The foundational work on this was published in the late 1970s and described what happens to people who are numerically rare in an organisational setting. Its core observations have held up unusually well.
Three consequences were identified, and each has since been studied further.
Heightened visibility
A person who is the only one of her kind in a group is more noticed, remembered and discussed than her peers.
The consequence is asymmetric. Errors are more memorable and more likely to be attributed to the category rather than to the individual. Successes are frequently attributed to the category as well — treated as evidence about the group rather than about the person.
This creates a performance environment in which the cost of a mistake is higher, which is precisely the condition the risk literature identifies as producing more conservative behaviour. A person taking fewer risks in that environment is responding correctly to the incentives.
The visibility also generates a workload that does not appear in any job description: being asked to represent, to appear in materials, to sit on committees requiring representation, and to advise on matters relating to the category. This is real time, it is rarely counted, and it comes out of the same hours as the actual work.
Contrast and boundary heightening
The presence of a rare individual tends to make the majority group more aware of its own commonality, which can produce exaggerated group behaviour and, in the original formulation, the informal exclusion of the rare member from social contexts where relationships and information are exchanged.
The information cost is the substantive one. A substantial proportion of workplace information moves informally, and exclusion from those channels is a genuine professional disadvantage rather than a social slight.
Assimilation to role
The third effect is the tendency for the rare individual to be perceived through a limited set of available stereotypes rather than as an individual — with a corresponding narrowing of the roles available to her.
Later research on role congruity has developed this: where the expectations attached to a category and those attached to a role conflict, occupants of that category in that role are evaluated less favourably, and the conflict is particularly sharp for leadership roles.
The numerical threshold
The original work proposed that these effects are strongest when a group makes up a small minority and diminish as proportion rises, with meaningful change occurring somewhere in the region of a fifth to a third.
The precise thresholds have not held up as fixed quantities, and later research suggests the dynamics are more contextual than the original formulation implied. But the general shape — that these are effects of rarity rather than of the trait itself, and that they attenuate as numbers rise — has considerable support.
The practical implication is important: the experience of being one of two is qualitatively different from being one of one, and the difference is larger than the numbers suggest.
The single-candidate finding
A related and much-cited result concerns hiring. Analyses of finalist pools have found that when only one candidate from an underrepresented group is in a pool, the probability of that candidate being selected is very low — and that it rises substantially when two or more are present.
The proposed mechanism is that a single non-default candidate is read as an exception to be justified, whereas two make the presence of such candidates unremarkable.
This finding has been influential because it implies an intervention that requires no change in evaluation standards at all — only in the composition of the shortlist.
What helps the individual
Documentation. Where attribution is unreliable, contemporaneous written records of contribution are the available correction, and they matter more in this position than in any other.
Explicit accounting of representational work. Being asked to sit on three committees for reasons of representation is a real allocation of time and can be raised as such — ideally in the same conversation as workload rather than as a grievance.
External reference groups. Where the immediate environment provides no comparison, professional networks outside it supply the calibration that the workplace does not. This is the most consistently reported benefit in qualitative studies of women in male-dominated technical roles.
And a clear-eyed reading of the situation. The single most reported relief in this literature is discovering that the experience is structural and documented rather than personal — that the difficulty of the position is a known property of the position.
What helps the organisation
Numbers, mostly. These are effects of rarity, and the reliable solution to rarity is to stop being rare.
Which is why interventions aimed at supporting the sole representative — while worth doing — are palliative. The condition being treated is the composition of the room, and everything else is management of its symptoms.