Self-Perception Theory: An S-Tier Behavioral Designer’s Guide

In 1967, Daryl Bem walked into the cognitive dissonance literature and proposed that the entire towering edifice that Festinger had built — the drive states, the discomfort, the motivation to reduce inconsistency — might not actually exist. People were not, Bem argued, contorting their attitudes to relieve internal tension. They were doing something far more mundane and far more disturbing: they were watching themselves act, and inferring what they must believe from what they had just done.

If Bem was right, the implications for behavioral design are not subtle. Every onboarding flow that asks a user to take one small action, every commitment device that sets a public stake in the ground, every habit-tracker that nudges a single first rep — none of these are working through the mechanism most product teams assume. They are not motivating an existing internal preference into expression. They are manufacturing the preference itself, retroactively, by giving the user a behavior to point at and reverse-engineer their own values from.

This post is a working designer’s guide to that mechanism: what Self-Perception Theory actually claims, what the experimental record actually shows, where the theory falls apart under scrutiny, and how the Octalysis Framework treats Bem’s insight as the load-bearing CD4 (Ownership & Possession) substrate underneath foot-in-the-door progressions, identity-led onboarding, and the entire family of behavior-first design moves that turn one click into a self.

Speed Run Notes

• The thesis. When internal cues are weak, ambiguous, or uninterpretable, people infer their own attitudes the same way an outside observer would — by watching their behavior and the situation it occurred in (Bem 1967, 1972).
• The original threat. Bem reanalyzed Festinger & Carlsmith’s $1/$20 forced-compliance study without invoking dissonance and produced the same predictions, suggesting the entire dissonance literature might be reframed as observer-style self-attribution.
• The 1977 resolution. Fazio, Zanna & Cooper showed both theories are correct in their own domain: dissonance fires when behavior is attitude-discrepant beyond a latitude of acceptance; self-perception fires when behavior is attitude-congruent or within that latitude.
• The flagship application. The foot-in-the-door technique (Freedman & Fraser 1966) — agree to a small request, then a large one — is the cleanest production-grade demonstration that Bem’s mechanism scales: people who comply with a small ask infer they are the kind of person who complies, and the inferred self-concept does the work.
• The empirical drag. The 1988 facial-feedback study (Strack, Martin & Stepper pen-in-mouth) was the textbook self-perception demonstration; the 2016 Registered Replication Report across 17 labs (n ≈ 1,894) failed to find the effect, and the broader replication record is mixed.
• The Octalysis design rule. Self-perception is the trait-level shadow of CD4 (Ownership & Possession): every action a user takes that they cannot easily attribute to coercion, payment, or social pressure becomes raw material the user uses to build the self-concept your product depends on. Design the visible behavior history; don’t only design the next click.

Table of Contents

Author Credibility: Yu-kai Chou

Yu-kai Chou is an S-Tier Behavioral Designer and the creator of the Octalysis Framework, the gamification design system now applied to products and experiences reaching over 1.5 billion users. His book Actionable Gamification is one of the most-cited works in the field, and he has been ranked the #1 Gamification Guru in the World.

He has advised MrBeast, LEGO, Microsoft, Porsche, Tesla, Stanford, Harvard, and governments including Ukraine on turning behavioral psychology into product mechanics that actually change user behavior.

Verify: Wikipedia · Google Scholar · Wikidata · LinkedIn

Self-Perception Theory sits in a particular load-bearing place in my own thinking about behavioral design. The argument I have been making for two decades — that the right mechanic is the one that lets the user do the identity rather than receive a label for it — runs straight through Bem’s 1967 paper. Every CD4 (Ownership & Possession) Game Technique I teach (Avatar #4, Collection Set #37, Endowed Effect #23, Alfred Effect #27) is a structured way to give the user a behavior history that supports the self-inference your product needs the user to make. The foot-in-the-door progressions inside Octalysis Level 2 onboarding flows, the visible commitment ladders inside the 4 Experience Phases, and the Black-Hat overshoot guardrails I keep insisting on for points-and-badges economies are all, structurally, applied Self-Perception Theory. This pillar is the academic floor under that practice.

What Is Self-Perception Theory

Self-Perception Theory is a 1967 social-psychological proposition by Daryl Bem that people determine their own attitudes, emotions, and internal states by observing their own behavior and the conditions under which it occurred — the same way an outside observer would. The theory’s two core claims, paraphrased almost verbatim from Bem (1972), are:

1. Individuals come to know their own attitudes, emotions, and other internal states partially by inferring them from observations of their own overt behavior and/or the circumstances in which this behavior occurs.
2. To the extent that internal cues are weak, ambiguous, or uninterpretable, the individual is functionally in the same position as an outside observer — an observer who must necessarily rely on those same external cues to infer the individual’s inner states.

The theory’s range of application is therefore not the whole of self-knowledge — it is the slice of self-knowledge where introspective cues are unreliable. People know directly that they are in pain, that they are hungry, that they prefer chocolate to anchovies. But for the long tail of attitudes, mild preferences, attributions, and self-concept beliefs that make up most of social cognition, Bem’s claim is that we are not reading off an internal record. We are observing ourselves the way a stranger would.

This is a deeply counterintuitive claim. Folk psychology insists on the priority of the inner state: we believe X, therefore we say X. Bem flips the causal arrow: we say X, therefore (under the right conditions) we infer we believe X. The popular shorthand — “we don’t sing because we’re happy; we’re happy because we sing” — captures the structural inversion but understates how unsettling the implication is for any model of agency that takes the self as a fixed object the conscious mind has direct access to.

Three boundary conditions sharpen the claim. First, the inference applies primarily when behavior is freely emitted — when no one is paying for it, no one is forcing it, no one is watching with an obvious agenda. Second, the inference applies primarily when introspective cues are absent or uninformative. Third, the inference is not infallible: people make systematic mistakes about their own attitudes the same way they make systematic mistakes about other people’s attitudes, because the data they have to work with is the same kind of data.

What Bem’s theory accomplishes, when those boundary conditions hold, is a clean re-derivation of a large fraction of attitude-change findings without invoking any motivational drive state at all. No dissonance, no discomfort, no tension to reduce — just a person walking back through the events of the last hour and updating their beliefs about themselves to match what they apparently did.

The Core Findings

The empirical record on Self-Perception Theory falls into three buckets: the original re-derivation of dissonance findings, the foot-in-the-door compliance literature, and the facial-feedback / posture-feedback literature on emotional self-attribution. Each tells a different chapter of the story.

Bem 1965: The Original Self-Persuasion Demonstrations

Before the 1967 theoretical paper, Bem (1965) ran a series of “interpersonal simulations” in the Journal of Experimental Social Psychology. Participants listened to recordings of people making attitudinal statements under conditions of either high incentive ($1) or low incentive ($20) — and were asked to estimate the speakers’ true attitudes. The result: observers used the size of the incentive as a discounting cue. They inferred that the speakers under high incentive were probably saying things they did not really believe, while the speakers under low incentive were probably saying things they did believe.

Bem’s argument was that participants in the original Festinger & Carlsmith (1959) forced-compliance study were doing the same thing — only the speaker and the observer were now the same person. A participant who had told a confederate that a boring task was interesting, in exchange for $1, looked back at their behavior, applied the discounting rule (no observer would believe a $1 incentive caused a sincere statement), and concluded the only remaining explanation was that they must have actually found the task somewhat interesting. The attitude shift that Festinger had attributed to dissonance reduction was, on Bem’s reading, a straightforward observer-style attribution from behavior to attitude.

Freedman & Fraser 1966: The Foot-in-the-Door

The flagship production-grade demonstration of Self-Perception Theory was published a year before the theoretical paper itself. Freedman & Fraser (1966) sent experimenters to suburban Palo Alto homes asking residents to allow a “huge, ugly” 36-square-foot Drive Carefully sign to be installed in their front yard. The naïve compliance rate was 17%.

The experimental condition first asked residents to do a tiny related favor — sign a petition supporting safe driving, or place a small 3-inch Drive Safely sticker in a window. Two weeks later, the same experimenters returned with the large-sign request. Compliance rose to 76% — more than four times the baseline rate.

Bem’s self-perception interpretation became the canonical explanation: residents who agreed to the small request inferred from their own behavior that they were the kind of person who supported civic safety causes. When the large request arrived, the inferred self-concept was already in place, and refusing the large request would have been inconsistent with the now-active identity. The residents had not been talked into anything. They had been given a behavior to point at, and their attitude had followed.

Burger (1999) reviewed roughly 60 follow-up foot-in-the-door studies in Personality and Social Psychology Review and found a small but reliable mean effect (Cohen’s d ≈ 0.20–0.30) with substantial heterogeneity. The technique works most reliably when the small request is large enough to require a real decision (not trivial), when it is congruent with the eventual large request (same domain), when the actor cannot easily attribute their compliance to external pressure, and when the labeling of the actor as a “helpful person” is made salient between requests. Each of those moderators is consistent with self-perception’s mechanism: the bigger and freer and more identity-loaded the small behavior, the richer the data for self-inference.

Laird 1974 / Strack et al. 1988: The Facial-Feedback Hypothesis

The most-cited extension of Bem’s framework into emotion research came from James Laird (1974) and later Fritz Strack, Leonard Martin & Sabine Stepper (1988). The argument: if people infer attitudes from behavior, they should also infer emotions from facial expressions, since the face is one of the most legible behavioral surfaces a person has to read. Laird had participants contract specific facial muscles under cover stories about electrode placement; participants who unknowingly produced smile-like configurations reported feeling happier than those who produced frown-like configurations.

Strack’s 1988 pen-in-mouth design became the textbook demonstration. Participants held a pen between their teeth (forcing a smile-like contraction) or between their lips (preventing a smile) while rating the funniness of Far Side cartoons. The teeth-condition participants rated the cartoons as funnier — a reported effect size that became one of the most-taught social-psychology demonstrations in undergraduate textbooks for nearly three decades.

The 2016 Registered Replication Report (Wagenmakers et al., Perspectives on Psychological Science) coordinated 17 labs running the original Strack protocol with n ≈ 1,894 participants. The pooled effect estimate was indistinguishable from zero. Subsequent meta-analyses (Coles, Larsen & Lench 2019, Psychological Bulletin, k = 138) recovered a small mean effect (g ≈ 0.13) but found that the effect was substantially smaller than the original studies suggested and that publication bias accounts for a meaningful portion of the historical record. The facial-feedback hypothesis is not zero — but it is not the bedrock self-perception demonstration the textbooks treated it as.

Lepper, Greene & Nisbett 1973: Overjustification as Self-Perception

The fourth canonical body of evidence is the Overjustification Effect — and it is structurally a Self-Perception Theory experiment. Lepper, Greene & Nisbett (1973) had nursery-school children draw with markers under three conditions: an Expected Reward condition (told in advance they would receive a Good Player Award), an Unexpected Reward condition (received the same award without being told beforehand), and a No Reward condition. Two weeks later, free-play observation showed the Expected-Reward children had cut their drawing time roughly in half compared to the other two conditions.

The self-perception reading: children in the Expected Reward condition watched themselves drawing for the award, applied the same discounting rule Bem had identified in 1965 (people who do things for prizes probably do not love the activity for its own sake), and reattributed their pre-experiment intrinsic interest to the extrinsic incentive. When the incentive disappeared, the inferred reason for drawing disappeared with it, and the free-choice behavior collapsed. Bem’s mechanism is the load-bearing causal account underneath the entire Cognitive Evaluation Theory branch of Self-Determination Theory and the empirical floor under the Black-Hat / White-Hat distinction in the Octalysis Framework. (For the full pillar on this finding and its 128-experiment Deci, Koestner & Ryan 1999 meta-analysis, see the Overjustification Effect guide.)

What Daryl Bem Got Right

Strip away the Festinger-versus-Bem palace politics of late-1960s social psychology and the substantive durable contributions of Self-Perception Theory come into focus. Bem’s paper survives because it solved real problems with the dissonance literature, and because the parts of it that survived survived for principled reasons.

The Parsimony Argument

Festinger’s dissonance theory required a motivational drive state — an internal psychological discomfort triggered by inconsistency that people were motivated to reduce. Bem showed that a wide swath of dissonance findings could be re-derived without that drive state, using only the more parsimonious assumption that people make attribution-style inferences about themselves. In the philosophy-of-science tradition, the cheaper theory wins when both make the same predictions. Bem’s theory was cheaper, and for a slice of the experimental record, it made identical predictions.

This was not idle theorizing. The dissonance literature had accumulated a serious problem: every time a dissonance prediction failed, the standard move was to invoke an additional dissonance-reduction route — selective exposure, attitude bolstering, denial of responsibility, trivialization. Each rescue made the theory unfalsifiable in its strongest form. Bem’s challenge forced the dissonance literature to specify exactly when the drive state actually fires, which led directly to the productive Fazio-Zanna-Cooper resolution that we discuss below.

The Discounting Principle Has Held Up

The single most replicable piece of Self-Perception Theory is the discounting principle: when behavior can be attributed to a salient external cause, observers (including the self-as-observer) reduce the weight they give to internal-attitude explanations. Kelley’s (1973) attribution-theoretic formalization of this principle treats it as a general computational property of social cognition, and four decades of attribution research has not unsettled it.

The practical implication for behavioral design is direct. Any time the user’s behavior can be plausibly explained by an external cue — a payment, a coercive nudge, a social-pressure surface, a public commitment they could not gracefully exit — the self-perception inference fails. The user discounts the behavior, attributes it to the external pressure, and does not update the self-concept. This is why coercive gamification rarely produces durable behavior change: the inference machinery refuses to assign the behavior to the self when an external cause is sitting in plain view.

The Asymmetry of Strong vs Weak Attitudes

Bem’s third durable contribution was the explicit acknowledgment that the theory’s domain is bounded by attitude strength. For attitudes that are strongly held, well-rehearsed, and supported by a rich internal record, behavior does not need to do the inferential work — the introspective cues are loud enough to read directly. Self-perception fires hardest in the long tail of weak, inchoate, ambiguous attitudes, which happens to be the empirical region most product designers actually operate in. Users do not arrive at most onboarding flows with a settled, articulate, deeply-held attitude toward your product category. They arrive curious, ambivalent, or neutral. That is precisely the attitude-strength region where Bem’s mechanism does its most consequential work.

This insight, alone, ought to be a load-bearing piece of every product team’s mental model. The user’s attitude toward your product is forming, not pre-existing. Whatever behavior you elicit in the first five minutes is going to be the data they reverse-engineer their attitude from. There is no neutral first interaction. The behavior is the attitude in formation.

Where Self-Perception Theory Falls Apart

The strongest parts of Self-Perception Theory survived nearly six decades of scrutiny. The weaker parts — and there are several — explain why the theory is best treated today as one component of a portfolio of attitude-change models, not as a unifying account that displaces dissonance theory the way Bem originally argued.

The Fazio-Zanna-Cooper Boundary Showed the Theory’s Domain Is Smaller Than Bem Claimed

The most consequential challenge to the unifying ambition of Self-Perception Theory came from Russell Fazio, Mark Zanna & Joel Cooper (1977) in the Journal of Experimental Social Psychology. Their experiment placed participants in either an attitude-discrepant condition (counter-attitudinal essay-writing on a topic they cared about, with low choice and a small payment) or an attitude-congruent condition (essay-writing on a topic toward which they were initially neutral). They predicted, and found, that arousal-based dissonance reduction would dominate in the discrepant condition, while observer-style self-perception inference would dominate in the congruent condition.

The result, replicated and extended in the dissonance and attribution literatures, established that both theories are correct in their proper domain. Dissonance fires when the behavior falls outside the latitude of acceptance — when the actor said or did something that genuinely contradicts a held attitude. Self-perception fires when the behavior is congruent with held attitudes or when no held attitude exists. Bem’s original ambition — the elimination of the dissonance drive state altogether — did not survive this work. The two theories now coexist as complementary mechanisms with separable empirical territories.

This is a real concession. Self-Perception Theory is not the unifying theory of attitude change Bem originally proposed; it is a precise local theory of how attitudes form when no attitude has yet formed. That is still consequential — most of the attitudes a designer cares about are exactly the ones that have not yet formed — but the historical revisionism that treats Bem’s paper as having “won” the debate is wrong.

The Facial-Feedback Replication Crisis Cut a Major Branch of the Theory’s Empirical Tree

The 2016 Registered Replication Report on Strack, Martin & Stepper (1988) was one of the most consequential events in the recent history of social psychology. Seventeen labs, running the original protocol with substantially more statistical power, recovered an effect indistinguishable from zero. The pen-in-mouth study had been a textbook fixture for nearly three decades; its failure to replicate at scale was a body blow to the broader self-perception-of-emotion literature that had grown up around it.

The follow-up evidence is more nuanced than “the effect is fake” but less reassuring than the original literature suggested. Coles, Larsen & Lench (2019) published a multi-decade meta-analysis in Psychological Bulletin covering 138 studies and 11,000 participants, recovering a mean effect of g ≈ 0.13 — present, reliable in the aggregate, but substantially smaller than the original studies suggested and substantially heterogeneous across paradigms. Coles et al. (2022) followed with a 19-laboratory many-labs project (n ≈ 3,900) using a directed facial action task and recovered a small but reliable effect (g ≈ 0.17), which the authors interpreted as suggestive of a real but context-dependent phenomenon.

The honest synthesis: facial feedback exists, but the original textbook narrative — that smiling reliably makes people happier and frowning reliably makes people sadder, in the way Strack 1988 implied — overstated the size and reliability of the effect. The self-perception interpretation of emotion remains a live hypothesis but a modest one, and the broader claim that emotional self-knowledge runs through a self-observer attribution process is far less empirically settled than the textbooks of the 2000s implied.

The Introspection Objection: Self-Perception Cannot Account for Direct Phenomenal Knowledge

The third durable critique of Self-Perception Theory is philosophical rather than empirical. Even granting that observers and selves use similar attributional logic for inferring attitudes, the theory has no account for the phenomenal character of direct introspective knowledge — the experience of knowing one is in pain, one is hungry, one is afraid, without needing to watch oneself behave to figure it out. Bem’s caveat (the theory applies “to the extent that internal cues are weak, ambiguous, or uninterpretable”) concedes the point but does not solve it. Where exactly is the boundary between the introspective channel and the self-observer channel? How does a person know which channel to consult? And does the self-observer channel produce attitudes that feel different, phenomenally, from attitudes formed through introspection?

Andersen & Ross (1984) and later work in the introspection-on-attitudes literature suggested that the self-observer channel produces attitudes that are weaker, less stable, and less predictive of subsequent behavior than the introspective channel. This implies that the attitudes designers manufacture through self-perception mechanisms — onboarding behaviors, foot-in-the-door progressions, micro-commitments — are real but fragile. They can be unmade as easily as they were made by exposing the user to a different behavior history. The implication for product design is that self-perception-built attitudes need to be reinforced over time through a continuous behavior record, not assumed to lock in after a single onboarding success.

The Brain on Self-Perception

The neural evidence sits inside the broader literature on self-referential processing, which has been mapped extensively using fMRI since the early 2000s. Northoff et al. (2006) meta-analyzed 27 imaging studies of self-referential judgment and isolated a “cortical midline structures” network anchored by the medial prefrontal cortex (mPFC), the posterior cingulate cortex (PCC), and the temporoparietal junction (TPJ). The same network shows up when subjects make trait-attribution judgments about close others (Mitchell, Macrae & Banaji 2006), with the dissociation between “judging self” and “judging other” tracking perceived similarity rather than category membership. The neural machinery does not draw a privileged boundary around the self the way Cartesian intuition assumes; it treats inferring one’s own attitudes as one instance of the more general inference problem of inferring any agent’s attitudes from observed behavior.

That structural overlap is itself evidence for the Bem proposal. If self-knowledge ran through a privileged introspective channel, the imaging signature should look qualitatively different from third-person attribution; it does not. What changes between self and other is the input data — autobiographical memory for the self case, observed behavior for the other case — not the inferential machinery downstream of the input.

Important caveat: the neuroscience here is correlational, not mechanistic. The cortical-midline activations do not by themselves adjudicate between self-perception and dissonance, because both theories predict similar self-referential network engagement whenever behavior-attitude inferences are being computed. The strongest neural evidence for Bem-style attribution is dissociation evidence. Lieberman, Jarcho & Satpute (2004) showed that deliberate “reflective” attribution engages mPFC while practiced “reflexive” attribution drops out of mPFC and into basal ganglia and amygdala circuits. That pattern is what Bem’s mechanism predicts: novel behavior produces conscious self-inference (lights up mPFC), repeated behavior produces a settled attitude that the user no longer needs to reconstruct (drops out of mPFC).

The design implication is direct. The user is doing self-inference most heavily early in the onboarding flow, when the behavior history is new and the self-concept is being built; this is the design window in which the visible behavior history surface matters most. By the time a user has six months of behavior history, the attitude has migrated out of the reflective inference circuit and into the practiced one — and the design has either built the self-concept it wanted (CD4 destination achieved) or built one the user is now stuck with.

Self-Perception Theory vs Other Theories

Self-Perception Theory occupies a specific place in a crowded ecosystem of attitude-change and self-knowledge theories. The best way to read it is in relation to the four nearest neighbors, each of which it overlaps with along a different dimension.

vs Cognitive Dissonance Theory (Festinger 1957)

The largest single contrast is the one Bem originally drew. Dissonance theory posits a motivated drive state triggered by inconsistency between cognitions; Self-Perception Theory posits a non-motivated inferential process from behavior to attitude. As discussed above, the Fazio-Zanna-Cooper (1977) resolution established that both theories are correct in their proper domain: dissonance for attitude-discrepant behavior outside the latitude of acceptance, self-perception for attitude-congruent or initially-neutral behavior. The two are now standard companions in the social-psychology curriculum, not rival accounts. (For the dissonance pillar, see Cognitive Dissonance: An S-Tier Behavioral Designer’s Guide.)

vs Self-Determination Theory (Deci & Ryan 1985)

Self-Determination Theory’s Cognitive Evaluation Theory branch is, in its causal mechanism, an applied Self-Perception Theory. Deci’s original 1971 puzzle-solving experiment and Lepper, Greene & Nisbett’s 1973 Overjustification Effect paper both relied on Bem’s discounting principle to explain why extrinsic rewards undermine intrinsic motivation: the user watches themselves doing the activity for the reward, applies the discounting rule, and reattributes their pre-reward engagement from “I do this because I want to” to “I do this because someone is paying me.” Where Self-Perception Theory provides the mechanism, Self-Determination Theory provides the developmental architecture (autonomy, competence, relatedness as basic psychological needs) and the regulation continuum (external, introjected, identified, integrated, intrinsic) that lets designers reason about which kinds of self-attributions support sustainable motivation. (For the SDT pillar, see Self-Determination Theory: An S-Tier Behavioral Designer’s Guide.)

vs MASK00022 (Heider 1958; Kelley 1973; Weiner 1985)

Attribution Theory is the broader parent class. Where Self-Perception Theory is specifically about how people attribute their own attitudes from their own behavior, Attribution Theory is the general account of how people attribute causes to behavior — their own, others’, past events, future events. Kelley’s (1973) covariation principle (consensus, distinctiveness, consistency) is the formal extension of Bem’s discounting principle to a richer space of inferential operations. Weiner’s (1985) attributional theory of motivation extends the attribution machinery to outcomes (success, failure) and emotional consequences (pride, shame, guilt). Self-Perception Theory is best understood as the attitude-domain instantiation of the attribution-theoretic toolkit. (For the broader attribution treatment, see Attribution Theory: An S-Tier Behavioral Designer’s Guide.)

vs Classical Conditioning & Operant Conditioning

The behaviorist tradition explains attitude formation through reward-and-punishment histories: people come to like what gets rewarded and avoid what gets punished, with no inferential step required. Self-Perception Theory shares the behaviorist commitment to taking observable behavior seriously and rejects the mentalistic apparatus of dissonance, but it parts ways on the inferential step: where behaviorism has people learning attitudes, Bem has people inferring attitudes. The two accounts make different predictions in the cases that matter most for design — particularly the Overjustification Effect, where reward histories produce attitude changes that are exactly opposite of what straight reinforcement would predict, and where Bem’s discounting rule provides the only account of the result. (For the operant-conditioning pillar, see Skinner’s Operant Conditioning: An S-Tier Designer’s Guide.)

Self-Perception Theory in the Real World

The applied record on Self-Perception Theory is large and mixed. The four domains where the theory has done its most consequential applied work are persuasion and compliance (foot-in-the-door, low-ball), behavior-change interventions (smoking, exercise, voting), brand and consumer attitudes, and behavioral economics (commitment devices, default opt-ins).

Persuasion and Compliance

The foot-in-the-door technique is the clearest applied success of Self-Perception Theory. Burger’s (1999) review and the subsequent meta-analytic literature consistently recover small-to-moderate effect sizes (Cohen’s d ≈ 0.20–0.30) across hundreds of compliance studies in fundraising, consumer behavior, public-health intervention, and political mobilization. Cialdini’s parallel commitment-and-consistency principle (in Influence, 1984) is the popular-press articulation of the same mechanism, repackaged for negotiation and sales practitioners.

The flip side is the door-in-the-face technique (Cialdini et al. 1975), which produces similar compliance outcomes through a different mechanism (reciprocal concession rather than self-attribution) and which therefore behaves differently across moderators. Door-in-the-face works best with the same requester, immediately; foot-in-the-door works across delays, with different requesters, in any direction so long as the small request was sincere. The mechanism difference matters: foot-in-the-door builds a self-concept the user takes with them; door-in-the-face builds a transactional obligation to the requester.

The low-ball technique (Cialdini, Cacioppo, Bassett & Miller 1978) is a third compliance procedure with a Bem-style mechanism: get an initial commitment under favorable conditions, then increase the cost — and the actor maintains the commitment because reversing it would contradict the now-active self-concept of “the kind of person who agreed to this.” Low-ball is a more aggressive applied version of foot-in-the-door, and the ethical concerns it raises (the favorable-conditions step is functionally a deception) are the reason most academic ethics review boards would not approve the original 1978 protocol today. The mechanism is the same; the ethical packaging is the only thing that differs.

Health and Civic Behavior Change

The applied health-behavior literature has used self-perception logic to design smoking-cessation, exercise-adoption, and voting-mobilization interventions. The clearest examples are the public-commitment paradigms: Pallak & Cummings (1976) had homeowners publicly commit (their names would appear in a published list) versus privately commit (no public record) to reduce energy consumption. Public committers reduced consumption substantially more in the short run, and the effect persisted for months — substantially longer than informational interventions alone produced. The self-perception interpretation: the public commitment behavior was unambiguously freely chosen, and the actor inferred the corresponding identity (“I am the kind of person who conserves energy”) from the public record.

The voting-mobilization literature provides the cleanest contemporary example. Bryan, Walton, Rogers & Dweck (2011) showed in two field studies that asking pre-election survey respondents about “being a voter” (noun-framed identity question) produced higher subsequent voting rates than asking about “voting” (verb-framed behavior question), consistent with self-perception’s emphasis on identity inference rather than behavior repetition. The effect has been partly contested in subsequent replication work (Gerber, Huber, Doherty, Dowling & Panagopoulos 2016, with smaller effects), but the underlying mechanism — that identity-loaded language activates self-perception circuitry differently from neutral behavioral language — survives even the more skeptical readings.

Brand and Consumer Attitudes

The consumer-behavior literature treats Self-Perception Theory as a foundational mechanism behind loyalty programs, sample-and-trial marketing, and the entire family of “small first purchase” tactics. A user who tries a free sample, attends a free webinar, signs up for a free trial, or buys a low-cost entry-level product is given a behavior to point at when their self-concept gets recruited into a larger purchase decision later. The mechanism is the same as Freedman & Fraser’s foot-in-the-door, repackaged for commercial conversion funnels.

The empirical record in commercial domains is mixed. The cleanest demonstrations come from controlled field experiments (Gneezy & List 2006 charity-asking, Karlan & List 2007 matching-grant) where the researchers can rule out the obvious confounds. Less-controlled marketing case studies often confuse foot-in-the-door effects with selection effects (the people who took the free sample were already higher-intent users) and produce inflated estimates of the technique’s marginal contribution. The honest read for designers is that the foot-in-the-door mechanism is real, the effect size is small-to-moderate, and the technique is reliably worth deploying when it costs little to implement and the small request is structured to produce genuine self-attribution.

Default Effects and Commitment Devices

The behavioral-economics literature on defaults (Madrian & Shea 2001 401(k) auto-enrollment; Johnson & Goldstein 2003 organ-donor opt-in vs opt-out; Thaler & Sunstein 2008 Nudge) is sometimes interpreted through a self-perception lens, though the more common reading is straight behavioral-economic friction (defaults stick because changing them is costly). The self-perception reading adds a second mechanism: a user who remained in the default has a behavior history of “remained in the default,” which under attribution-theoretic logic produces an inferred self-concept (“I am the kind of person who is comfortable with this default”) that subsequently makes deliberate opt-out feel costly in identity terms as well as in friction terms. The two mechanisms are difficult to disentangle empirically, and the design implication is the same regardless: defaults are not neutral; they are doing identity work whether or not the designer intended them to.

The commitment-device literature (Beshears, Choi, Laibson & Madrian 2015 review in the Annual Review of Economics; Ariely & Wertenbroch 2002 self-imposed deadlines; stickk.com behavioral commitment contracts) operates explicitly on self-perception logic: the user commits publicly to a future behavior, the public commitment becomes the behavior to point at, and the inferred identity (“I am the kind of person who keeps their commitments”) supplies the motivational pull through to execution. The empirical record is genuinely mixed — commitment devices work for some people on some behaviors, and not others — but the moderators map cleanly onto Bem’s boundary conditions: works best when the commitment is freely chosen, public, identity-loaded, and not easily attributable to external pressure.

The Elephant in the Room

The honest practitioner’s question about Self-Perception Theory is not whether the mechanism is real — fifty-eight years of foot-in-the-door, low-ball, and overjustification research have settled that — but whether the mechanism is ethically deployable in production behavioral design. The theory describes a cognitive process by which people manufacture self-concepts from behaviors that were elicited under external incentives the user could be talked out of attributing to themselves. That is, by any honest reading, a mechanism for engineered self-deception. And the techniques that operationalize it (foot-in-the-door, low-ball, public commitment, default opt-in, identity-noun framing) are routinely used to nudge users into states that serve the designer’s interests at least as much as they serve the user’s.

The most-discussed ethical concerns in the applied literature: foot-in-the-door is the mechanism behind cult-recruitment progressions; low-ball is the mechanism behind price-anchoring negotiation tactics; public commitment is the mechanism behind multi-level marketing identity capture; default opt-ins are the mechanism behind the worst dark-pattern subscription traps the FTC has prosecuted. The theory itself is not the problem — Self-Perception Theory is descriptive, not prescriptive. The problem is that the mechanism the theory describes is genuinely consequential, and consequential mechanisms attract bad actors.

The Octalysis Black-Hat / White-Hat distinction is the design discipline that emerges from taking this seriously. A White-Hat self-perception design move asks for a small commitment to a behavior the user genuinely values, then gives the user a record of having done it that supports the self-concept they already wanted to build. The Octalysis pillar on the Overjustification Effect describes the canonical Black-Hat overshoot mode: any reward layered onto a CD3-driven activity that pushes the user’s self-attribution from “I do this because I want to” to “I do this because I’m getting paid” damages the very self-concept the design was supposed to be building. The Cognitive Evaluation Theory audit on every reward is, structurally, the design discipline of asking whether the self-perception inference your mechanic is producing is the inference the user actually wants to be making.

The practical heuristic that follows: if your mechanic depends on the user not noticing what you are doing, you are operating in the Black-Hat zone. If your mechanic works equally well whether or not the user understands the inferential machinery underneath it, you are operating in the White-Hat zone. The foot-in-the-door technique is in the gray zone — it works whether or not the user sees the mechanism, but it works substantially better when the user has not been primed to discount their own compliance as engineered. That ambiguity is why mature behavioral-design practice tends to disclose the mechanism in advance (“we’ll start with a small step, then build up”) and let the user opt into the progression rather than be slid down it. The disclosure does not eliminate the foot-in-the-door effect — the small commitment is still genuine — but it eliminates the engineered-deception component of the mechanism, and the technique still produces meaningful compliance lift.

How to Apply Self-Perception Theory with the Octalysis Framework

Self-Perception Theory is the trait-level shadow of Core Drive 4 (Ownership & Possession) inside the Octalysis Framework. Where Core Drive 4 is the design surface that builds users’ sense that something is theirs, Self-Perception Theory describes the inferential mechanism by which that ownership feeling actually forms — the user watches themselves act, applies the discounting rule to rule out external causes, and updates the self-concept to match the behavior. Every CD4 mechanic is, structurally, a self-perception scaffold.

The eight Core Drives map onto Self-Perception Theory along three structural axes: which Core Drives are the inferential destinations (where the inferred self-concept lands), which Core Drives are the behavioral sources (where the behavior gets generated), and which Core Drives are the discounting risks (where external attributions get inserted that block the self-perception inference). Designing self-perception flows means routing the user’s behavior history through each of these axes deliberately.

CD4 Ownership & Possession as the Inferential Destination

The cleanest mapping is between Self-Perception Theory and Core Drive 4 (Ownership & Possession). The four canonical CD4 Game Techniques — Avatar #4, Collection Set #37, Endowed Effect #23, Alfred Effect #27 — are all explicit self-perception scaffolds. The Avatar (Game Technique #4) is a customization surface where every choice the user makes gets recorded in a visible representation; the user looks at the avatar, sees a record of choices that nobody coerced or paid for, and infers the self-concept those choices imply. The Collection Set (Game Technique #37) gives the user a behavioral record across time — every collected item is a behavior the user took freely — and the inferred identity (“I am a collector of X”) accumulates with the set. The Endowed Effect (Game Technique #23) puts something nominally in the user’s possession before any commitment has been made; the user observes themselves treating the object as theirs, and the resulting self-attribution is “I own this,” which is the Bem-derivable result.

The Alfred Effect (Game Technique #27) — the personalized-treatment surface where the system remembers the user’s preferences and reflects them back — is the most architecturally interesting case. The user is shown a record of their own past behavior dressed up as the system’s “knowledge of them.” The self-attribution the user makes from that record (“the system knows I am the kind of person who…”) is precisely the inference Self-Perception Theory predicts: the user is treating their own behavior history as evidence about themselves, with the system serving as the visible-mirror surface that makes the history available for self-inspection.

CD2 Accomplishment and CD3 Creativity as the Behavior Sources

Where CD4 is the inferential destination, Core Drive 2 (Development & Accomplishment) and Core Drive 3 (Empowerment of Creativity & Feedback) are the most reliable behavior sources for self-perception flows. CD2 mechanics (Step-by-Step Tutorials #13, Quick Win #36, Crowning #16, Pattern Recognition #43) generate a steady stream of small completed behaviors the user can point at. CD3 mechanics (Meaningful Choices #11, Boosters #31, Real-Time Control #52) generate freely-chosen creative behaviors that are particularly hard to attribute to external coercion — exactly the data Bem’s mechanism most readily uses.

The structural design move is to route CD2 and CD3 behaviors through CD4 visibility surfaces. A user who completes a tutorial step (CD2) does the self-perception work most reliably when the completion is recorded in a visible behavior history (CD4 Collection Set, profile badge, public progress bar). A user who makes a creative choice (CD3) does the self-perception work most reliably when the choice is recorded in a persistent artifact (CD4 Avatar, saved configuration, public portfolio). The mechanic is not the behavior alone or the visibility alone; it is the routing of the behavior into the visibility surface.

CD6, CD7, CD8 as the Discounting Risk Zones

Core Drive 6 (Scarcity & Impatience), Core Drive 7 (Unpredictability & Curiosity), and Core Drive 8 (Loss & Avoidance) are the design surfaces where Bem’s discounting rule most reliably blocks self-perception. A behavior elicited by a Countdown Timer (Game Technique #65, CD6) gives the user a visible external cause to attribute the behavior to (“I only did that because the timer was about to expire”); the inferred self-concept does not form. A behavior elicited by a loot-box variable-reinforcement schedule (CD7) gives the user a visible external cause (“I only did that because the schedule was telling me to”); the self-perception inference fails. A behavior elicited by an avoidance-of-loss surface (CD8) gives the user a visible external cause (“I only did that because I would have lost something”); the self-perception inference is blocked.

The design implication is direct: the more your CD6 / CD7 / CD8 surface is doing the work of motivating a behavior, the less self-perception will. The user is telling themselves the behavior happened because of the external pressure, not because of who they are. The 4 Experience Phases supply the staging: in Discovery and Onboarding, where the user is forming their self-concept of “what kind of user I am,” self-perception work belongs in CD2/CD3/CD4 routing. CD6/CD7/CD8 mechanics belong in Scaffolding (where the self-concept is already in place and the design needs to keep the user engaged through difficult middle stretches) and in Endgame (where the experienced user can read the mechanics for what they are and the self-perception layer has already done its work).

The Octalysis Self-Perception Audit

Treat every behavior your mechanic elicits as raw material for a self-perception inference. The audit is four questions per behavior delivery: does the user have a visible record of having taken this behavior freely; can the behavior be attributed to a salient external cause that the discounting rule will use to block the self-inference; does the inferred self-concept your mechanic is building support the regulation continuum the design is targeting; and does the cumulative behavior history across the surface produce a self-concept the user would endorse if they read the history back to themselves outside the product context.

The audit cuts both ways. On the affirmative side, the audit tells you to design visible behavior histories deliberately — not just for the next click, but for the user’s eventual self-inference from the accumulated record. On the protective side, the audit tells you to watch for any mechanic that produces behavior at the cost of the inferred self-concept the design is supposed to be building: that is the Black-Hat overshoot the Octalysis Framework keeps warning about, and Self-Perception Theory is the academic floor under that warning.

Practical Steps to Apply Self-Perception Theory

Eight concrete moves that fall out of the audit, in roughly the order a working designer encounters them.

1. Start the user with a small, freely-chosen, identity-loaded behavior. The first interaction matters disproportionately because it is the first piece of evidence the user has about themselves in your product context. Make the small ask substantial enough to require a real decision (not a default click), congruent with the larger ask the design will eventually make, and free of obvious external coercion. The CD3 (Empowerment of Creativity & Feedback) Game Techniques are the cleanest behavior source: a meaningful choice, a creative configuration, a personally relevant input.
2. Make the behavior visible to the user. A behavior the user cannot inspect cannot do self-perception work. The CD4 (Ownership & Possession) visibility surfaces — Avatar #4, Collection Set #37, Endowed Effect #23, Alfred Effect #27 — are the cleanest visibility scaffolds. The behavior history the user can scroll back through is the data they will reverse-engineer their attitude from.
3. Use foot-in-the-door progressions, but disclose the progression. The technique is empirically robust at small-to-moderate effect sizes; it is also ethically suspect when used covertly. Disclose the progression in advance (“we’ll start small and build up”) and let the user opt into the larger commitment. The technique still works (the small commitment is still genuine), and the disclosure removes the engineered-deception component.
4. Eliminate external causes the discounting rule can latch onto. Every salient external cause in the surface is a candidate explanation the user can use to discount their own behavior. Watch especially for CD6 (Countdown Timer #65), CD8 (loss-avoidance prompts), CD5 social-pressure surfaces, and obvious payment / reward layers. Each of these reduces the size of the self-perception inference even as it increases the immediate compliance.
5. Use identity-noun framing in copy. Bryan, Walton, Rogers & Dweck (2011) and a decade of follow-up work on identity-language framing has shown that noun-form identity language (“be a voter”) activates self-perception circuitry differently from verb-form behavior language (“vote”). The effect sizes are smaller and more contested in the recent replication record than the original studies implied, but the underlying mechanism — that identity-loaded language activates self-attribution — is consistent with Bem’s account and worth deploying as a free design lever.
6. Reinforce the self-concept through continuous behavior history. Self-perception-built attitudes are weaker and less stable than introspection-grounded attitudes. The design implication is that the self-concept the onboarding flow built has to be reinforced over time through a continuous record of subsequent behaviors that support it. The Collection Set #37 and the Alfred Effect #27 surfaces are the cleanest reinforcement mechanics.
7. Run the Cognitive Evaluation Theory audit on every reward. Every extrinsic reward the design layers onto a behavior is a potential discounting cue that lets the user reattribute their pre-reward intrinsic interest to the reward — the Overjustification Effect. The audit is whether the reward reads informationally (affirms competence, autonomy, relatedness) or controllingly (treats the user as a transactional agent). Verbal informational rewards do not damage self-perception. Tangible expected engagement-contingent rewards do.
8. Measure the self-concept, not just the immediate behavior. Standard product analytics measure the engagement-during-reward; Self-Perception Theory predicts the consequential differences will show up after the reward is removed and the user has to act from the inferred self-concept alone. Long-tail free-choice behavior, multi-month retention conditional on incentive removal, and the differential between users who internalized the self-concept and users who learned to expect the reward — these are the measurements that diagnose whether your self-perception flow is actually working.

Closing Thoughts

Self-Perception Theory survived the dissonance wars not by displacing Festinger but by claiming a smaller, sharper, more empirically defensible territory: the formation of attitudes when no settled attitude yet exists. That is the empirical region most product designers actually operate in, which is why the theory has remained a load-bearing piece of behavioral-design practice even as the unifying-theory ambition Bem originally proposed has been pared back.

The right read for the working designer is not “self-perception explains everything” or “self-perception is unreliable since the facial-feedback collapse.” The right read is that Bem identified a real, narrow, consequential mechanism — people manufacture self-concepts from behaviors they cannot easily attribute to external causes — and that the mechanism scales from the individual onboarding flow up through the multi-month retention curve. The design implication is that the visible behavior history is the single most underdesigned surface in most products. Designers spend disproportionate effort on the next click, the next reward, the next push notification — and disproportionately little effort on the artifact the user can scroll back through and use as evidence about themselves.

The discipline that follows is concrete. Start the user with a small freely-chosen behavior. Make the behavior visible. Disclose foot-in-the-door progressions instead of running them covertly. Eliminate external attributions that block the discounting rule. Use identity-noun framing where natural. Reinforce the self-concept through continuous behavior history. Audit every reward for its functional significance. Measure free-choice behavior and post-incentive retention, not just engagement during the reward period. Treat the visible behavior surface as a first-class design object, not as an analytics afterthought. The mechanism Bem identified is real; the question is whether the design is building the self-concept the user actually wants to have, or one the design needed to install. If you want the design-side framework that operationalizes this insight across every product surface, that’s the Octalysis Framework — and the Core Drive most directly downstream of Self-Perception Theory is Core Drive 4: Ownership & Possession.

Frequently Asked Questions

What is Self-Perception Theory in simple terms?

Self-Perception Theory, proposed by Daryl Bem in 1967, holds that when internal cues are weak or ambiguous, people figure out their own attitudes the same way an outside observer would — by watching their own behavior and the situation it occurred in. The popular shorthand is “we don’t sing because we’re happy; we’re happy because we sing,” and the theory’s clearest applied example is the foot-in-the-door technique: comply with a small request, infer that you are the kind of person who supports the cause, and become much more likely to agree to a larger related request later.

How is Self-Perception Theory different from Cognitive Dissonance Theory?

Festinger’s dissonance theory posits a motivated drive state — psychological discomfort triggered by inconsistency that people are motivated to reduce. Bem’s Self-Perception Theory posits a non-motivated inferential process from behavior to attitude. The Fazio, Zanna & Cooper (1977) resolution showed both theories are correct in their proper domain: dissonance fires when behavior contradicts a held attitude (outside the latitude of acceptance); self-perception fires when behavior is congruent with held attitudes or when no held attitude exists. The two theories now coexist as complementary mechanisms rather than rival accounts.

Is the foot-in-the-door technique actually backed by Self-Perception Theory?

The foot-in-the-door technique was demonstrated by Freedman & Fraser in 1966, a year before Bem’s theoretical paper. Bem’s self-perception interpretation became the canonical mechanism explanation: residents who agreed to a small civic-safety request inferred they were the kind of people who supported civic-safety causes, and the inferred self-concept supplied the motivational pull for the much larger subsequent request. Burger’s 1999 review found small-to-moderate effect sizes (Cohen’s d ≈ 0.20–0.30) across roughly 60 follow-up studies, with the effect strongest when the small request is substantial, identity-congruent, and free of obvious external pressure.

Did the facial-feedback hypothesis survive replication?

The 2016 Registered Replication Report on Strack, Martin & Stepper’s 1988 pen-in-mouth study coordinated 17 labs (n ≈ 1,894) and recovered an effect indistinguishable from zero. The original textbook claim — that smiling reliably makes you happier and frowning reliably makes you sadder — has not survived in its original strong form. Subsequent meta-analyses (Coles, Larsen & Lench 2019, k = 138; Coles et al. 2022 19-laboratory project) recovered small mean effects (g ≈ 0.13–0.17) consistent with a real but context-dependent and considerably smaller phenomenon than the original studies implied.

What is the Overjustification Effect and how does it relate to Self-Perception Theory?

The Overjustification Effect (Lepper, Greene & Nisbett 1973) is the finding that paying nursery-school children to draw cut their free-play drawing in half once the rewards were removed. The self-perception interpretation: children watching themselves draw for a reward applied Bem’s discounting rule and reattributed their pre-reward intrinsic interest to the extrinsic incentive. The Deci, Koestner & Ryan (1999) meta-analysis pooled 128 experiments and found a reliable d ≈ -0.40 negative effect on free-choice intrinsic motivation for tangible expected engagement-contingent rewards. The Overjustification Effect is structurally Self-Perception Theory applied to the case where the external cause is the reward.

What does Self-Perception Theory imply for product onboarding?

The design implication is that onboarding flows are not just teaching users how to use the product — they are manufacturing the user’s self-concept of “what kind of user I am.” Whatever behavior the onboarding elicits in the first few minutes becomes the data the user will reverse-engineer their attitude toward the product from. The right design moves are to start with small freely-chosen identity-loaded behaviors, make those behaviors visible to the user in a persistent record (the CD4 Ownership Game Techniques inside the Octalysis Framework), and avoid layering external pressure (CD6 Scarcity, CD8 Loss-Avoidance) onto the early flow because that pressure gives the discounting rule something to latch onto and blocks the self-inference.

How does Self-Perception Theory connect to identity-noun framing in voting and behavior change?

Bryan, Walton, Rogers & Dweck (2011) showed in two field studies that asking pre-election survey respondents about “being a voter” (noun-framed identity language) produced higher subsequent voting rates than asking about “voting” (verb-framed behavior language), consistent with self-perception’s emphasis on identity inference rather than behavior repetition. The original effect sizes have been partly contested in subsequent replication work (Gerber et al. 2016 found smaller effects), but the underlying mechanism — that identity-loaded language activates self-attribution circuitry differently from neutral behavioral language — has held up across the more skeptical readings.

Can Self-Perception Theory be used ethically in product design?

Yes, but the discipline matters. The mechanism the theory describes — manufacturing self-concepts from behaviors the user cannot easily attribute to external causes — is genuinely consequential and has historically been used in cult-recruitment progressions, multi-level marketing capture, and dark-pattern subscription traps. The Octalysis Black-Hat / White-Hat distinction is the design discipline that emerges from taking this seriously: a White-Hat self-perception flow asks for a small commitment to a behavior the user genuinely values, makes the resulting behavior history visible, and lets the user opt into the larger progression. Disclosing the foot-in-the-door progression in advance does not eliminate the effect (the small commitment is still genuine) but does eliminate the engineered-deception component of the technique.

What are the boundary conditions that make Self-Perception Theory work?

The theory fires most reliably under four conditions. First, internal cues are weak or ambiguous — the attitude has not yet formed, or the existing attitude is uncertain. Second, the behavior is freely chosen — no one is paying for it, forcing it, or watching with an obvious agenda the actor can use to discount their own behavior. Third, the behavior is identity-loaded enough to be diagnostic — a trivial click does not produce a self-attribution, but a meaningful choice does. Fourth, the behavior is visible to the actor in a persistent way — a behavior the user cannot inspect cannot do self-perception work. Outside these conditions the effect attenuates rapidly, which is why most failed self-perception interventions are failed boundary-condition checks.

Is Self-Perception Theory still considered current in social psychology?

The theory’s core mechanism — observer-style attribution from behavior to attitude under conditions of weak introspective cues — has held up across nearly six decades of empirical work and remains a standard component of social-psychology curricula. The strong unifying claim Bem originally proposed (that Self-Perception Theory could replace dissonance theory entirely) did not survive the Fazio-Zanna-Cooper 1977 boundary-condition resolution. The textbook facial-feedback demonstration that anchored the emotion-self-perception literature did not survive the 2016 Registered Replication Report. What remains is a precise, narrow, consequential local theory of attitude formation that continues to do real work in the foot-in-the-door, low-ball, public-commitment, and identity-framing literatures.

References

1. Bem, D. J. (1965). An experimental analysis of self-persuasion. Journal of Experimental Social Psychology, 1(3), 199–218.
2. Bem, D. J. (1967). Self-perception: An alternative interpretation of cognitive dissonance phenomena. Psychological Review, 74(3), 183–200.
3. Bem, D. J. (1972). Self-perception theory. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology (Vol. 6, pp. 1–62). Academic Press.
4. Festinger, L., & Carlsmith, J. M. (1959). Cognitive consequences of forced compliance. Journal of Abnormal and Social Psychology, 58(2), 203–210.
5. Fazio, R. H., Zanna, M. P., & Cooper, J. (1977). Dissonance and self-perception: An integrative view of each theory’s proper domain of application. Journal of Experimental Social Psychology, 13(5), 464–479.
6. Freedman, J. L., & Fraser, S. C. (1966). Compliance without pressure: The foot-in-the-door technique. Journal of Personality and Social Psychology, 4(2), 195–202.
7. Burger, J. M. (1999). The foot-in-the-door compliance procedure: A multiple-process analysis and review. Personality and Social Psychology Review, 3(4), 303–325.
8. Cialdini, R. B., Cacioppo, J. T., Bassett, R., & Miller, J. A. (1978). Low-ball procedure for producing compliance: Commitment then cost. Journal of Personality and Social Psychology, 36(5), 463–476.
9. Cialdini, R. B., Vincent, J. E., Lewis, S. K., Catalan, J., Wheeler, D., & Darby, B. L. (1975). Reciprocal concessions procedure for inducing compliance: The door-in-the-face technique. Journal of Personality and Social Psychology, 31(2), 206–215.
10. Lepper, M. R., Greene, D., & Nisbett, R. E. (1973). Undermining children’s intrinsic interest with extrinsic reward: A test of the overjustification hypothesis. Journal of Personality and Social Psychology, 28(1), 129–137.
11. Deci, E. L., Koestner, R., & Ryan, R. M. (1999). A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 125(6), 627–668.
12. Strack, F., Martin, L. L., & Stepper, S. (1988). Inhibiting and facilitating conditions of the human smile: A nonobtrusive test of the facial feedback hypothesis. Journal of Personality and Social Psychology, 54(5), 768–777.
13. Wagenmakers, E.-J., Beek, T., Dijkhoff, L., Gronau, Q. F., Acosta, A., Adams, R. B., Jr., et al. (2016). Registered Replication Report: Strack, Martin, & Stepper (1988). Perspectives on Psychological Science, 11(6), 917–928.
14. Coles, N. A., Larsen, J. T., & Lench, H. C. (2019). A meta-analysis of the facial feedback literature: Effects of facial feedback on emotional experience are small and variable. Psychological Bulletin, 145(6), 610–651.
15. Bryan, C. J., Walton, G. M., Rogers, T., & Dweck, C. S. (2011). Motivating voter turnout by invoking the self. Proceedings of the National Academy of Sciences, 108(31), 12653–12656.
16. Mitchell, J. P., Banaji, M. R., & Macrae, C. N. (2005). The link between social cognition and self-referential thought in the medial prefrontal cortex. Journal of Cognitive Neuroscience, 17(8), 1306–1315.

Related Reading

• The Octalysis Framework: A Complete Guide to Gamification
• Cognitive Dissonance: An S-Tier Behavioral Designer’s Guide
• The Overjustification Effect: An S-Tier Behavioral Designer’s Guide
• Self-Determination Theory: An S-Tier Behavioral Designer’s Guide
• Attribution Theory: An S-Tier Behavioral Designer’s Guide
• Locus of Control: An S-Tier Behavioral Designer’s Guide
• Reactance Theory: An S-Tier Behavioral Designer’s Guide
• Actionable Gamification (book)

WOULD YOU LIKE YU-KAI CHOU TO WORK WITH YOUR ORGANIZATION?