In 1998, the psychologists Brad Bushman and Roy Baumeister ran an experiment that directly tested a popular idea: that people who lash out or dominate a room are secretly suffering from low self-esteem. Participants first filled out a standard self-esteem questionnaire and a separate measure of narcissism. Then each one wrote a short essay, and a supposed fellow participant, actually a research assistant, gave it either a mild compliment or a harsh insult. Afterward, each participant got the chance to blast that same person with unpleasant noise through headphones, choosing the volume and duration themselves.
Plain self-esteem score, on its own, predicted almost nothing about how loud or long someone chose to blast the person who had insulted them. What did predict it was narcissism. People who scored high on narcissism and got insulted delivered dramatically more aggressive blasts than anyone else in the study. The paper, published in the Journal of Personality and Social Psychology, described this as threatened egotism: aggression aimed less at defending a low opinion of oneself than at defending an inflated one that has just been challenged.
This is one influential pair of experiments, not the final word on why people act loud or domineering, and it directly complicates the tidy version of the claim that gets repeated online. The people most likely to blast a critic were not the insecure ones by a simple self-esteem measure. They were the ones who already rated themselves unusually highly and could not tolerate having that rating questioned.
Not all high self-esteem is the same
A separate line of research helps explain why “high self-esteem” alone is such an unreliable predictor. The psychologist Michael Kernis and his colleagues at the University of Georgia distinguished between what they called secure and fragile forms of high self-esteem, describing the difference in a 2008 paper in the Journal of Personality. Participants whose self-esteem was stable over time, and whose stated self-regard matched a separate implicit measure of self-worth, showed especially low levels of verbal defensiveness when challenged. Participants whose high self-esteem was unstable, dependent on daily events, or mismatched with their implicit self-worth showed considerably more defensiveness under the same conditions.
Put next to the Bushman and Baumeister findings, a more precise picture starts to form. It is not that confidence causes loudness or that insecurity causes quiet. It is that a self-regard which depends on winning every exchange, staying on top, or not being questioned tends to produce a much louder, more defensive reaction when that exchange does not go the way the person needs it to. A steadier form of self-regard, one that does not rise and fall with each individual interaction, does not require the same defense.
Where the “hiding something” idea gets tested more directly
A different body of research tests a related but distinct question: not what happens when someone is challenged, but what happens when their sense of identity is threatened more broadly. The sociologist Robb Willer and colleagues ran a set of four studies, described in a 2013 paper in the American Journal of Sociology, testing what they called the masculine overcompensation thesis. In the first experiment, men who were given false feedback suggesting they scored as more feminine than average responded by expressing more support for war, more homophobic attitudes, and more interest in buying an SUV than men given no such feedback. Women given equivalent feedback about scoring as more masculine showed no comparable shift.
A later study in the same paper found that men who felt threatened this way also expressed a stronger desire to move up in status hierarchies, and a large survey sample found that men who believed social change was undermining men’s status reported more support for domination-oriented and combative attitudes generally. This is a single research team’s line of work, tested across four separate studies rather than one, and later researchers attempting to replicate parts of it have reported mixed results on some specific measures, so it should be read as a real and repeatedly tested pattern rather than a settled law about all display of dominance.
What “hiding” actually seems to mean here
Taken together, this research does not support the specific idea that loud people are secretly low in self-esteem, which is the version that circulates most widely and is not really what any of these studies found. What it does support is a narrower and, in some ways, more useful claim: displays of dominance, whether raising your voice, escalating a conflict, or performing extra confidence, tend to show up most reliably not when someone feels good about themselves in a stable way, but when a specific, threatened version of their self-image is on the line and needs defending in that moment.
That is a meaningfully different claim from “they’re secretly insecure deep down.” It says the volume is often a response to a live threat to a specific self-image, whether that image is about intelligence, status, or masculinity, rather than a permanent window into someone’s private self-worth. The same person can hold a stable, secure sense of themselves in most areas of life and still react loudly the moment one particular, precarious piece of that self-image is challenged.
What this does not prove
None of these studies were run outside a laboratory or survey setting, and none of them followed the same people over time to see whether early loudness predicts later insecurity or the reverse. The Bushman and Baumeister aggression measure was a noise blast in a controlled experiment, not a real-world confrontation, and the masculine overcompensation studies focused specifically on threats to gender identity rather than dominance display in general. None of this research licenses a confident read on any specific loud person in your own life. What it offers instead is a clearer account of the actual mechanism the popular claim is gesturing toward: not simple low self-esteem, but self-regard that has become dependent on winning a particular contest, and reacts accordingly when that contest is lost.
It is also worth being honest about what these three lines of research do not have in common. Bushman and Baumeister measured aggression in response to a direct personal insult. Kernis and his colleagues measured verbal defensiveness under a milder kind of challenge. Willer and his colleagues measured a broader set of attitudes following a threat to gender identity specifically, not a one-on-one confrontation at all. They point toward a shared underlying pattern rather than describing one single, unified mechanism that has been proven to operate identically across every kind of loud or dominant behavior. Reading them together is a reasonable synthesis, not a claim that a single study settled the question.