The Nuclear Reactor

Proximity Creates Attraction — the Science Is Not Soft

54%

of workers report having experienced attraction to a colleague at some point in their career

This is not weakness or moral failure. Robert Zajonc’s mere exposure effect (1968) established that repeated exposure to any stimulus reliably increases how much we like it — faces included. Add shared goals, shared stress, and mutual evaluation, and you have conditions more intimacy-generating than most social contexts, including bars, apps, and parties.

Festinger, Schachter & Back (1950) studied a housing complex and found that 65% of residents’ close friendships were with immediate neighbors. Physical proximity was the single largest predictor of closeness — stronger than shared interests, background, or personality. Offices replicate this effect five days a week.

Most of It Never Converts — and the Job Is Better for That

5:1

ratio of people who felt workplace attraction to those who acted on it

The gap between feeling and acting is where most of the value lives. Unacted-on attraction between colleagues is a documented driver of higher effort, sharper communication, more careful presentation, and better mood — what organizational psychologists call positive affect spillover.

Think of it as plutonium in a nuclear reactor: the energy is real, it makes the whole system run hotter, and the job of professionalism — like cadmium control rods — is not to eliminate the energy but to keep it doing useful work.

36%

of workers have dated a coworker at some point in their career

31%

of those who dated a coworker went on to marry them

~6%

left or changed jobs because of a workplace romance that ended badly

CareerBuilder Workplace Romance Survey (2012, 2018). Based on nationwide U.S. samples of employed adults.

The Risk Architecture — Variance, Not Probability

Risk matrix: power dynamic vs. career risk

This is a variance problem, not a probability problem. Most workplace relationships don’t ruin careers — but when they do, the damage is asymmetric and rapid. The right question is not “will this work?” but “am I willing to restructure my professional life if it doesn’t?”
Power dynamic is the primary structural variable. Peer–peer relationships have a very different risk profile than manager–report. Most company policies target the latter specifically — not out of moralism, but because the institutional liability is real and separate from whatever the individuals actually feel.
Industry culture matters. A startup’s tolerance for blurred personal/professional lines is not the same as a hospital’s or a law firm’s. Know your environment before you know your intentions.
Disclosure policies exist; use them. Post-2017, approximately 78% of Fortune 500 companies have some form of romance disclosure requirement. If one exists at your company and you’re in it, the cover-up carries more career risk than the relationship.
Some power couples have dominated. The outlier successes are real — but they are outliers. The people who came out ahead had genuine alignment, managed perception carefully, and often worked in cultures tolerant of it. They also got lucky with timing.

Bottom Line

Romance is not a hard no in professional networks. But the energy that makes it possible is also what makes the job worth doing — and most of the time, the better move is to let it run the reactor.

The workplace generates genuine attraction because it is designed to: repeated proximity, shared purpose, and mutual evaluation are exactly the conditions under which human bonding evolved. Over half of workers have felt it with a colleague. Most never act. Of those who do, a meaningful fraction end up married — and a small but real fraction end up changing jobs. The expected value of acting depends entirely on two things: the structural risk (peer or hierarchical, disclosed or hidden, industry culture or policy environment) and the honest answer to one question — is this a possible forever thing? If the answer is yes, the risk calculus changes. If the answer is “I think so” or “maybe,” the reactor is probably better off running at operating temperature. The tension, managed well, is an asset. The meltdown is not recoverable.

Sources: Zajonc, “Attitudinal Effects of Mere Exposure,” Journal of Personality and Social Psychology (1968) · Festinger, Schachter & Back, Social Pressures in Informal Groups (1950) · Rosenfeld, Ruan & Falcon, “Disintermediating your friends,” PNAS (2019) / Stanford HCMST data 2017–2019 · CareerBuilder Workplace Romance Survey (2012, 2018), nationwide U.S. employed-adult samples · SHRM Workplace Romance Policy data (estimated, 2023–2024) · Funnel middle stages (serious relationship, marriage) are CareerBuilder-derived calculations; “positive affect spillover” framing from Fisher et al. (2008) organizational affect literature. Some figures are directional industry-sourced estimates.