The Barry Manilow T-Shirt
In the late 1990s, a Cornell undergraduate volunteered for what she thought was a study on memory. She showed up at the lab, was handed a t-shirt, and told to put it on before entering a room full of other students completing questionnaires.
The t-shirt featured a large image of Barry Manilow.
The researchers — Thomas Gilovich, Victoria Medvec, and Kenneth Savitsky — had pre-tested the shirt with a separate group of Cornell students. They had asked these students to rate, on a 1-to-13 scale, how embarrassed they would be to wear the shirt across campus. The Manilow shirt scored a 9.5. It was, by college-undergraduate consensus, embarrassing.
The participant wore the shirt into the room. She sat for a few minutes among five or six other students who were busy filling out unrelated questionnaires. Then she was pulled out and asked a single question: of the people in that room, what percentage do you think will be able to recall what was on your shirt?
Her answer, and the answers of the dozens of participants who followed her through the same procedure, averaged about 46 percent. Roughly half of the people in the room, the t-shirt wearers believed, would remember Barry Manilow’s face.
The researchers then walked into the room and asked the other students. The actual recall rate was about 23 percent.
The participants had overestimated how much they stood out by a factor of two.
That gap — between how visible we feel and how visible we actually are — is the spotlight effect. Gilovich, Medvec, and Savitsky introduced it formally in 2000, and despite a quarter-century of the replication crisis tearing through social psychology and toppling more famous effects, this one has survived. It replicates across cultures, paradigms, populations, and decades. It is one of the more robust findings the field has produced.
It is also one of the most useful, if you actually take its implications seriously.
What Gilovich, Medvec, and Savitsky Did
The 2000 paper in the Journal of Personality and Social Psychology was not a single experiment but a series of five. The Barry Manilow study is the most famous, but the structure of the paper matters: each study isolated a different prediction the spotlight effect makes.
Study 1 was the Manilow shirt. The headline finding was the 46 percent versus 23 percent gap, but the paper also tested whether participants who were given more time to acclimate (and presumably feel less conspicuous) showed smaller overestimates. They did, but only modestly. The effect persisted.
Study 2 ran the same paradigm but with shirts featuring people the experimenters predicted students would feel good about wearing — Bob Marley, Martin Luther King Jr., Jerry Seinfeld. The prediction was that the spotlight effect should hold for positive distinctiveness too, not just embarrassment. It did. Participants overestimated how many others would recall their flattering shirt by roughly the same factor.
This was important. If the effect were purely about embarrassment, the explanation would be motivational — anxiety amplifying perceived attention. But the same overestimate appeared when participants wanted to be noticed. The mechanism had to be cognitive, not just affective.
Studies 3 through 5 extended the paradigm beyond appearance. Study 3 had participants engage in a group discussion and then estimate how much their contributions — both positive and negative — would be recalled by other group members. Study 4 looked at video game performance in groups. Study 5 tested estimates of how much others would notice variability in one’s behavior over time.
Across all five studies, the same pattern held. People consistently overestimated the salience of their own actions and appearance to others, by roughly a factor of two.
The proposed mechanism was anchoring and adjustment. We have unmediated access to our own self-consciousness — we know, vividly, that we are wearing the shirt, that we made the comment, that we stumbled over the word. From that anchor of high salience to ourselves, we adjust downward to estimate how salient we are to others. But the adjustment is insufficient. We don’t adjust far enough.
This is the same general mechanism Daniel Kahneman and Amos Tversky had identified two decades earlier in their work on anchoring biases more broadly. The spotlight effect is, in this framing, just one specific application of a general feature of human cognition: starting points are sticky.
The 2002 Replication and Extension
Two years later, Gilovich, Kruger, and Medvec published a follow-up in the Journal of Experimental Social Psychology that tightened the original findings and pushed on the mechanism.
The 2002 paper tested whether people overestimate not just how visible they are at a single moment but how variable their visible behavior appears over time. If you have a bad day at work and a good day at work, do other people notice the swing as much as you do?
The answer was no. People perceive the manifest variability of their own actions and appearance — how much they appear to fluctuate to outside observers — as substantially greater than it actually is. Just as the original spotlight effect found that we overestimate the visibility of any single moment, the 2002 extension found that we overestimate the visibility of the contrast between our moments.
This matters for one of the most common forms the spotlight effect takes in practice: replaying a single bad moment and assuming it was a defining moment in someone else’s perception of you. The data say no. Not only did they notice it less than you think, but they also noticed less of a contrast between that moment and your other moments.
The 2002 paper also addressed a methodological concern. Some critics had wondered whether the Cornell undergraduate population was uniquely prone to self-consciousness, or whether the Barry Manilow shirt was somehow uniquely salient. Gilovich and colleagues ran the paradigm with different stimuli, different populations (including older non-students), and different timing structures. The effect persisted. Effect sizes varied somewhat — the Manilow shirt’s 2x overestimate sat near the high end — but the direction was always the same. People always overestimated, sometimes by a small margin, sometimes by a large one. They never underestimated.
The Illusion of Transparency
Before the 2000 paper, Gilovich and Savitsky had been working on what they called the illusion of transparency — the belief that one’s internal states are more visible to others than they actually are. A 1998 paper had shown that liars believe their deception is more obvious than observers actually perceive it to be, and that anxious public speakers believe their anxiety is more visible than audiences detect.
The 1999 Current Directions in Psychological Science paper connected the two phenomena explicitly. The spotlight effect is about overestimating the visibility of one’s actions and appearance — external things. The illusion of transparency is about overestimating the visibility of one’s internal states — feelings, intentions, attempts at deception. Both share the same egocentric anchoring mechanism: we start from our own vivid experience and adjust insufficiently.
The unification matters because it predicts where the effect should be strongest. Any time there is a large gap between what we experience internally and what we display externally, the overestimate should be larger. A liar feels the deception intensely; an observer sees only the surface. A nervous speaker feels the racing heart; the audience sees only the standing-still body. The internal-versus-external gap is exactly where the anchoring goes wrong.
Clinically, this turned out to matter a great deal.
Savitsky 2001 — The Mistakes Extension
In 2001, Savitsky, Epley, and Gilovich published a paper in the same journal as the original that pushed the spotlight effect into the territory people actually care about: failures, shortcomings, and mishaps.
The Barry Manilow shirt is funny but low-stakes. What about the cringe-inducing things — the comment in a meeting that landed wrong, the typo in an email to the whole company, the moment you mispronounced someone’s name? Do we overestimate how much others judge us for these too?
The 2001 paper ran a series of studies. In one, participants were asked to imagine themselves committing various social blunders — tripping in public, giving a bad presentation, saying something tactless — and to predict how harshly observers would judge them. They were then compared to actual observers who watched recordings of these events.
The pattern was again robust. Self-judgers consistently predicted that observers would view them more harshly than observers actually did. The gap was substantial. People who imagined themselves making a mistake estimated negative judgments from observers at levels far exceeding the actual judgments those observers reported.
The paper proposed an explanation rooted in the same anchoring mechanism but with an additional twist: when we imagine ourselves making a mistake, we cannot avoid simulating how it feels from the inside. That simulation includes vivid awareness of all the things we did wrong, all the ways we wished we had behaved differently, all the self-criticism we are inflicting on ourselves in the moment. Observers experience none of this. They see only the surface. The asymmetry of access — full from the inside, partial from the outside — produces the systematic overestimate.
The Savitsky paper also suggested an asymmetric effect: positive moments are also overestimated in visibility, but the gap is smaller than for negative moments. We are particularly prone to overestimating how much our failures are noticed. This asymmetry maps onto loss-aversion patterns Kahneman and Tversky had documented elsewhere — negative outcomes loom larger across the board.
Clinical Implications — Social Anxiety and CBT
This is where the spotlight effect stopped being a curiosity and became therapeutically important.
Social anxiety disorder, as it appears in the clinical literature, is characterized in significant part by the conviction that one’s social behavior is being constantly evaluated and that one’s flaws are being constantly observed. The clinical phenomenology maps almost directly onto an extreme form of the spotlight effect. The socially anxious patient believes their blushing is conspicuous, their stumbling over words is glaring, their nervousness is broadcast on their face. Frequently, none of these things are as visible as the patient believes.
Cognitive-behavioral therapy for social anxiety has, since at least the early 2000s, integrated direct interventions targeting the spotlight effect and the illusion of transparency. A common protocol involves having patients make explicit predictions before a social interaction (how nervous I think I will look, how badly I think this will go), then collecting feedback from actual observers afterward (rated independently), then comparing. The systematic gap between prediction and reality becomes therapeutic data. The patient sees, repeatedly, that their internal experience of visibility does not match the external reality.
This protocol works in part because the spotlight effect is itself so robust. If the underlying bias were small or unreliable, CBT couldn’t use it as a teaching target. The fact that almost every patient consistently overestimates their visibility, in predictable directions, by predictable magnitudes, is what makes the intervention reliable.
Behavioral experiments in CBT — for instance, deliberately spilling something at a coffee shop and then asking the patient to count how many people noticed — are essentially Gilovich’s Barry Manilow study repurposed as therapy. The student version goes: predict the recall, then measure the recall, then notice the gap. The therapeutic version goes: predict the judgment, then test the judgment, then notice the gap.
Twenty-five years of clinical practice have not weakened the underlying empirical effect. If anything, the consistency of the result across thousands of patients in therapy has provided one of the largest informal replications of any social psychology finding ever produced.
Workplace and Performance Applications
Outside of clinical contexts, the spotlight effect has implications for performance anxiety in workplace settings — and these implications are mostly unused.
Public speaking is the obvious case. The standard advice that “the audience can’t see how nervous you are” is, it turns out, empirically accurate. Gilovich and Savitsky’s transparency work showed that the gap between perceived and actual visibility of anxiety is large. Speakers who feel their hands are shaking visibly are observed by audiences who notice nothing of the kind. The advice has been right for decades; the spotlight effect explains why.
Performance reviews and feedback contexts are another. Employees who made a single visible mistake — a public presentation that went poorly, a project that missed a deadline — frequently believe the mistake defines their reputation in the organization. Managers, asked to recall the same employee’s recent year, often barely mention the incident. The asymmetry of attention is structural: the employee replays the moment internally many times; the manager experienced it once.
Personal-brand strategy is a case where the spotlight effect cuts in counterintuitive directions. Founders and operators worry intensely about the visibility of their imperfect content, their inconsistent posting, their off-brand moments. The base rate of how much any of this is actually noticed by the audience is substantially lower than the worrier believes. The implication is not that quality doesn’t matter, but that the marginal cost of imperfection is being systematically overestimated, and the marginal cost of inaction (not posting because it isn’t perfect) is being systematically underestimated.
Risk-taking decisions also bend under spotlight analysis. People avoid making bold moves — pitching the unconventional idea, sending the cold email, applying for the stretch role — in part because they overestimate how much the failure of those moves will be visible to and remembered by others. The empirical reality is that most failures are noticed by fewer people, less intensely, for shorter duration, than the actor believes. This isn’t a license to make sloppy moves. It is a correction factor for the calibration of personal risk.
The Strategist Takeaway
The spotlight effect is one of the rarer findings in social psychology that combines three properties: it is empirically robust, it is mechanistically explained, and it is practically actionable.
For someone making decisions under social visibility — which is most decisions — the spotlight effect implies a specific correction:
When you are trying to predict how much others will notice your action, take your initial estimate and roughly halve it.
This is not a precise heuristic; the actual overestimate factor varies across contexts. But it is approximately right, and it is approximately right in a predictable direction. You will almost never err by underestimating how much others notice you. You will almost always err by overestimating.
The applications follow:
If you are calibrating how much to invest in avoiding a possible embarrassment, halve your estimate of the embarrassment’s visibility and recalculate.
If you are debating whether to take a public risk, halve your estimate of how visible the failure case would be and recalculate.
If you are managing social anxiety in any of its forms — from clinical anxiety disorders to mundane awkwardness aversion — the spotlight effect provides empirical justification for a specific cognitive correction: your internal experience is more vivid than your external presentation.
If you are spending energy worrying about a single bad moment from your past — a meeting comment, an awkward interaction, a public mistake — apply the 2002 finding: not only did fewer people notice than you think, but the contrast between that moment and your other moments is also smaller in their perception than in yours.
If you are building a personal brand or public-facing professional presence, the spotlight effect predicts that you are over-investing in perfection and under-investing in frequency. The audience notices imperfection less than you fear. The audience notices absence more than you assume.
If you are evaluating how a competitor or peer will react to a strategic move, the spotlight effect cuts the other way: you may be overestimating how much they will notice and react to your move. This applies to product launches, pricing changes, hiring announcements, and content strategy shifts. The competitor’s spotlight is on themselves, not you.
The unifying principle is that the asymmetry of self-attention versus other-attention is large, systematic, and consistently underweighted by the unaided intuition. Twenty-five years after Gilovich, Medvec, and Savitsky introduced the effect, and after most of social psychology’s better-known findings have failed to replicate, the Barry Manilow shirt still does the same job. You feel more conspicuous than you are. Others are looking at themselves.
Sources
- Gilovich, T., Medvec, V. H., & Savitsky, K. (2000). The spotlight effect in social judgment: An egocentric bias in estimates of the salience of one’s own actions and appearance. Journal of Personality and Social Psychology, 78(2), 211-222. https://doi.org/10.1037/0022-3514.78.2.211
- Gilovich, T., Kruger, J., & Medvec, V. H. (2002). The spotlight effect revisited: Overestimating the manifest variability of our actions and appearance. Journal of Experimental Social Psychology, 38(1), 93-99. https://doi.org/10.1006/jesp.2001.1490
- Savitsky, K., Epley, N., & Gilovich, T. (2001). Do others judge us as harshly as we think? Overestimating the impact of our failures, shortcomings, and mishaps. Journal of Personality and Social Psychology, 81(1), 44-56. https://doi.org/10.1037/0022-3514.81.1.44
- Gilovich, T., & Savitsky, K. (1999). The spotlight effect and the illusion of transparency: Egocentric assessments of how we are seen by others. Current Directions in Psychological Science, 8(6), 165-168. https://doi.org/10.1111/1467-8721.00039
Related
- The False-Consensus Effect — The companion egocentric bias, where we overestimate how much others share our views.
- The Dunning-Kruger Effect — Another self-perception bias from the late 1990s, but one with substantially weaker post-replication standing.
- Hindsight Bias — The other classic egocentric distortion, applied to past events rather than present visibility.
- Confirmation Bias — The selective attention bias that compounds with the spotlight effect when we look for evidence others noticed our mistakes.
- Festinger’s Social Comparison Theory — The 1954 foundation for thinking about self-perception in social contexts.
FAQ
Did the spotlight effect replicate?
Yes, and substantially. The 2000 paper has been replicated across populations, cultures, and paradigms for a quarter-century. Effect sizes vary, but the direction is consistent: people overestimate the visibility of their own actions and appearance, never the reverse. It is one of the more robust findings in social psychology.
Is the spotlight effect just anxiety?
No. The 2000 paper specifically tested this. When the t-shirt featured someone participants felt good about wearing (Bob Marley, MLK, Jerry Seinfeld), the same overestimate appeared. The effect operates on positive visibility too, which means the mechanism is cognitive (anchoring on self-perspective) rather than purely affective.
How does the spotlight effect relate to the illusion of transparency?
They share a mechanism — egocentric anchoring with insufficient adjustment — but apply to different content. The spotlight effect is about overestimating the visibility of external things (actions, appearance). The illusion of transparency is about overestimating the visibility of internal states (feelings, intentions, anxiety). Gilovich and Savitsky’s 1999 paper formally unified them.
Is the 2x overestimate the consistent finding?
Roughly, but not always. The Barry Manilow shirt produced a 46% versus 23% gap, which is a 2x overestimate. Other paradigms have produced smaller and larger overestimates. The 2x figure is a rule-of-thumb anchor for the typical magnitude, not a precise universal constant. The consistent finding is the direction, not the exact ratio.
How is the spotlight effect used in clinical practice?
Cognitive-behavioral therapy for social anxiety uses spotlight-effect-style behavioral experiments as a core intervention. Patients make explicit predictions about how visible their anxiety, mistakes, or shortcomings will be in upcoming social situations, then test those predictions against actual observer feedback. The reliable gap between predicted and actual visibility provides therapeutic data.
Does the spotlight effect apply to digital and remote contexts?
The original studies were in-person. Research on the spotlight effect in digital contexts — social media posting, video calls, online presentations — has generally found the same direction of overestimation, though the magnitudes can differ. The mechanism (anchoring on self-perspective) should apply regardless of medium, and the empirical evidence largely supports that.
What is the practical takeaway for risk-taking?
Halve your estimate of how visible the failure case will be. If you are deciding whether to make a public move — pitch, post, apply, present — and your hesitation is driven by how much you fear others will notice the failure, the spotlight effect suggests you are overestimating that visibility by a substantial margin. This is not license for sloppiness; it is a calibration correction for risk aversion driven by social visibility fears.