 2017). Stimulus Correspondence, as presented here, adopts this event-specific approach. It does not require researchers to reproduce every feature of the marketplace, but evaluates whether the respondent is presented with perceptual information that reasonably corresponds to what was available during the particular marketplace event the survey is intended to represent.
Presented stimuli could conceivably include auditory, tactile, olfactory, and visual information. Because many contemporary retail contexts involve digital commerce, and because internet-based surveys make variables easier to control and less expensive to administer, the stimuli presented to respondents in many contemporary confusion studies are typically visual (e.g., a photograph of the product in a digital retail context such as Amazon). To the extent this reasonably mirrors the context of the alleged infringing context, this may be sufficient. Nonetheless, researchers should also consider whether other senses are typically involved and whether they might affect the level of confusion observed within this context.
For the sake of both simplicity and exposition, let us stay within the visual domain. Understanding the visual information physically available is important in suggesting what material the individual has an opportunity to attend to, and whether or not the stimuli in the survey match well with the stimuli in the real world. Moreover, understanding what information is missing matters because incomplete information carries its own perceptual consequences. Perception does not simply register whatever information a stimulus happens to supply. When visual information is missing, our minds may draw on expectations, category knowledge, and prior experience to fill in the rest (Cohen & Basu, 1987; Loken, 2006; Loken et al., 2008). When a feature is occluded, distant, glimpsed briefly, or never attended to, the result is not necessarily experienced as a gap but often a complete experience. That is, an incomplete encounter with a product does not necessarily feel incomplete to the person having it, which is exactly why Stimulus Correspondence matters as much as it does.
Work on schema congruity points in the same direction, suggesting that even when a stimulus is at odds with an active product schema, people tend to interpret the discrepancy in a manner consistent with their existing schema rather than recognize it as something new (Mandler, 1982; Meyers-Levy & Tybout, 1989). Similarly, Peracchio and Meyers-Levy (1994) demonstrate that when a product photo in an advertisement is cropped ambiguously, consumers mentally fill in the missing part. The parallel to a post-sale context is apparent, suggesting a bystander who only partly sees a product does not necessarily experience an object with missing features.
With a view to understanding stimulus correspondence further, and staying within the visual domain, there are a number of pragmatic questions that need to be considered regarding the stimuli presented: (1) whether the stimuli are static or involve motion, (2) viewing distance, (3) viewing angle, (4) lighting, (5) product orientation, (6) product occlusion, (7) product condition, and (8) the surrounding context. This list is not meant to be exhaustive, but simply to provide a more detailed treatment of visual stimuli than trademark infringement surveys typically receive, and to highlight points on which point-of-sale and post-sale confusion studies may differ. Let us consider how point-of-sale and post-sale confusion may differ on these dimensions alone.
Point-of-Sale Stimulus Correspondence
Given the prevalence of internet-based surveys, this section approaches standard methods for point-of-sale confusion surveys from that perspective. Here are some characteristics that would define a typical internet-based survey:
- Product description page almost always features close-up shots of the product, perhaps supplemented by intermediate shots (e.g., close-up shots of a running shoe alongside intermediate shots of a person wearing them).
- Photographs are from multiple perspectives (e.g., for shoes: top, bottom, medial and lateral sides, front and back). Each shot is typically taken face-on or at a shallow angle.
- Images are professionally shot at high resolution under studio lighting that illuminates all features and minimizes shadows.
- A context shot may leave some aspect of the product out of view, but occlusion is largely moot given the multiple unimpeded views the professionally shot product otherwise provides.
- Products shown are exclusively new and unused, apart from occasional shots of a model who has just put them on; they are not creased, dirty, or worn.
- Vast majority of shots are captured against a neutral, typically white, background that isolates the product and keeps it front and center.
- Most images on a product description page are static. Even where a video shows the product in use, motion-related concerns about losing detailed imagery of the product are not particularly problematic, given the presence of other photographs in the carousel that present this detail.
Post-Sale Stimulus Correspondence
Post-sale confusion begins with an act of marketplace observation by a third party. A commuter may notice a handbag crossing a railway platform, a traveler may observe luggage moving through an airport concourse, or a spectator may glimpse branded footwear during a road race. In these contexts, the observer encounters the product as one element within a broader environment rather than as an isolated object presented for deliberate examination. Accordingly, the visual information available is often drastically different than the point-of-sale context just described.
- As observers move farther from a product, fine-grained (possibly source-identifying) features become progressively less available.
- Many products reveal their allegedly distinctive features only from particular viewpoints. Logos, stitching patterns, trim, ornamentation, or other claimed trade dress elements may appear on one side while remaining partially or completely concealed on another.
- In post-sale observation contexts, the observer does not control the orientation or the viewing angle. As such, the product orientation and angle should match those characterizing the reference marketplace event.
- Lighting and image fidelity likewise should be evaluated according to representational correspondence rather than technical quality. For example, studio lighting present for point-of-sale stimuli may reveal details unavailable under everyday illumination.
- Partial obstruction is often a defining characteristic of post-sale observation. A sleeve may obscure a portion of a watch. The fold present in a garment may obscure a design element. A person present may be standing in front of a handbag. In a context where motion is present, body movement continuously alters which product features remain visible. Removing these conditions may produce an accurate depiction of the product while simultaneously creating an inaccurate representation of the marketplace event.
- A feature visible in a point-of-sale context may also be completely unavailable when the product is in use. Examples include a label on underwear (i.e., a product not worn externally) or the lining inside a jacket if it is being worn.
- Post-sale products frequently display ordinary wear. This might include creasing, fading, dirt, stretching, scuffing, or deformation. Such changes may impact the ability to perceive otherwise distinctive features.
- Context also constitutes part of the perceptual event rather than merely decorative background. Products encountered in airports, workplaces, sporting events, or public streets have a more complex visual field than products displayed against neutral backgrounds. Objects, people, signage and the like add to the complexity of the scene and some elements may serve as primary draws of attention.
- Many post-sale encounters involve products that move through the observer’s visual field, rotate with body movement, or become visible only intermittently as viewpoints continuously change.
Component 3: Cognitive Correspondence
Observers do not simply react to the products around them. They allocate attention, interpret information, and regulate cognitive effort according to the goals they are pursuing, the competing demands of the surrounding activity, and the anticipated value of further processing. Two research procedures can expose respondents to identical marketplace information while producing fundamentally different psychological events. Cognitive Correspondence addresses whether a research procedure reproduces the attentional priorities, cognitive demands, task orientation, and information-processing strategies characterizing the marketplace event to which the study seeks to generalize.
Biased instructions or question wording, commonly referred to as demand effects, may implicitly or explicitly signal what the researcher is looking for. Sometimes the wording might be subtle and no ill-formed intent may be present, but if present these can be particularly problematic. Perhaps for this reason they have received a great deal of attention in survey design in a range of contexts (References). In other contexts, the issue is less about telegraphing what the “appropriate” answer or behavior might be, but rather about soliciting cognitive resources that would not otherwise be invoked in the context in question.
Cognitive resources are not distributed uniformly across everything that is visible. Instead, attention and information processing are shaped by the observer’s goals, expectations, competing task demands, and the anticipated value of further processing (Bettman, 1979; Bettman et al., 1998; Kahneman, 1973; Payne et al., 1993). In marketplace settings, shoppers’ visual attention follows their goals and a product’s relevance to the task at hand rather than simply its visual salience (Chandon et al., 2009; Milosavljevic et al., 2012). Consumers also adapt both the amount and form of processing to situational demands (Orquin & Loose, 2013; Schmitt, 2024). Much of this allocation occurs without deliberate scrutiny and is governed by situational goals rather than conscious attempts to inspect everything within view (Chartrand & Fitzsimons, 2011). The implication of this is that direction within the survey regarding what to process will have an impact, hence the importance of aligning such guidance with the characteristics of the marketplace context under study.
The eyewitness literature extends this principle in a way that is particularly relevant to post-sale bystander encounters. Hyman et al. (2018) argue that what governs an eyewitness’s later report is the individual’s attentional state during the original event, not simply whether the person or object was technically within view. Wulff and Hyman (2022) similarly found that participants whose attention was directed elsewhere were not only more likely to miss an event but were also more likely to select the wrong person when subsequently asked to identify the person involved. Importantly, they were more likely to make an incorrect identification than to report that they had not obtained a sufficient look.
This distinction has important implications for confusion research. A bystander may encounter an allegedly infringing product while simultaneously attending to a conversation, a phone, traffic, another person, or some other activity. A subsequent survey that isolates the product and directs the respondent’s attention toward its appearance can therefore create a substantially different cognitive event. The survey respondent is now being asked to inspect and evaluate features that the marketplace observer may never have attended to in the first place. A response obtained under those conditions may reflect expectations or inferences generated by the survey task rather than information actually processed during the marketplace encounter.
A survey can therefore faithfully reproduce the physical features of an alleged marketplace event while eliciting a different psychological one. Researchers should first specify the circumstances in which the alleged confusion is claimed to occur. This includes what consumers or bystanders are likely to be doing, what information is relevant to those goals, and where their attention would ordinarily be directed. Researchers can then evaluate the stimuli, instructions, exposure conditions, and questions according to how well they reproduce those cognitive conditions.
Defining this event requires identifying the activity in which the alleged observation occurs, the observer’s primary objective, the competing demands on cognitive resources, whether the product is expected or unexpected, the relevance of source identification to the activity, and whether the circumstances provide any independent reason to inspect or compare the product. These characteristics define the cognitive role the product ordinarily occupies within the marketplace event and establish the benchmark for judging Cognitive Correspondence. They are not refinements to be added after the survey is designed. Rather, they define the psychological event the survey is intended to reproduce. The relevant question is therefore not simply whether marketplace information is physically available. It is whether the reference cognitive event gives that information a reason to become cognitively significant.
Point-of-Sale Cognitive Correspondence
In a point-of-sale context, the goal within the survey is typically to emulate the focused attention and cognitive processing required when engaged in shopping behavior.
- Given the nature of the task, a consumer comparing handbags before purchase may deliberately search for logos, materials, construction details, and other diagnostic features, because product evaluation is the purpose of the encounter.
- Hence, instructions prior to inspecting visual stimuli within a survey might reasonably lead with, “On the next screen you will see a display of a product within a retail setting. Please review this information as you normally would on a retail site of this nature.” As such, the task and real-world context would be in reasonable alignment.
Post-Sale Cognitive Correspondence
Consider the same handbag encountered post-sale, perhaps while another individual walks through an airport, sits in a meeting, or waits in a restaurant.
- The observer’s cognitive relationship to a given object differs substantially because the goals organizing attention and processing have changed. Instructions for any survey must ensure that any attention and cognitive processing of the product in question (e.g., a handbag) would be incidental.
- Achieving the appropriate attentional balance is nontrivial. Clearly, though, a post-sale survey that instructs respondents to inspect a product as part of its procedure would be grossly errant.
Component 4: Temporal Correspondence
Marketplace behavior unfolds through time. Consumers observe products for finite periods, retain information over varying intervals, and form judgments under temporal conditions established by the marketplace encounter itself. Temporal Correspondence concerns whether the research procedure reproduces this temporal structure. This matters because perception and cognition necessarily occur within time. Consumers cannot observe information after the marketplace opportunity has ended, and later judgments must rely on whatever information remains mentally available to them at that point. Temporal Correspondence therefore builds on cognitive correspondence by placing additional parameters that need to be considered for any study examining consumer confusion. Because marketplace encounters differ in their temporal structure, a procedure that accurately reproduces one consumer experience may materially misrepresent another. Two components requiring consideration are exposure time and retention interval.
Exposure Duration
This concerns how long a person can view the stimulus in question, for example, the time spent reviewing a product on an Amazon webpage, or the length of time viewing a handbag worn by another person in a grocery store. It stands to reason that consumer judgments are shaped not only by the information available but also by the opportunity to acquire it. It has long been established that information acquisition is inherently constrained by the temporal and environmental conditions of the marketplace encounter (Bettman, 1979; Bettman et al., 1998; Payne et al., 1993). Contemporary consumer research likewise treats attention and information acquisition as adaptive responses to the demands and constraints of the consumption environment rather than as processes operating under ideal observational conditions (Huddleston et al., 2018; Perfecto & O’Donnell, 2026; Robinson et al., 2026). Consumers accordingly adapt their search strategies to temporal constraints rather than maximizing information acquisition in every circumstance (Gigerenzer & Gaissmaier, 2011). As Robinson et al. (2026) argue more broadly, temporal characteristics are not incidental features of consumer behavior but fundamental to how consumption experiences are structured and interpreted. Observers direct attention toward information that is useful within the limited opportunities afforded by the environment rather than attempting exhaustive inspection (Tatler et al., 2011; Wedel & Pieters, 2008). Hence, the selection of the exposure time is consequential.
Retention Interval
This is the time between the end of stimulus exposure and the point at which the respondent is asked to recall or judge information from that exposure. Cognitive psychology increasingly shows that remembered events change as temporal distance increases. Elapsed time affects the accessibility of information, its contextual specificity, and the balance between detailed recollection and generalized representation (Howard, 2018; Sekeres et al., 2018; Xue, 2018; Yonelinas et al., 2019). Consumer research similarly finds that delayed evaluations rest on informational conditions different from those supporting immediate judgments. As specific details become less accessible, consumers may reconstruct product information from broader impressions (Kim et al., 2009; Montgomery & Unnava, 2009; Si & Dai, 2022; Walters & Hershfield, 2020). Research on everyday brand memory further demonstrates how little visual detail may survive routine, repeated exposure to a familiar logo even without a delay, much less the additional delay separating exposure from subsequent recall or recognition (Mayas et al., 2024; Whatley et al., 2023).
Interference Between Stimulus Exposure and Memory Measurement. Consideration of the retention interval would be incomplete without considering the nature of the interference that occurs during it. An interference task, sometimes called a distractor or filler task, is an unrelated activity inserted between exposure to a stimulus and a subsequent memory-based question. Among other things, a person could be asked to watch a video, do mathematical calculations, or write prose. Requiring respondents to complete an intervening task shifts the measure toward information that survives in longer-term memory. Without some means of preventing rehearsal, a measure described as memory may instead capture immediate maintenance rather than storage (Lynch & Srull, 1982). Applied research reinforces this point. Burke and Srull (1988) and Kent and Allen (1994) both found that competing brand information presented between an advertisement and a later memory test measurably reduced what consumers reported about the original advertisement relative to an undelayed test. Together, this literature supports incorporating a delay and some form of intervening activity into procedures intended to approximate real consumer memory.
Some degree of caution must also be exercised when selecting an interference task, ideally finding a task that has a counterpart in the marketplace being studied. For example, performing mathematical tasks in the period between exposure and the memory task might achieve the goal of erasing material in working memory, but would introduce its own artificial element. As research has noted, added procedural elements should be calibrated to the particular real-world event a study seeks to represent rather than assumed to improve validity simply because they are established research practices (Holleman et al., 2020; Morales et al., 2017).
Point-of-Sale Temporal Correspondence
Given the nature of the point-of-sale context, the following are noteworthy:
- Exposure time in point-of-sale consumer confusion surveys is often left to the respondent, which accommodates individual differences.
- The retention interval is often short because the purchase of the allegedly infringing product will generally occur soon after the consumer views it. That is, the nature of the transaction itself generally requires the consumer to be in close temporal proximity to the allegedly infringing product when making the purchase decision, either physically in a brick-and-mortar store or on the relevant product description page when placing the product in an online shopping cart. For the same reason, an interference task may be unnecessary in most point-of-sale contexts.
Post-Sale Temporal Correspondence
For a post-sale context, the context is remarkably different:
- Exposure may be extremely brief (e.g., a traveler may briefly notice another passenger’s luggage while boarding an aircraft) or considerably longer (e.g., sitting beside someone using an allegedly infringing brand of pen throughout a flight). Variation in exposure time is therefore likely to be substantially greater in post-sale versus point-of-sale contexts. Therefore, it is particularly important to establish a defensible exposure duration before conducting the survey (i.e., what is a typical context in which this is seen, and how long is exposure in this context). Even here, the high level of variability leaves a researcher subject to concerns that the research context may have been chosen to suit the needs of the research itself (e.g., intentionally selecting a context where the exposure time would be longer).
- The retention interval is also likely to be substantially longer in a post-sale context. Assuming an observer actually notices the allegedly infringing product and forms a mistaken impression regarding its source, the interval between the observation and the point at which that impression influences subsequent behavior, such as a later purchase, may be measured in months, or even years. As such, many observers may not retain the experience at all.
- Given the potentially extended retention interval, substantially more interference is also likely to occur in a post-sale context than at point of sale. Capturing this interference within a survey is difficult and may, in some circumstances, be impossible. For example, it is not clear how one would approximate the type and amount of interference that would occur during a much longer retention interval, such as the ten months between observing a product and making a subsequent purchase.
Component 5: Comparative Correspondence
Net consumer confusion is a function of the confusion level generated by the alleged infringing mark minus the confusion level generated by the control. A control, by definition, should be identical to the test stimulus (e.g., the junior mark) except for the specific feature(s) under investigation (Cook & Campbell, 1979; Ericksen & Pittaoulis, 2014; Shadish et al., 2002). If the concern is that the junior mark has a logo that contains a confusingly similar purple triangle, the control should be identical to the junior mark but without the purple triangle (e.g., removed, or replaced with a line of a different shape and color). A control that is otherwise different from the infringing mark (e.g., by removing non-infringing characteristics) so that it generates almost no confusion of its own will inflate the reported confusion attributed to the junior mark.
An underpinning assumption here is that respondents are assigned to the test and control conditions at random, so that the two groups differ only in the stimulus each received. With the advent of digital technologies for image creation and the generally accepted use of online surveys, such randomization is commonplace and does not pose much of a challenge from a logistical perspective. As will be discussed when considering the post-sale environment, moving things outside of the digital realm, though, could pose more of a challenge from a control perspective.
Point-of-Sale Comparative Control
Consider the following regarding a point-of-sale context:
- Comparative Correspondence requires that the test and control conditions reproduce the same purchasing environment and differ only with respect to the allegedly infringing feature(s).
- This principle governs not just the physical appearance of the stimuli but the broader informational context in which they are encountered (e.g., additional information on a product description page).
- Though not trivial, the nature of the task is reasonably straightforward.
Post-Sale Comparative Control
The same experimental requirement applies in a post-sale study: the test and control conditions should differ only with respect to the allegedly infringing feature or features.
- Applying that principle is potentially more demanding here, though, because post-sale observers generally encounter products outside the controlled environment, which may be characterized as rich in product information.
- To the extent that one can represent such a context digitally (e.g., a concert), this burden may be lifted, assuming that product imagery can be substituted between test and control.
- Of course, the parsimony achieved through the digital representation may come at a cost. Whereas one could argue that browsing for items on a website (i.e., the type of context for point-of-sale confusion) is primarily visual, this argument is harder to make in a “real-world” environment where scents, sounds, and other distractions outside the visual domain may be present.
- Conversely, moving observation into the “real-world” (e.g., having a person go to a concert) not only presents logistical issues in setting things up but may also make random assignment to conditions much more challenging.
- An additional control issue emerges for post-sale confusion, should the underlying legal theory suggest that damage is created due to a perception of diminished quality associated with the junior mark relative to the senior mark. In this case, the senior mark is also required as a control in order to judge whether material harm is likely.
Component 6: Behavioral Correspondence
Behavioral Correspondence considers whether the outcomes selected for measurement correspond to the marketplace behavior on which the researcher’s conclusions depend. This does not mean actual purchase behavior is the only relevant measure, but it does suggest that the selected construct should correspond to the behavior the conclusion actually requires. A measure may possess excellent psychometric properties yet provide limited ecological evidence when used to support conclusions about a particular behavior.
Within consumer research, researchers routinely substitute a subordinate construct (e.g., purchase intention or willingness-to-pay) for a behavior they actually care about (e.g., a completed purchase). The substitution is often necessary in survey contexts, given that having people actually complete a sale within the survey would not be tenable in the vast majority of contexts. So, the issue comes down to understanding what level of attenuation occurs as one moves from a statement regarding purchase intention through to the time a person actually takes out their wallet and makes the purchase.
The concern in bridging a subordinate behavior with a superordinate one is longstanding. In their examination of the attitude-behavior relationship, Ajzen and Fishbein (1977) put forward the compatibility principle, suggesting that the linkage between attitudes and behaviors will be greater as the specificity in the attitude measured increases. For example, if you are trying to predict blood donation, the attitude-behavior linkage will be stronger if you ask, “do you intend to donate blood at a Red Cross location on Tuesday of next week?” rather than “do you plan on giving blood?” This has been empirically backed by the meta-analysis of Sheppard et al. (1988), and further understanding has been gained through examining different contexts (see, e.g., Jamieson & Bass, 1989; Morwitz et al., 2007). However, research by Sheeran (2002) cautioned that even well-specified intentions may still perform poorly at predicting what appears to be associated behavior, and large variations continue to be found (Morwitz & Munz, 2021; Sheeran & Webb, 2016). The implication is not that researchers should abandon subordinate constructs, but that the inferential distance in predicting an end-state activity likely increases with the number of assumptions underpinning the attitude-behavior chain.
Point-of-Sale Behavioral Correspondence
Consider the following regarding the point-of-sale context:
- The prospective buyer encounters the allegedly confusing stimulus at or near the time of purchase. If the purchaser mistakes the product’s source, sponsorship, or affiliation when deciding whether to buy it, the confusion is now embedded as part of the decision-making process itself.
- As such, within a point-of-sale confusion study, because the respondent is the prospective purchaser and the decision would be temporally proximate to the confusion, relatively few inferential steps are needed to connect confusion to material harm.
Post-Sale Behavioral Correspondence
Because post-sale confusion relies on a naïve observer and a series of assumptions to establish some form of harm, there are a number of assumptions and inferential steps required to connect the dots.
- Whereas just measuring net confusion in a point-of-sale study makes sense for the reasons noted previously, because post-sale deals with a naïve observer in a non-purchasing context, simply collecting a measure of net confusion would seem to be insufficient and thereby problematic. There is virtually no connection between the level of confusion present and the possibility of an errant decision in that moment. Any harm is separated by both time and a number of assumptions.
- Perhaps one of the more implicit assumptions is that confusion in this context must lead to something bad. However, what if a person was confused regarding the source of the junior mark, but liked the appearance of the product and was inspired to purchase the senior mark? Perhaps after seeing it the confused individual deduced that this brand must be in vogue and decided to purchase from the senior mark because of its popularity from a fashion perspective.
- The marketplace consequence in a post-sale confusion claim relies on assumptions of exclusion from a future consideration set, reduced willingness to pay, diminished exclusivity and/or tarnished prestige.
- Behavioral Correspondence is therefore especially consequential when research advances a multistage causal account: (1) people will see the junior mark in use; (2) they will then confuse the junior mark with the senior mark; (3) a belief will exist that the quality is of inferior quality and/or lower prestige; and (4) beliefs about quality and/or prestige will impact purchase behavior at a later point in time.
- When a theory of harm contains multiple behavioral stages, the survey should specify which stages are measured, which are inferred, and what evidence supports the proposed causal chain.
General Discussion
The preceding sections make a straightforward argument in an area where methodology has largely proceeded without one. Post-sale confusion cannot be assessed with point-of-sale survey instruments that have simply been relabeled. The observer is not the purchaser, the exposure is not deliberate, the timing is not compressed into the moments before a transaction, and the behavior at stake is not a completed sale but some more distant marketplace consequence. Each of these differences carries methodological weight, and the Marketplace Correspondence Framework (MCF) developed here makes that weight explicit, dimension by dimension, rather than leaving ecological validity as a single global judgment rated along some continuum.
The framework does not require researchers to invent new constructs or abandon the Eveready and Squirt traditions that have become familiar in trademark litigation. Instead, it asks that each methodological choice, including the universe sampled, the stimuli selected, the instructions provided, the duration of exposure and retention, the construction of the control condition, and the dependent variables measured, be tied to the reference marketplace event in which the alleged confusion is said to occur. The discipline to consider these matters before a survey is developed and administered may be especially important in post-sale research, because familiar point-of-sale conventions are relatively easy to administer online and have a long history of acceptance in litigation. Familiarity, however, does not establish that those procedures adequately represent the bystander encounter at issue or the consequences claimed to follow from it.
Implications for Litigants, Experts, and Courts
For those seeking to establish post-sale confusion, the implication is not that survey evidence is futile, but that the evidentiary burden is considerably heavier than is often recognized. A survey offered in support of such a claim should, at a minimum, specify the reference marketplace event and justify the universe drawn from it, document how stimuli, instructions, exposure, and retention were calibrated to that event, construct controls that hold the surrounding context constant, and measure, rather than assume, the downstream beliefs on which the theory of harm depends. Consideration also needs to be given to including the senior mark as a control as a reference for any claims that the junior mark is harming the senior mark in this regard. Where a link in the causal chain cannot be measured, the survey should acknowledge that the link rests on inference.
For those evaluating such evidence, the framework offers a structured basis for inquiry. Rather than asking only whether a survey followed an accepted format, counsel, experts, and courts can ask whether each dimension of correspondence was addressed and what follows when it was not. This inquiry fits comfortably within the existing analysis of admissibility and weight under Federal Rule of Evidence 702, and within the long-standing principle that survey evidence gains weight as it approximates the conditions under which ordinary people encounter the marks (McCarthy, 2025; Swann, 2022).
Taken as a whole, the framework advances a relatively modest proposition with less modest consequences. The evidentiary value of a survey depends in meaningful part on how well the conditions of the research correspond to the marketplace event the survey is intended to represent, and that correspondence is better understood as the product of several separately assessable dimensions than as a single property of the research design. Point-of-sale confusion research has developed over decades of litigation and has produced conventions that approximate the purchase context reasonably well. Post-sale confusion research has not developed to the same extent, and the analysis presented here suggests that the gap reflects more than a lack of attention. It reflects the nature of the phenomenon itself.
[1]From a procedural perspective, small sample sizes and non-response bias (see, e.g., Groves & Peytcheva, 2008) do not automatically render a sample unrepresentative. A small sample may still include members who are representative of the universe, and non-response bias affects sampling frame error only when it systematically excludes specific characteristics of the universe that should otherwise be included. A related but separate question is whether the substance of respondents’ answers (e.g., their extremity) differs from what non-respondents would have provided.