On February 24, 2003, Richard E. Fairfax, director of OSHA's Directorate of Enforcement Programs, signed a memo to the agency's regional administrators and State Plan designees that anyone who has ever fielded a complaint about a cold office should keep in a drawer. "As a general rule, office temperature and humidity are matters of human comfort," it begins. "OSHA has no regulations specifically addressing temperature and humidity in an office setting." A few paragraphs later comes the sentence that closes the question: "OSHA cannot cite the General Duty Clause for personal discomfort."
The memo does give a number. OSHA recommends temperature control between 68 and 76 °F (20 and 24.4 °C) and humidity between 20% and 60%, and it names a second source of guidance, ASHRAE Standard 55, the engineers' standard on thermal comfort. It is a recommendation and it stays one, so in an American office the question of who is cold lands on the people who design, furnish and run the floor.
One version of that question has grown a research literature of its own: why women, sharing a room with male colleagues, voice more complaints about it, and do so above all when it runs cool. My reading of that literature is that men and women want almost the same number on the dial, and that they part ways over how well each tolerates a room that misses it. Some of that gap is settled by where a desk lands relative to a window wall, a supply diffuser or a cold slab, and by whether whoever sits there can adjust their own corner.
The standard in OSHA's memo rests on a disputed model
The challenge to the science behind that standard came from the Netherlands. In a 2015 Nature Climate Change paper, Boris Kingma and his Maastricht University colleague Wouter van Marken Lichtenbelt argued that the rules for indoor climate go back to a thermal comfort model built empirically in the 1960s, and the citation they attached to that claim is ASHRAE 55, 2010 edition: the same standard OSHA's memo names as its second source. A central input of that model is metabolic rate, the heat the body produces. Its default values, the pair wrote, were drawn from the average man, and for women they may run high by as much as 35%. What worried them was the energy bill: in their view, buildings may be inherently inefficient, in energy terms, at keeping women comfortable.
Their own data covered sixteen women at desk work, with no male group beside them. That was one objection from Bjarne Olesen, a professor at Denmark's technical university, writing in the REHVA Journal later that year. He then went back to the experiments underneath ASHRAE 55 and its international counterparts, where subjects of both sexes took part in equal numbers, dressed alike and seated at the same light task. Pooling Nevins (1966) with Fanger, 488 men settled on an average of 77.7 °F (25.4 °C) and 488 women on 78.4 °F (25.8 °C). A 1975 study by Fanger and Langkilde recorded 77.0 °F for men and 77.2 °F for women (25.0 and 25.1 °C). Olesen granted that women are frequently the colder ones in air-conditioned offices during the summer months, and put that down in most cases to wardrobe: they dress for the season, and the men beside them keep the suit and tie.
A wider lens came from Finland. Sami Karjalainen, of the VTT research center, reviewed the evidence for Indoor Air in 2012. Across most of the work he read, the neutral point, where the body registers neither heat nor chill, did not differ significantly by sex. Even so, clearly over half the studies, in labs and in real buildings, caught women reporting more discomfort under identical conditions, and his meta-analysis puts the ratio of women to men in expressing thermal dissatisfaction at 1.74. Put simply, women react more strongly when temperature strays from the optimum, most of all on the cool side, and on average need more say over their own thermal conditions, so the number on the dial barely moves between the sexes while the tolerance for missing it does.
ASHRAE writes its target down: a substantial majority
The society is open about what its standard aims for. According to ASHRAE's own fact sheet on the 2023 edition, Standard 55 specifies conditions that are satisfactory "to a substantial majority (>80%) of occupants within a space." The same sheet, citing a 2021 paper on twenty years of occupant surveys, puts the share of building occupants dissatisfied with their thermal environment at roughly 40%. It adds that building codes generally consider only air temperature, and it states that "increasing the level of personal control over the thermal environment also increases thermal comfort."
A building can therefore meet the standard in full and still leave up to one occupant in five unsatisfied. Put that next to Karjalainen and I read the remainder as something that also depends on where people are seated.
Where does the average stop holding? ISO 7730, which Olesen lists among the standards resting on those same experiments, gives four answers under the heading of local discomfort: moving air, a gap in temperature between head and ankles, flooring that feels too hot or too cold underfoot, and radiant asymmetry, meaning the body loses heat to a surface on one side faster than it does on the other. In its 2023 edition ASHRAE 55 introduced a new way to assess the head-to-ankle gradient. Standard 55 is written for most of a room, and the people it misses sit in particular chairs.
A January walk across the floor plan
The desk against the window wall. Half the body faces a cold pane, the other faces a warm room: radiant asymmetry, in the language of ISO 7730.
The desk under the supply diffuser. It sits in the airstream, and drafts come first on the standard's list.
The corner over an unheated slab. The feet sit at one temperature and the head at another, and the floor itself is one of the four causes.
The desk nobody can adjust. Karjalainen's finding that women need more control and ASHRAE's sentence on personal control point at the same seat.
None of these seats is forbidden, and the thermostat cannot see any of them. That is the reason, in my view, to decide where each unit goes in the same meeting in which the office workstations are chosen, while the floor plan is still a drawing. Control is the second decision. If those who feel a deviation most also need the most say over it, a well-planned workspace & open plan sets aside areas that feel different, one out of reach of glass and air currents, a second where the air circulates more, and lets each person choose where to settle for the day. Flooring deserves a line in the same brief, since the standard names it among the four.
What rides on it was tested, within limits, by Tom Chang and Agne Kajackaite, whose 2019 paper carries the title "Battle for the thermostat." In Berlin, 543 students spent roughly an hour on tests in a lab set, session by session, anywhere from about 61 to 91 °F (16 to 33 °C). Warmer sessions lifted the women's scores on math and verbal problems and nudged the men's down. For women, every added 1.8 °F (1 °C) went with 1.76% more correct math answers; the drop among men was minor and fell short of statistical significance. The authors read part of the gain as extra effort, because the women handed in more answers at an unchanged error rate, and temperature left the cognitive reflection test untouched. These were students in a lab for an hour, a long way from a Tuesday at a desk by the glass. It still shows the dial on the wall reaching the work done under it. One more benefit escapes every study cited here: an office where the setpoint is no longer something colleagues bargain over.
OSHA left the office thermostat to the people who run the building, and ASHRAE says plainly that it designs for a substantial majority. What remains is settled with the floor plan open on the table, placing each desk where the average is true for whoever will sit there in January and giving everyone else a place to move. A good plan makes the seat you were assigned as comfortable as the one you would have chosen.
Quick answers
Are women really colder at work than men?
Karjalainen's 2012 review in Indoor Air found that most studies show no significant sex difference in neutral temperature, yet women react more to departures from it, above all toward the cool side, and the ratio of women to men in expressing thermal dissatisfaction is 1.74. A 2015 Nature Climate Change paper from Maastricht University adds that default metabolic values, taken from the average man, may run up to 35% high for women.
Is there a legal office temperature in the United States?
Not at the federal level under OSHA. A 2003 OSHA interpretation letter says the agency has no regulations specifically addressing office temperature and humidity and cannot cite the General Duty Clause for personal discomfort. OSHA recommends 68 to 76 °F (20 to 24.4 °C) and 20% to 60% humidity, and names ASHRAE Standard 55 as a second source of guidance.
What is ASHRAE 55 designed to achieve?
ASHRAE's fact sheet on the 2023 edition says the standard sets combinations of temperature, radiation, humidity, air speed, clothing and activity acceptable to a substantial majority, more than 80%, of the people in a space. The same edition brought in a new method for judging the temperature gradient from head to ankle.
How do you plan desks so nobody is stuck in a cold spot?
Before seats are assigned, go over the plan with the four local discomfort factors of ISO 7730 in mind: air movement, head-to-ankle gradient, floor temperature and radiant asymmetry. Then provide zones with different conditions and let people move between them; ASHRAE notes that more personal control over the thermal environment raises thermal comfort.