Nine hundred and seventy-six volunteers, half of them women, all in identical lightweight clothes, all sitting at the same undemanding task. Pooled from Nevins's work of 1966 and Fanger's, their preferred temperatures came to 25.4 °C for the men and 25.8 °C for the women. A later trial by Fanger and Langkilde, published in 1975 with 32 volunteers of each sex, produced 25.0 and 25.1 °C. Those experiments are the foundations beneath ISO 7730 and its sister standards, a point Bjarne Olesen made in the REHVA Journal in 2015.
Olesen, at the time a professor at DTU, the Danish technical university, was replying to a paper from Maastricht printed that same year in Nature Climate Change. Its authors, Kingma and van Marken Lichtenbelt, pointed out that rules on indoor climate depend on a comfort model dating from the 1960s, whose default metabolic values were taken from an average man and could overstate a woman's metabolism by up to 35 per cent; buildings, on that account, might be inherently non-energy-efficient in giving women comfort. The measurements behind their case came from sixteen women doing sedentary work. The study included no men, and Olesen wrote that it should have. The unit itself has a history. Byrne and colleagues recalled in 2005, in the Journal of Applied Physiology, that one met was first worked out from the oxygen a single 40-year-old man of 70 kg used at rest, and that in a large, heterogeneous group of subjects it overshot actual resting consumption by 35 per cent on average.
My reading of the quarrel is simple. On the figure to dial in, the sexes stand close together; the real divergence is in what a departure from that figure costs each person. Part of the chill women describe at work is settled beyond any thermostat's reach: by how near a desk sits to glazing, to a supply grille or to a cold stretch of floor, and by whether the person at that desk has any means of altering the air around it.
Sensitivity to drift is where the sexes diverge
Sami Karjalainen, of Finland's VTT research centre, reviewed the research on gender and thermal comfort for the journal Indoor Air in 2012. Neutral temperature, the point where nobody feels warm or cool, showed no significant gap between men and women in most of the studies he reviewed. Yet clearly more than half the laboratory and field investigations recorded greater dissatisfaction among women sharing an identical environment with men, and his pooled calculation gives a ratio of 1.74. Women, in his summary, react more strongly to a departure from the optimum, notably once the air cools, and on average need a greater say in their own conditions, including the means to adjust them. Olesen offers a seasonal footnote: where summer air conditioning leaves women chillier than men, he puts most cases down to clothing, since women adapt their clothing to the season better, while men stay in suit and tie.
British practice reads the thermometer beside the worker
The legal text in Great Britain is brief. Regulation 7 of the Workplace (Health, Safety and Welfare) Regulations 1992 says that throughout working hours, temperatures in indoor workplaces "shall be reasonable". Since 2002 it has also demanded adequate thermal insulation where the kind of work and its physical activity make that necessary, and that "excessive effects of sunlight on temperature shall be avoided". The regulation also requires enough thermometers for people at work to check conditions themselves.
HSE's Approved Code of Practice, L24, second edition, 2013, carries the detail. Departing from it is not an offence as such, though a court may treat it as proof of a breach unless the employer demonstrates compliance by another route. Paragraph 60 frames the goal as "reasonable comfort without the need for special clothing". Paragraph 61 supplies a floor of normally 16 °C, 13 °C for rigorous physical effort, then concedes that such figures may still fall short of comfort once air movement, humidity and clothing are considered. Its closing instruction is precise: readings "should be taken close to workstations, at working height and away from windows". No legal upper limit exists; HSE explains that heat in many workplaces comes from the work rather than the weather. Should reasonable comfort prove unattainable across a whole workroom, paragraph 65 calls for local heating or cooling.
I regard paragraph 61 as the Code's most valuable line for anyone planning offices, because it moves the thermometer to where somebody actually works. The Code already reads the temperature at the desk, so where each desk goes belongs on the layout drawing.
Four checks for each spot on the drawing
Alongside its whole-body comfort index, ISO 7730 names four sources of localised discomfort: draught; a vertical gradient that leaves the head in warmer air than the feet; flooring that runs either too warm or too chilly; and radiant asymmetry, meaning that the body trades heat unequally with surfaces on opposite sides of it. Mapped onto an office, each becomes a test of a location. On a January morning, a desk hard against a wide pane gives up warmth on one flank to cold glass. A desk below a ceiling diffuser takes the airflow. A desk tucked into a corner above a cold slab keeps its user's feet cooler than the rest of them. In each case the desk sits where ISO 7730 locates discomfort, whatever the thermostat's average says, and Karjalainen's findings imply that the gap bears hardest on the more sensitive. The practical step follows from paragraph 61: take the four checks to the drawing and settle where the office workstations will stand before the order goes in, not once the delivery lorry has left.
Control is the second matter, and L24 has already named the remedy: where a workroom cannot be made comfortable throughout, local heating or cooling (paragraph 65). Karjalainen adds that women, on average, need a greater say over their own conditions. A workspace and open plan zoned from the outset gives that say a place: a sheltered area away from glass and supply air, a brisker one, and the freedom to move between them instead of one setting that everybody argues over.
Before the furniture: five stops with a thermometer
Stop one, desk height. Where L24 wants readings: near the workstation, clear of windows (paragraph 61).
Stop two, the glazing. One flank turned towards a cold pane, which ISO 7730 calls radiant asymmetry.
Stop three, beneath the grilles. Draught, one of the standard's four local complaints.
Stop four, down by the ankles. Vertical temperature difference, plus the floor underfoot (ISO 7730).
Stop five, the occupant. Karjalainen's finding that women, on average, need more individual control (2012).
One laboratory experiment has tested temperature against performance, and its caveats matter as much as its result. Chang and Kajackaite, reporting in PLOS ONE in 2019, tested 543 Berlin students in a laboratory held at a different setting each session, somewhere between roughly 16 and 33 °C. Warmer rooms lifted women's scores in maths and verbal exercises and nudged men's slightly downwards; every additional degree went with a 1.76 per cent increase in correct maths answers from women, whereas the male decline was slight and statistically insignificant. Extra effort seems partly responsible, the authors suggest, because the women's output of answers rose while their error rate did not. Scores on a cognitive reflection test did not move. Sessions lasted an hour on average, subjects were students, and it was room temperature that changed, never desk position; I treat this as a sign that warmth can affect performance, nothing more.
L24 has already shown facilities managers where to stand with their thermometers. Pass that same instruction to whoever draws the plan, before a single desk is ordered, and judge the outcome in paragraph 60's own terms: can everybody on that floor work comfortably during the first cold week of the year in what they wore to get there?
Quick answers
Do women really feel the cold more at work?
In most studies reviewed by Karjalainen (2012), neutral temperature showed no significant difference between the sexes, yet women proved more sensitive to temperatures straying from the optimum, chiefly towards the cool end, and were more likely to express thermal dissatisfaction, with a pooled ratio of 1.74. Kingma and van Marken Lichtenbelt (2015) argued that the comfort model's standard metabolic figures, derived from an average man, may overstate women's metabolism by up to 35 per cent.
How cold may a UK office legally be?
Regulation 7 of the Workplace (Health, Safety and Welfare) Regulations 1992 demands a reasonable temperature. HSE's Code of Practice (L24, 2013) sets a normal minimum of 16 °C, or 13 °C for rigorous physical effort, measured near workstations, at working height and away from windows. HSE states that no legal maximum exists.
Are comfort standards biased towards men?
Bjarne Olesen (REHVA Journal, 2015) reported that the experiments beneath ISO 7730 recruited men and women in equal numbers, and that in identical clothing and activity they preferred nearly identical temperatures, 25.4 °C for men and 25.8 °C for women. Kingma and van Marken Lichtenbelt note that the model's default metabolic values come from an average man.
Where should desks go to avoid cold spots?
Check each position for the four kinds of local discomfort listed in ISO 7730: draught, a head-to-foot temperature gradient, a floor that is uncomfortably cold or warm, and radiant asymmetry, for instance beside a large window. If one workroom cannot be made comfortable throughout, paragraph 65 of L24 asks for local heating or cooling.