{ "version": "3.1", "collectionStatus": "Demonstrations and proposed protocols; no pooled data", "lenses": [ { "id": "attention", "name": "Attention", "question": "What gets into awareness?", "terms": "Salience · interruption · novelty · motion · contrast · gaze · cognitive load", "body": "Look at what enters the foreground, what interrupts a task, and what becomes difficult to ignore. A shelf, a notification, and a headline can all arrange the order of attention." }, { "id": "perception", "name": "Perception", "question": "What does context change?", "terms": "Framing · sensory context · expectation · anchoring · cross-modal effects", "body": "Explore how the surroundings of an object or a message change its interpretation. Separate what changed in the stimulus from what changed in the judgement." }, { "id": "desire", "name": "Desire", "question": "Where does wanting come from?", "terms": "Scarcity · status · imitation · association · anticipation · identity", "body": "Examine the cues around wanting: who appears to want something, what it seems to represent, and whether availability changes its appeal." }, { "id": "choice", "name": "Choice", "question": "Who arranged the alternatives?", "terms": "Defaults · friction · ordering · decoys · complexity · commitment", "body": "Trace the options, their starting states, and the effort required to choose differently. Ask whether the person can see and meaningfully resist the intervention." }, { "id": "memory", "name": "Memory", "question": "What remains after the moment?", "terms": "Repetition · familiarity · narrative · peak-end effects · reconstruction", "body": "Compare what was presented, what felt familiar, and what was remembered. Remembered experience is an object of study in its own right." }, { "id": "adaptation", "name": "Adaptation", "question": "What happens when the environment learns?", "terms": "Personalisation · ranking · prediction · affect inference · AI", "body": "Follow the loop from observation to prediction, adaptation, and influence. Distinguish what a system actually knows from the authority its interface appears to claim." } ], "studies": [ { "id": "invisible-default", "number": "001", "title": "The invisible default", "deck": "The starting point is already part of the choice.", "kind": "Demonstration", "status": "Ready to try", "lens": "choice", "tags": [ "Interfaces", "Defaults", "Friction" ], "minutes": "1 min", "engine": "default", "observation": "A preference screen can arrive with one answer already selected.", "hypothesis": "A preselected option may be retained more often than an option that requires an explicit selection.", "mechanism": "Defaults may communicate a suggested norm, reduce effort, or go unnoticed. Those are competing explanations, not interchangeable conclusions.", "experiment": "Choose a reading cadence twice before the reveal. The first screen preselects one option; the second leaves both unselected, with the same wording and order. The cadence order varies between visits and stays fixed throughout the comparison and replay. The informed replay leaves both rounds unselected.", "result": "The reveal compares both selections and whether the initial default was retained. This is a within-visit comparison, not a control-group experiment. There are no pooled results.", "alternatives": "You may prefer that cadence, remember your first answer, seek consistency, notice the presentation change, or recognise the premise. Fixed order and repetition prevent attributing a difference to the default alone.", "realWorld": "Account preferences, checkout options, cookie interfaces, and enrolment forms.", "ethics": "A default can reduce effort or take advantage of inattention. Who benefits, what the choice commits someone to, and how easy it is to change all matter.", "everyday": "In a real service, a preselected choice may affect what you receive, share, or pay for. It may stay in place without being pointed out. Check the selected options before continuing, especially when the choice has consequences beyond the screen.", "countermeasure": "Clear the initial selection. State your preference before viewing the interface. Compare the options with equal effort.", "sources": [ "defaults", "defaultsMeta" ], "evidence": [ { "key": "defaults", "text": "Donation choices differed with the starting condition in Johnson and Goldstein’s studies. This makes the preselected answer here worth inspecting: leaving a setting alone can itself determine an outcome. The paper concerns donation agreement, not proof that this fictional reading preference—or actual transplant outcomes—will change." }, { "key": "defaultsMeta", "text": "Across 58 studies, defaults generally shifted choices, but their effects varied substantially. The review considers perceived endorsement and the status quo among the explanations. It supports asking why a starting state matters here, while cautioning against treating every retained default as inattention or predicting an individual response." } ], "entryTitle": "Choose your reading rhythm", "entryDeck": "Make two small choices, then look back at them.", "entryContext": "A fictional reading preference" }, { "id": "price-context", "number": "002", "title": "The price that changes nothing", "deck": "Judge four offers. Then compare what changed.", "kind": "Demonstration", "status": "Ready to try", "lens": "perception", "tags": [ "Commerce", "Anchoring", "Scarcity" ], "minutes": "3 min", "engine": "price", "observation": "A price can be accompanied by a crossed-out reference, a popularity badge, or a stock message.", "hypothesis": "Changing only this surrounding claim may change stated value, urgency, or confidence.", "mechanism": "A reference price supplies a comparison point. Scarcity and popularity are different candidate mechanisms and are presented as separate variants.", "experiment": "Judge four offers one at a time: a crossed-out €79 reference beside €29 first, then a popularity badge, a stock claim, and finally €29 alone. Give value, urgency, and confidence ratings for each before the explanation. The notebook and €29 price stay the same. The replay removes all surrounding claims from four offers.", "result": "The reveal compares all submitted ratings with the exact offer shown. Differences or identical ratings describe this visit; they do not establish an effect or immunity. An early reveal preserves completed ratings and marks unrated offers explicitly.", "alternatives": "Order, repeated exposure to the notebook, remembering earlier ratings, suspicion, prior prices, the study title, and expectations about the task may affect the comparison. The fixed order cannot isolate a causal effect within one visit. Starting with promoted offers is a presentation choice; we have not established that it delays recognition of the pattern.", "realWorld": "Retail labels, sale pages, restaurant menus, travel listings, and subscription comparisons.", "ethics": "The price reference, stock claim, and popularity label are invented within this disclosed demonstration. In ordinary commerce, their truth and relevance are central to judging the intervention.", "everyday": "An everyday offer can lead to a real purchase, and you may never receive an explanation of how it was presented. Compare the actual total with what the product is worth for your purpose. Look for a verifiable basis for a discount or stock claim rather than treating the label itself as evidence.", "countermeasure": "Cover the reference and badges. Decide what the object is worth for your purpose before comparing offers.", "sources": [ "anchors", "coherentAnchors", "wine" ], "evidence": [ { "key": "anchors", "text": "Tversky and Kahneman described how an initial reference can influence later estimates. A crossed-out price gives the notebook just such a comparison point. This is the background for the anchoring question; the paper does not validate our particular discount, stock message, or popularity badge." }, { "key": "coherentAnchors", "text": "These experiments found that arbitrary starting points could influence initial valuations even when later choices followed a coherent pattern. That matters when several offers feel consistently judged: consistency alone does not show that the reference point was irrelevant. Their procedures differ from the fixed sequence of offers here." }, { "key": "wine", "text": "Participants tasted the same wines under different price labels. Higher labels increased reported pleasantness in the study, alongside changes in a brain signal associated with experienced pleasantness. Price context can reach beyond a bargain calculation; this does not establish a neural effect from viewing our notebook." } ], "entryTitle": "Four price offers", "entryDeck": "Judge four offers, then compare your responses.", "entryContext": "Fictional notebook offers" }, { "id": "where-did-you-look", "number": "003", "title": "Where did you look?", "deck": "The same information can enter awareness in a different order.", "kind": "Demonstration", "status": "Ready to try", "lens": "attention", "tags": [ "Visual perception", "Salience", "Recall" ], "minutes": "1 min", "engine": "salience", "observation": "Two research notes with equivalent descriptions and differently phrased titles can receive very different visual emphasis.", "hypothesis": "A more visually prominent option may be selected or recalled more often.", "mechanism": "Contrast can contribute to salience, alongside task goals, expectations, position, and meaning.", "experiment": "Choose a research card in two consecutive presentations. The first highlights one card; the second gives both equal emphasis. Each choice advances automatically. The titles use different phrasing for the same topic, with the same description and reading time. Titles and card order stay fixed across both rounds and replay. Optional recall follows both choices, then one reveal compares them.", "result": "The reveal preserves both choices and optional recall. It does not know where your eyes looked or whether emphasis caused a decision. A second response is not an independent control observation.", "alternatives": "Preference for one title’s phrasing, reading order, colour preference, chance, remembering the first choice, seeking consistency, noticing the contrast, or guessing the hypothesis. The titles are intended to have similar appeal, but that equivalence has not been measured. The fixed sequence cannot isolate an emphasis effect.", "realWorld": "Supermarket shelves, streaming rows, news pages, app stores, and checkout actions.", "ethics": "Visual emphasis can support a person’s task or steer them toward another party’s objective. Visibility of alternatives and freedom to choose remain relevant.", "everyday": "Outside this demonstration, there may be no second presentation with equal emphasis. A prominent option can lead to a real decision while alternatives remain easy to overlook. Pause to inspect the other options and their terms before treating the most noticeable one as the best fit.", "countermeasure": "Equalise the presentation and compare the underlying descriptions. Replay after disclosure cannot establish that awareness has made you immune.", "sources": [ "salience", "selectivity" ], "evidence": [ { "key": "salience", "text": "Itti and Koch modelled how visual features can contribute to the selection of conspicuous locations. The model helps explain why contrast is a candidate influence on these cards. It does not tell us where a visitor looked: a click or recalled headline is not an eye-movement measurement." }, { "key": "selectivity", "text": "In a controlled visual-search task, an irrelevant colour singleton slowed some searches for a shape target. The relative distinctiveness of colour and form mattered. This gives a more precise basis for the emphasis comparison: conspicuous information can compete with the task, but a bright element does not invariably win attention." } ], "entryTitle": "Two research notes", "entryDeck": "Choose what you would open in two short rounds.", "entryContext": "Two research notes" }, { "id": "choice-architecture", "number": "004", "title": "Choice architecture", "deck": "Six possibilities. Then twenty-six. What changes?", "kind": "Demonstration", "status": "Ready to try", "lens": "choice", "tags": [ "Interfaces", "Complexity", "Decision confidence" ], "minutes": "3 min", "engine": "choice", "observation": "Larger assortments add possibilities while changing the work of comparison.", "hypothesis": "Assortment size may change deferral, confidence, and satisfaction, depending on the task and the person.", "mechanism": "Search effort, preference uncertainty, and trade-offs are possible contributors. More choice is not universally worse.", "experiment": "Choose a fictional notebook colour from six and twenty-six options in randomised order. The six are randomly selected from the twenty-six, and their relative order is preserved between presentations. The replay uses the same six twice. Rate confidence and satisfaction after each choice; deferral is always available.", "result": "The reveal compares your two choices and ratings. Order and learning effects remain, even when presentation order is randomised.", "alternatives": "A favourite colour exists only in the larger set, screen size, motor effort, practice, fatigue, or preference for variety.", "realWorld": "Menus, plan comparisons, product catalogues, recommendation grids, and configuration screens.", "ethics": "Reducing options can clarify a decision or remove valuable alternatives. The purpose and availability of the omitted options matter.", "everyday": "In everyday decisions, you may not be able to replay the choice with a smaller set, and postponing it may also have consequences. Identify what matters to you before comparing options, then check that simplifying the list has not removed something you need.", "countermeasure": "Set two or three criteria before opening the full assortment. Narrow the set deliberately and keep a route back to the rest.", "sources": [ "assortment", "choice" ], "evidence": [ { "key": "assortment", "text": "Across shopping and laboratory tasks, Iyengar and Lepper found situations in which extensive choice reduced motivation compared with smaller assortments. Their work motivates measuring deferral and satisfaction alongside selection. Our colour task changes the materials and procedure, so it cannot reproduce their result simply by offering more options." }, { "key": "choice", "text": "This meta-analysis found an average choice-overload effect close to zero, with considerable variation between studies. That is an essential counterweight to the familiar “less is more” story. Your preferences, the available options, and the work of comparing them matter; a larger grid need not reduce confidence." } ], "entryTitle": "Choose a notebook colour", "entryDeck": "Make a choice in each of two assortments.", "entryContext": "Fictional notebook colours" }, { "id": "manufactured-popularity", "number": "005", "title": "Manufactured popularity", "deck": "What if the crowd is part of the interface?", "kind": "Demonstration", "status": "Ready to try", "lens": "desire", "tags": [ "Social influence", "Ratings", "Credibility" ], "minutes": "2 min", "engine": "social", "observation": "Visible engagement can accompany content before a person evaluates it.", "hypothesis": "A popularity cue may change stated quality, credibility, or willingness to share identical content.", "mechanism": "A crowd cue may be read as information, social approval, or a signal of relevance.", "experiment": "Rate the same passage three times: 18,400 likes first, then 23 likes, then no count. The reveal follows all three responses. The informed replay removes every count.", "result": "The reveal compares each set of ratings with its actual count or absent count. The counts are fictional stimuli, not visitor activity. This sequence provides no causal estimate.", "alternatives": "Distrust of counts, interest in the content, order, repetition, remembered ratings, expectations, and unfamiliarity with the platform may affect the comparison.", "realWorld": "Reviews, feeds, trending lists, download counters, queues, and best-seller labels.", "ethics": "Popularity can help discover useful material. Fabricated or selectively displayed popularity can misrepresent the evidence available to the viewer.", "everyday": "On an everyday page, you may receive no explanation of who was counted or what a number represents. Even a real engagement count does not establish quality, accuracy, or suitability for you. Inspect the content and the basis of the count before relying on its apparent popularity.", "countermeasure": "Form an initial judgement with the count covered. Ask who was counted, how they were selected, and what the number actually measures.", "sources": [ "social", "culturalMarket" ], "evidence": [ { "key": "social", "text": "A randomised field experiment altered an initial rating and observed later votes. Prior social information could bias subsequent evaluations, making the first visible signal part of the process. This grounds the question behind our invented like counts, without supplying an expected effect size for this passage or visitor." }, { "key": "culturalMarket", "text": "An experimental music market compared independent choices with choices informed by others’ downloads. Social influence increased both inequality and unpredictability of success. The implication extends beyond a single like count: visible history can help create the popularity it appears merely to report. Song quality still mattered." } ], "entryTitle": "A field note, three times", "entryDeck": "Read a short passage and give your assessment.", "entryContext": "A short passage" }, { "id": "room-changes-product", "number": "006", "title": "The room changes the product", "deck": "When the surroundings become part of the thing itself.", "kind": "Field study", "status": "Protocol", "lens": "perception", "tags": [ "Physical environment", "Sensory influence", "Context" ], "minutes": "Field protocol", "engine": null, "observation": "People encounter an object together with lighting, sound, materials, spacing, and other surroundings.", "hypothesis": "Changing one environmental variable may change the rating of an otherwise unchanged object or experience.", "mechanism": "Expectation and sensory context are candidate explanations. This protocol does not assume that every sensory intervention produces an effect.", "experiment": "Use the same object and room. Change one predeclared variable, such as lighting, while keeping placement and instructions fixed. Counterbalance order across consenting participants and record device or room conditions. Compare object ratings and environment ratings separately.", "result": "The four generated scenes are a visual illustration based on the homepage object. They are not field observations or a controlled comparison. No participant data have been collected in this implementation.", "alternatives": "Adaptation, order, temperature drift, social expectations, visible equipment, or a physical change to the object itself.", "realWorld": "Restaurants, retail spaces, packaging displays, museums, listening rooms, and hospitality.", "ethics": "Agree the scope of participation before changing the environment. Make the intervention reversible and disclose exactly what changed. Account for sensory comfort and individual access needs.", "countermeasure": "Repeat the rating in a neutral setting. Ask separately how the object and the environment felt.", "sources": [ "musicTaste", "wine" ], "evidence": [ { "key": "musicTaste", "text": "Separate groups’ descriptions of wine reflected the emotional character of the music playing while they drank. This directly motivates asking whether a room changes judgements of the object itself. It is evidence about a particular auditory and tasting setup, not proof that the four illustrated displays change anyone’s valuation." }, { "key": "wine", "text": "Holding wine constant while varying its stated price helped separate the drink from expectations surrounding it. That is the useful design principle for this field protocol: preserve the actual object and manipulate a defined contextual cue. Our generated scenes change several features together, so they remain an illustration." } ] }, { "id": "algorithmic-repetition", "number": "007", "title": "Algorithmic repetition", "deck": "Familiarity can arrive before a reason for liking.", "kind": "Demonstration", "status": "Ready to try", "lens": "memory", "tags": [ "Algorithmic environments", "Repetition", "Familiarity" ], "minutes": "2 min", "engine": "exposure", "observation": "A name can become familiar through repeated appearances beside unrelated tasks.", "hypothesis": "Repeated exposure may influence later recognition or preference, but the two should be measured separately.", "mechanism": "Exposure and familiarity are candidate contributors. A short task cannot establish how an actual recommender system shapes a person over time.", "experiment": "Complete four simple comparisons. A fictional label repeats in the surrounding display. Choose a preferred label before answering a separate familiarity question. Label order varies between visits and stays fixed for both questions and replay. Replay removes the repeated label.", "result": "The reveal reconstructs the exposures and preserves the preference submitted before the familiarity question, including an incomplete final response. There is no personalised ranking or inferred psychological profile.", "alternatives": "The sound of a name, deliberate memorisation, position, guessing the hypothesis, or preference formed before the task.", "realWorld": "Recommendation feeds, retargeting, repeated packaging, jingles, and recurring names in media.", "ethics": "Repetition can help learning or crowd out alternatives. The frequency, purpose, and ability to leave affect the person’s agency.", "everyday": "Outside this task, you may have no convenient record of how often a name or idea has appeared. Familiarity alone does not establish that something is accurate or a good fit. Ask what evidence supports your preference beyond recognising it.", "countermeasure": "Distinguish “I recognise it” from “it meets my criteria”. Seek alternatives outside the repeated stream.", "sources": [ "exposure", "exposureMeta" ], "evidence": [ { "key": "exposure", "text": "Zajonc’s foundational work examined whether repeated encounters could make a stimulus more positively evaluated. It motivates the repeated name in this task, including the possibility of liking without an elaborate reason. Exposure is one candidate influence; a single familiar name does not demonstrate a newly created preference." }, { "key": "exposureMeta", "text": "This review found that the relationship between repetition and liking was not simply an unlimited upward trend. Across the analysed studies it showed a curved pattern, with differences between conditions. That is why the reveal keeps familiarity and preference separate, and why “show it more” is not a universal rule." } ], "entryTitle": "A few quick comparisons", "entryDeck": "Compare some numbers, then make a small choice.", "entryContext": "Number comparisons and fictional names" }, { "id": "human-machine-authority", "number": "008", "title": "Human versus machine authority", "deck": "Same recommendation. A different name behind it.", "kind": "Demonstration", "status": "Ready to try", "lens": "adaptation", "tags": [ "AI", "Authority", "Trust" ], "minutes": "3 min", "engine": "authority", "observation": "The stated source of advice can change how the advice is received.", "hypothesis": "Identical advice may receive different trust ratings when attributed to a customer, expert, algorithm, AI assistant, or similar users.", "mechanism": "Perceived expertise, similarity, objectivity, or prior experience may influence reliance. The source label alone cannot establish actual competence.", "experiment": "Consider the same fictional recommendation in three presentations. One randomly selected human attribution and one randomly selected machine attribution appear first, in shuffled order. The unattributed recommendation appears last. Neither source category is assumed to be more persuasive. The notebooks have equivalent everyday-use titles with different phrasing and identical specifications. Their titles and the randomly assigned recommendation stay fixed across all rounds and replay. Choose an option and rate trust each time. The reveal follows all three rounds; replay removes all source labels.", "result": "Your choices and trust ratings are compared alongside the exact invented attributions. No real expert, AI service, or crowd generated the recommendation. Differences do not establish which label caused a response.", "alternatives": "Preference for one notebook title’s wording, prior attitudes toward AI, a preferred letter, suspicion, order, repetition, and remembering earlier responses may affect the comparison.", "realWorld": "AI assistants, expert badges, product recommendations, decision support, and “people like you” suggestions.", "ethics": "A useful source label clarifies provenance and limits. Unsupported authority can create confidence without evidence.", "everyday": "An everyday recommendation may affect a consequential choice without revealing how it was produced. A credible-sounding label does not replace the reasoning behind it. Check the source, supporting evidence, and limits before relying on the recommendation.", "countermeasure": "Read the reasoning without the label. Ask what evidence supports the recommendation and how accuracy was assessed.", "sources": [ "authority", "aversion" ], "evidence": [ { "key": "authority", "text": "Logg and colleagues found greater reliance on advice attributed to algorithms in several estimation tasks. The attribution comparison here asks a related question about the source label around a recommendation. It does not test the advice’s accuracy, and findings about numerical estimates do not automatically extend to every AI-mediated decision." }, { "key": "aversion", "text": "After seeing an algorithm make errors, participants could become less willing to use it even when it outperformed a human forecaster. This counterpoint matters: reliance depends on experience as well as attribution. Our short comparison holds the recommendation constant and cannot reproduce the history of trust or disappointment." } ], "entryTitle": "One recommendation", "entryDeck": "Consider the advice and make your own choice.", "entryContext": "Fictional notebook recommendations" }, { "id": "awareness-and-agency", "number": "009", "title": "Does the reveal give agency back?", "deck": "Knowing the mechanism is a hypothesis about protection.", "kind": "Experiment", "status": "Proposed · not recruiting", "lens": "adaptation", "tags": [ "Disclosure", "Transfer", "Agency" ], "minutes": "Research protocol", "engine": null, "observation": "A person can understand an intervention and still encounter it again in a different setting.", "hypothesis": "A mechanism-specific explanation may change responses to a new example compared with a general explanation; the direction and durability are open questions.", "mechanism": "Awareness, strategy use, demand effects, and learning may all contribute. A change in choice alone is not a complete measure of agency.", "experiment": "Pre-register parallel groups receiving a specific debrief or a matched general explanation. Use a new task for transfer and a later follow-up. Define outcomes, exclusions, sample size, consent, withdrawal, retention, and analysis before recruiting.", "result": "Proposed protocol. No participants are being recruited and no research results exist here.", "alternatives": "Task practice, memory for the prior answer, social desirability, fatigue, attrition, and recognising the researcher’s expectations.", "realWorld": "Education about defaults, platform literacy, consumer decisions, and awareness of adaptive systems.", "ethics": "Both groups should receive a complete debrief. Evaluate whether an intervention improves understood choice rather than merely reversing the preferred option.", "countermeasure": "Test transfer to unfamiliar material rather than treating an immediate replay as proof of resistance.", "sources": [ "transparency", "training" ], "evidence": [ { "key": "transparency", "text": "In a contribution experiment, explaining a default’s influence or purpose did not reliably eliminate its effect. Disclosure and influence can coexist. That makes this protocol’s central question testable rather than settled: recognising a cue need not remove it, although a null difference does not prove disclosure never matters." }, { "key": "training", "text": "Game and video interventions reduced several measured biases, with benefits still observed at a later follow-up. Practice and feedback mattered. This offers a constructive direction for the protocol: test transfer to new problems over time, rather than treating an improved immediate replay as evidence that awareness alone protects someone." } ] }, { "id": "useful-friction", "number": "010", "title": "A pause with a purpose", "deck": "Can a small interruption protect a deliberate choice?", "kind": "Intervention", "status": "Design protocol", "lens": "choice", "tags": [ "Friction", "Reversibility", "Prosocial design" ], "minutes": "Design protocol", "engine": null, "observation": "An irreversible action may be easy to trigger while its consequence is difficult to undo.", "hypothesis": "A clear confirmation or a reversible action may reduce unintended outcomes while preserving task completion.", "mechanism": "Friction can create a moment to inspect consequences. It can also cause fatigue, exclusion, and repeated automatic dismissal.", "experiment": "Compare immediate deletion, a contextual confirmation, and an undo window in a fictional task. Measure intended completion, unintended action, comprehension, recovery, and burden. Use no real destructive actions.", "result": "Design protocol only. No claim of benefit has been measured.", "alternatives": "Practice, salience of the prompt, differences in reading ability, motor access, and unfamiliarity with the task.", "realWorld": "Deleting work, sending messages, changing access, or confirming a consequential choice.", "ethics": "The person’s goal defines success. A protective pause should not become a trap or make the alternative unnecessarily difficult.", "countermeasure": "Keep consequences explicit, preserve an exit, and favour reversibility where it is available.", "sources": [ "warnings", "accuracy" ], "evidence": [ { "key": "warnings", "text": "Browser records covering millions of security warnings showed that responses varied with warning type and browser. Warnings were not uniformly ignored. For a protective pause, the lesson is to measure comprehension and behaviour in context; this observational study does not establish that our proposed confirmation or undo design will help." }, { "key": "accuracy", "text": "Experiments found that briefly directing attention toward accuracy could improve the quality of subsequently shared news. A useful interruption can focus attention on a relevant goal. This is an analogy for a consequential-action pause, not evidence that adding a delay to deletion is effective; that needs its own comparison." } ] }, { "id": "the-reason-after-the-choice", "number": "011", "title": "The reason after the choice", "deck": "Make a small choice. Then look closely at its explanation.", "kind": "Demonstration", "status": "Ready to try", "lens": "choice", "tags": [ "Choice blindness", "Introspection", "Agency" ], "minutes": "2 min", "engine": "blind", "journey": "reasons", "observation": "An account of a choice can sound coherent even when the outcome being explained has been changed.", "hypothesis": "A participant may offer a reason for the displayed notebook without reporting that it differs from the notebook they selected.", "mechanism": "Choice blindness describes failures to detect a mismatch between a choice and its presented outcome. A fluent explanation need not give direct access to how the original choice was made.", "experiment": "Choose between two notebooks. The next screen deliberately presents the other one as your choice. You can flag the mismatch immediately or continue with an optional explanation. The replay keeps the actual selection.", "result": "The record preserves your original selection, the displayed outcome, and whether you used the mismatch button. Not using that button does not establish that you failed to notice. Your explanation is not analysed.", "alternatives": "You may notice but play along, misclick, have little preference, recognise the premise, or read the title before starting. This single browser task is not a replication of the face-choice experiment.", "realWorld": "Use this question when reviewing a decision summary, a recommendation, or an account of why something happened. These are reflection contexts, not outcomes established by this task.", "ethics": "An interface must not rewrite a consequential decision. This disclosed, fictional demonstration changes no purchase or preference outside the task and shows the original selection in its reveal.", "everyday": "In an everyday transaction, a mismatch between your intention and the final summary may affect real money, access, or consent. You may receive no explanation or chance to replay. Check the final details against your original choice before confirming.", "countermeasure": "Keep an independent record of the decision and compare it with the outcome before explaining it. A plausible reason deserves the same scrutiny as an implausible one.", "sources": [ "blindness", "marketBlindness" ], "evidence": [ { "key": "blindness", "text": "Participants sometimes explained a face they had been handed after a concealed switch, without reporting that it differed from their choice. The result separates a fluent explanation from reliable access to the original decision. Our disclosed notebook adaptation changes the stakes, materials, and procedure and cannot diagnose anyone’s self-knowledge." }, { "key": "marketBlindness", "text": "In a supermarket task, people chose between jams or teas before the experimenters sometimes substituted the outcome. Some substitutions went unreported when people sampled and explained the supposed choice. This extends the question beyond portraits into everyday sensory preferences; it does not mean people never notice a mismatch." } ], "entryTitle": "A notebook for your thoughts", "entryDeck": "Choose a notebook and tell us a little more.", "entryContext": "A fictional notebook choice" }, { "id": "the-word-that-wasnt-there", "number": "012", "title": "The word that wasn’t there", "deck": "A short encounter with the difference between a feeling and a record.", "kind": "Demonstration", "status": "Ready to try", "lens": "memory", "tags": [ "Recognition", "Related words", "Confidence" ], "minutes": "2 min", "engine": "memory", "journey": "memory", "observation": "Related words can bring an associated word to mind even when that word was absent from the list.", "hypothesis": "The unpresented word “sleep” may be recognised as present after reading words associated with it.", "mechanism": "Related meanings can activate an absent word; remembering the theme may be confused with remembering a particular item. Several accounts of the DRM effect exist.", "experiment": "Read a related-word list at your own pace, hide it, then judge six words as present, absent, or uncertain. The recognition-question order varies between visits and is preserved in replay. Inspect the exact list afterward. The informed replay leaves the list available as an external record.", "result": "The reveal compares each answer with the actual list and preserves uncertainty. Recognising an absent word is an error on this task; neither an error nor a correct answer establishes your general memory reliability.", "alternatives": "Prior knowledge, reading time, guessing, language familiarity, distraction, and the study title can all affect the response. Self-paced text, one list, and immediate testing differ from the published procedure.", "realWorld": "An invitation to check notes and ask open questions when recollections differ. Word-list findings cannot determine whether any particular autobiographical memory or testimony is accurate.", "ethics": "This uses ordinary words and does not ask you to recover or reinterpret a personal event. A disagreement with the list is shown without shaming or a diagnostic label.", "everyday": "Here, there is an exact word list to inspect. Everyday disagreements may have no complete record and can affect trust between people. This task cannot establish whether a particular personal memory is accurate. Where possible, compare accounts with an independent record without assuming dishonesty.", "countermeasure": "Compare recollection with an independent record when one exists. In the replay, consult the list. That illustrates a practical support; it does not test whether awareness alone protects memory.", "sources": [ "drm", "memory" ], "evidence": [ { "key": "drm", "text": "In two experiments, studying related words led people to recall or recognise an associated word that had not appeared. Confidence could accompany the error. The significance here is that a feeling of remembering can be checked against a presentation record; this result does not adjudicate a particular autobiographical memory." }, { "key": "memory", "text": "Stadler and colleagues published norms for word lists used to elicit false recall and recognition. Their materials include the associates underlying this demonstration. The list makes the procedure inspectable, but our self-paced presentation and recognition questions differ: the paper’s group rates are not predictions or scores for you." } ], "entryTitle": "A short word list", "entryDeck": "Read a list, then answer a few questions.", "entryContext": "An ordinary word list" }, { "id": "the-scent-before-the-cup", "number": "013", "title": "The scent before the cup", "deck": "The shop can reach you before the product does.", "kind": "Field study", "status": "Protocol", "lens": "perception", "tags": [ "Physical environment", "Scent", "Expectation" ], "minutes": "Field protocol", "engine": null, "observation": "Coffee aroma can cross a shop boundary before someone decides to approach, enter, order, or taste anything.", "hypothesis": "A controlled coffee-like ambient scent at the entrance may change noticing, approach, entry, or expected quality even when the coffee, menu, service, and visible shop remain unchanged.", "mechanism": "Learned associations, expectation, appetite, and arousal are candidate explanations. The protocol does not assume the scent is unnoticed, that the effect is nonconscious, or that a stronger smell is better.", "experiment": "Counterbalance matched time blocks with no added scent, a low-intensity scent inside only, and the same scent at the threshold. Keep the menu, prices, signage, staffing, music, lighting, and coffee-production schedule fixed. Pretest perceived intensity; log weather, wind, doors, queues, footfall, and ventilation. Count passers-by, glances, approaches, and entries in aggregate. With consent, ask a separate intercept sample about expected quality and desire before ordering. In an unscented room, let a blinded panel evaluate the same standardised coffee so approach and expectation are not mistaken for product quality.", "result": "This is a proposed field protocol. No scent has been dispersed and no observations have been collected. The design must be reviewed with the venue, staff, safety specialists, and an appropriate ethics process before use.", "alternatives": "Natural brewing or bakery aromas, wind and ventilation, weather, time of day, caffeine demand, regular customers, visible queues, promotions, staff behaviour, and changing pedestrian traffic could explain a difference. Unstable scent plumes and unmatched blocks should be excluded under predeclared rules.", "realWorld": "Coffee shops, bakeries, hotel lobbies, food halls, transport hubs, retail entrances, and scent-enabled advertising.", "ethics": "Ambient scent can affect people who never chose to enter. Use ingredients and exposure levels reviewed for public settings; account for asthma, migraine, pregnancy, sensory sensitivity, staff exposure, and nearby businesses. Avoid individual tracking, prevent nuisance outside the agreed boundary, and disclose and debrief where practicable.", "everyday": "The smell near a shop may be produced, intensified, or placed independently of the item you eventually buy. It can shape approach and expectation without telling you anything about bean quality, freshness, preparation, labour, price, or whether you will enjoy the cup.", "countermeasure": "Step outside the scented zone before deciding. Separate “this smells inviting” from “this product meets my criteria,” and, where it matters, compare the product blind or against independent evidence.", "evidence": [ { "key": "storeScent", "text": "In simulated retail environments, participants evaluated scented and unscented settings differently. The study makes ambient odour a legitimate variable to isolate, but it did not test a real coffee-shop threshold or prove that scent improves purchases or product quality." }, { "key": "coffeeScent", "text": "A coffee-like scent changed expectations and analytical-task performance in the reported experiments. That coffee-specific association is relevant before a cup is tasted; it is not evidence that doorway scent causes entry, improves flavour, or operates outside awareness." }, { "key": "mallScent", "text": "Controlled field work in a shopping centre produced a more qualified picture, including limited or absent behavioural effects under stronger controls. This belongs beside the positive findings because scent is not a dependable lever in every environment and field confounds can create an appealing story." } ], "sources": [ "storeScent", "coffeeScent", "mallScent" ] }, { "id": "the-tempo-in-the-aisle", "number": "014", "title": "The tempo in the aisle", "deck": "A room can change pace without giving an instruction.", "kind": "Field study", "status": "Protocol", "lens": "attention", "tags": [ "Sensory influence", "Audio", "Temporal influence" ], "minutes": "Field protocol", "engine": null, "observation": "Background music changes over time while people move, wait, browse, and estimate how long an experience has taken.", "hypothesis": "Music tempo may change walking pace, dwell time, perceived duration, or abandonment while leaving the route and available products unchanged.", "mechanism": "Rhythmic entrainment, arousal, attention, and mood are candidate explanations. Tempo is not interchangeable with genre, familiarity, preference, volume, or musical fit.", "experiment": "Counterbalance no-music, slower-tempo, and faster-tempo blocks across comparable periods. Match loudness, genre, familiarity, valence, track duration, and playback location as closely as possible; where feasible use tempo-adjusted versions of the same material. Hold staffing, queue system, layout, pricing, and promotions fixed. Record aggregate route time and stopping, plus consenting participants’ perceived duration and comfort. Log footfall, time, events, and operational changes before analysing.", "result": "This is a proposed field protocol with no observations. It does not predict that fast music always speeds people up or that longer dwell necessarily benefits the visitor or venue.", "alternatives": "Crowding, queues, staff pace, shopping purpose, age, mobility, hearing, music preference, promotions, weather, and time of day may change movement or duration independently of tempo.", "realWorld": "Supermarkets, restaurants, gyms, casinos, waiting rooms, exhibitions, public transport, and hospitality spaces.", "ethics": "Keep volume and repetition within comfortable limits, provide quieter routes or times, include staff consent and exposure, and treat accessibility and sensory needs as design requirements rather than noise in the data.", "everyday": "The pace that feels natural in a room may partly belong to the room. Faster movement, longer browsing, or a distorted sense of waiting does not necessarily reveal a stable preference or deliberate plan.", "countermeasure": "Notice the soundtrack and check elapsed time against a clock. Set a route, budget, or stopping rule before entering when pace and duration matter.", "evidence": [ { "key": "musicTempo", "text": "In a supermarket field experiment, background-music tempo was associated with changes in traffic pace and sales volume. It motivates measuring movement in place, but one store study does not establish a universal tempo effect or separate every feature of the music and shopping day." }, { "key": "musicTaste", "text": "Music playing during wine tasting changed how separate groups described the wine in this experiment. That result broadens the protocol beyond walking speed: sound may alter the experience people attribute to an object. It does not show that tempo alone caused those ratings." } ], "sources": [ "musicTempo", "musicTaste" ] }, { "id": "the-crowd-that-makes-a-crowd", "number": "015", "title": "The crowd that makes a crowd", "deck": "Attention can become visible—and contagious.", "kind": "Field study", "status": "Protocol", "lens": "desire", "tags": [ "Social influence", "Crowds", "Attention" ], "minutes": "Field protocol", "engine": null, "observation": "A few people who stop and look can become information for everyone approaching after them.", "hypothesis": "A larger visible group attending to the same neutral display may increase the proportion of passers-by who glance, slow, stop, or approach.", "mechanism": "Social proof, informational uncertainty, imitation, curiosity, and simple visual obstruction are competing explanations. Following a gaze does not establish agreement, desire, or unawareness.", "experiment": "With venue approval, counterbalance blocks using zero, one, and five trained adults who stop at fixed positions and attend to the same neutral display for the same duration. Record aggregate counts of eligible passers-by, glances, slowing, stops, and approaches without faces or identifiers. Keep the display, sightline, actor behaviour, and block length fixed; log weather, local events, pedestrian density, obstructions, and existing queues.", "result": "This is a proposed field protocol. No actors have been placed and no public behaviour has been observed. The protocol should be preregistered with safety, stopping, and exclusion rules.", "alternatives": "A larger group is more visually salient, blocks the path, creates sound, signals a queue, or makes the display harder to see. Time, events, weather, and pedestrian composition can also change stopping.", "realWorld": "Shop windows, street performances, restaurant queues, trade-show booths, viewing counters, livestream audiences, trending lists, and visible engagement metrics.", "ethics": "Do not obstruct paths, manufacture demand for a real transaction, create alarm, record identifiable people, or include minors in analysis. Use a neutral setting, short blocks, venue approval, public notice or debrief where feasible, and an immediate stop rule for congestion or distress.", "everyday": "Other people’s attention can be genuine while still being a weak guide to quality or relevance. A visible crowd may describe what earlier people noticed—or may be the intervention that caused later people to notice it.", "countermeasure": "Ask what drew the first people and whether the crowd supplies evidence about the thing itself. Inspect the object or claim separately from the number of people attending to it.", "evidence": [ { "key": "crowdDrawing", "text": "On a busy street, Milgram and colleagues varied the size of a group looking upward and recorded whether passers-by adopted the looking behaviour. Larger groups drew more looking in that setting. The procedure supports a crowd-size comparison, not a claim that every crowd signals value or forces conformity." }, { "key": "culturalMarket", "text": "In an online music market, showing other people’s downloads made success more unequal and less predictable. It provides a digital counterpart to a street crowd: visible attention can help produce the popularity it appears to report, while the underlying songs still mattered." } ], "sources": [ "crowdDrawing", "culturalMarket" ] } ], "sources": { "blindness": { "title": "Johansson et al. (2005) — Failure to detect mismatches between intention and outcome", "url": "https://pubmed.ncbi.nlm.nih.gov/16210542/", "note": "Participants in a face-choice task sometimes explained a substituted outcome without reporting the mismatch. Our notebook task changes the materials, timing, and disclosure; it is an educational adaptation." }, "memory": { "title": "Stadler, Roediger & McDermott (1999) — Norms for word lists that create false memories", "url": "https://link.springer.com/article/10.3758/BF03211543", "note": "Following Roediger and McDermott’s 1995 work, this paper reports related-word lists and false recall and recognition. This self-paced, single-list demonstration has a different procedure." }, "conformity": { "title": "Asch (1956) — Studies of independence and conformity", "url": "https://doi.org/10.1037/h0093718", "note": "A historical line-judgement study of a person facing a unanimous majority. Its setting and sample limit generalisation; both conformity and independence were observed." }, "agency": { "title": "Maoz et al. (2019) — Neural precursors of decisions that matter", "url": "https://pubmed.ncbi.nlm.nih.gov/31642807/", "note": "This study compared arbitrary actions with deliberate donation choices. It raises a limit on generalising neural findings from one kind of decision to another; it does not settle free will." }, "defaultsMeta": { "title": "Jachimowicz et al. (2019) — When and why defaults influence decisions", "url": "https://doi.org/10.1017/bpp.2018.43", "type": "Meta-analysis" }, "wine": { "title": "Plassmann et al. (2008) — Marketing actions can modulate neural representations of experienced pleasantness", "url": "https://pubmed.ncbi.nlm.nih.gov/18195362/", "type": "Behavioural and fMRI experiment" }, "coherentAnchors": { "title": "Ariely, Loewenstein & Prelec (2003) — Coherent Arbitrariness", "url": "https://academic.oup.com/qje/article-abstract/118/1/73/1917051", "type": "Valuation experiments" }, "selectivity": { "title": "Theeuwes (1992) — Perceptual selectivity for color and form", "url": "https://link.springer.com/article/10.3758/BF03211656", "type": "Visual-search experiments" }, "assortment": { "title": "Iyengar & Lepper (2000) — When choice is demotivating", "url": "https://pubmed.ncbi.nlm.nih.gov/11138768/", "type": "Field and laboratory experiments" }, "culturalMarket": { "title": "Salganik, Dodds & Watts (2006) — Experimental study of inequality and unpredictability in an artificial cultural market", "url": "https://pubmed.ncbi.nlm.nih.gov/16469928/", "type": "Online market experiment" }, "exposureMeta": { "title": "Montoya et al. (2017) — A re-examination of the mere exposure effect", "url": "https://pubmed.ncbi.nlm.nih.gov/28263645/", "type": "Meta-analysis" }, "aversion": { "title": "Dietvorst, Simmons & Massey (2015) — Algorithm aversion", "url": "https://pubmed.ncbi.nlm.nih.gov/25401381/", "type": "Forecasting experiments" }, "musicTaste": { "title": "North (2012) — The effect of background music on the taste of wine", "url": "https://bpspsychub.onlinelibrary.wiley.com/doi/10.1111/j.2044-8295.2011.02072.x", "type": "Sensory-context experiment" }, "training": { "title": "Morewedge et al. (2015) — Debiasing Decisions", "url": "https://journals.sagepub.com/doi/abs/10.1177/2372732215600886", "type": "Longitudinal training experiments" }, "transparency": { "title": "Bruns et al. (2018) — Can nudges be transparent and yet effective?", "url": "https://www.sciencedirect.com/science/article/abs/pii/S0167487017307845", "type": "Default-disclosure experiment" }, "warnings": { "title": "Akhawe & Felt (2013) — Alice in Warningland", "url": "https://www.usenix.org/conference/usenixsecurity13/technical-sessions/presentation/akhawe", "type": "Large-scale observational field study" }, "accuracy": { "title": "Pennycook et al. (2021) — Shifting attention to accuracy can reduce misinformation online", "url": "https://pubmed.ncbi.nlm.nih.gov/33731933/", "type": "Survey and field experiments" }, "marketBlindness": { "title": "Hall et al. (2010) — Magic at the marketplace", "url": "https://portal.research.lu.se/en/publications/magic-at-the-marketplace-choice-blindness-for-the-taste-of-jam-an/", "type": "Supermarket choice experiment" }, "drm": { "title": "Roediger & McDermott (1995) — Creating false memories", "url": "https://doi.org/10.1037/0278-7393.21.4.803", "type": "Word-list memory experiments" }, "storeScent": { "title": "Spangenberg, Crowley & Henderson (1996) — Improving the Store Environment: Do Olfactory Cues Affect Evaluations and Behaviors?", "url": "https://journals.sagepub.com/doi/10.1177/002224299606000205", "type": "Simulated retail experiments" }, "coffeeScent": { "title": "Madzharov et al. (2018) — The impact of coffee-like scent on expectations and performance", "url": "https://www.sciencedirect.com/science/article/abs/pii/S0272494418302615", "type": "Coffee-scent experiment" }, "mallScent": { "title": "Teller & Dennis (2012) — The effect of ambient scent on consumers’ perception, emotions and behaviour", "url": "https://www.tandfonline.com/doi/abs/10.1080/0267257X.2011.560719", "type": "Shopping-centre field experiments" }, "musicTempo": { "title": "Milliman (1982) — Using Background Music to Affect the Behavior of Supermarket Shoppers", "url": "https://journals.sagepub.com/doi/10.1177/002224298204600313", "type": "Supermarket field experiment" }, "crowdDrawing": { "title": "Milgram, Bickman & Berkowitz (1969) — Note on the Drawing Power of Crowds of Different Size", "url": "https://psycnet.apa.org/doi/10.1037/h0028070", "type": "Street field experiment" }, "defaults": { "title": "Johnson & Goldstein (2003) — Do Defaults Save Lives?", "url": "https://business.columbia.edu/faculty/research/do-defaults-save-lives", "note": "The authors investigated default conditions and donation agreement. This demonstration concerns a fictional interface preference; it is not a replication or a claim about actual donation outcomes." }, "anchors": { "title": "Tversky & Kahneman (1974) — Judgment under Uncertainty", "url": "https://www.science.org/doi/10.1126/science.185.4157.1124", "note": "The paper describes heuristics including anchoring. It does not establish the strength of the price, stock, or popularity cues in this particular demonstration." }, "salience": { "title": "Itti & Koch (2000) — A saliency-based search mechanism", "url": "https://www.sciencedirect.com/science/article/pii/S0042698999001637", "note": "A computational account of visual salience provides background. Button selection and free recall in this demonstration are not eye tracking and do not validate that model." }, "choice": { "title": "Scheibehenne, Greifeneder & Todd (2010) — A meta-analytic review of choice overload", "url": "https://archive-ouverte.unige.ch/unige:76440", "note": "Across the reviewed studies, the average effect was close to zero with substantial variation. More options do not reliably produce overload in every setting." }, "social": { "title": "Muchnik, Aral & Taylor (2013) — Social Influence Bias", "url": "https://www.science.org/doi/10.1126/science.1240466", "note": "A randomised field experiment found that prior ratings affected later rating behaviour. Effects in that setting do not provide an expected effect size for this local example." }, "exposure": { "title": "Zajonc (1968) — Attitudinal Effects of Mere Exposure", "url": "https://web.mit.edu/curhan/www/docs/Articles/biases/9_J_Personality_Social_Psychology_1_%28Zajonc%29.pdf", "note": "This foundational paper examines exposure and attitudes. A short exposure task here illustrates the question; familiarity does not guarantee preference." }, "authority": { "title": "Logg, Minson & Moore (2019) — Algorithm Appreciation", "url": "https://www.sciencedirect.com/science/article/pii/S0749597818303388", "note": "The authors report experiments involving reliance on algorithmic advice. Task, expertise, experience, and wording can change reliance; an AI label is not a measure of actual accuracy." } } }