Project Lukewarm

Project Lukewarm

Overview
Agency	United States Department of Unnecessary Research
Component	Bureau of Thermal Indifference
Duration	June 3, 1991 – September 30, 2019
Directors	Ingeborg Salt-Wheeler (1991–2004) Calvin J. Ostermeyer (2004–2019)
Participants	48,212 occupants
Test sites	1,377 rooms in 44 states
Cumulative cost	$31.4 million
Result
Finding	No temperature eliminates complaints
Nearest approximation	70.9 °F (21.6 °C)
Complaint rate at optimum	11.4%

Project Lukewarm was a research programme of the United States Department of Unnecessary Research (DUR) that sought to identify the exact indoor temperature at which no occupant of a shared space would complain. Conducted by the department's Bureau of Thermal Indifference between 1991 and 2019, it was among the longest-running studies in the department's history, involving more than 48,000 participants in 1,377 rooms at a cumulative cost of approximately $31.4 million.

The programme's final report concluded that no such temperature exists. The temperature producing the fewest complaints, 70.9 °F (21.6 °C), still drew objections from 11.4 percent of occupants, roughly evenly divided between those who found it too warm and those who found it too cold. The finding, sometimes called the Lukewarm limit, has been cited in workplace thermostat disputes, in the Office Microwave Accords and in at least two state legislatures.

Background

Complaints about office temperatures had been among the most common categories of constituent correspondence received by the Office of Marginal Inquiry, the DUR's predecessor, which logged 9,400 such letters between 1961 and 1988. Existing engineering standards defined thermal comfort as a range rather than a point, and acknowledged that some proportion of occupants would always be dissatisfied. The founders of Project Lukewarm regarded this as an unexamined assumption. Ingeborg Salt-Wheeler, the programme's first director, wrote in its 1991 charter that "a range is merely a temperature that has not yet made up its mind".

The project satisfied the three statutory tests of the Unnecessary Research Act of 1987 because, according to the department's grant memorandum, "the answer is unlikely to exist, and if found, nobody will set their thermostat to it".

Methodology

Phase I: the Fixed Room studies (1991–1998)

The first phase took place in 64 conference rooms in federal buildings across nine states. Each room was held at a single temperature for a period of two weeks, after which the temperature was adjusted by 0.1 °F. Occupants recorded complaints on standardised Form BTI-7 ("Thermal Objection"), which offered the options "too warm", "too cold", "both, at different times" and "other".

Phase I encountered difficulties with what researchers termed "pre-emptive objection": a tendency of participants to submit complaints before entering the room, based on the temperature they expected. Approximately 14 percent of BTI-7 forms received during the phase were dated earlier than the occupant's first recorded entry. A further 3,100 complaints were attributed to the forms themselves, which were printed on unusually cold paper.

Phase II: Blind thermostat trials (1999–2008)

The second phase introduced placebo thermostats, which could be adjusted by occupants but were not connected to any heating or cooling equipment. Researchers found that 62 percent of occupants who adjusted a placebo thermostat reported feeling more comfortable within ten minutes, regardless of the direction of adjustment. The finding was considered potentially useful and was referred to the department's Inspector General, who ruled in 2004 that it did not constitute "unauthorized necessity" because placebo thermostats were already in widespread use in commercial buildings.

Salt-Wheeler retired in 2004 and was succeeded by Calvin J. Ostermeyer, a former building engineer at the Hollis Annex.

Phase III: Convergence (2009–2019)

The third and largest phase used 1,180 rooms of varying sizes and functions, including offices, waiting rooms, school classrooms, a bowling alley in Lesser Tumbleton, Minnesota and the passenger cabin of a regional bus. Temperatures were controlled to within 0.02 °F by purpose-built units designated "Goldilocks III". Complaints were recorded electronically by means of two buttons, one blue and one red.

Ostermeyer's team applied an iterative narrowing procedure: after each six-month round, the temperature range tested was reduced to the band in which complaints were lowest. The table below shows the results of selected rounds.

Phase III convergence rounds (selected)

Round	Year	Range tested (°F)	Optimum (°F)	Complaint rate at optimum	Too warm : too cold
1	2009	64.0–78.0	71.2	19.8%	47 : 53
4	2011	68.0–74.0	70.6	15.1%	41 : 59
8	2013	69.5–72.5	71.0	13.0%	52 : 48
12	2015	70.4–71.4	70.8	12.2%	49 : 51
17	2017	70.80–71.00	70.92	11.6%	50 : 50
21	2019	70.88–70.94	70.91	11.4%	50 : 50

By round 17 the complaint rate had ceased to fall in any meaningful way, and the ratio of "too warm" to "too cold" complaints had reached parity. Ostermeyer later described this as "the point at which the room begins to argue with itself".

Findings

The final report, published on September 30, 2019, set out four principal conclusions:

1. No indoor temperature produced a complaint rate of zero in any room with more than two occupants.
2. The temperature producing the fewest complaints was 70.91 °F, rounded in the report to 70.9 °F.
3. At this temperature, complaints were evenly split between "too warm" and "too cold", and approximately 2 percent of occupants submitted both.
4. Complaint rates were more closely associated with the identity of the person who last touched the thermostat than with the temperature.

The fourth finding, which the report called the Last Hand Effect, drew considerable attention. Rooms in which a thermostat had been visibly adjusted by a colleague generated complaint rates up to 34 percent higher than identical rooms at the same temperature in which no adjustment had been observed. The report concluded that "the problem is not the temperature; the problem is each other".

Criticism

The programme's methods were criticised throughout its existence. A recurring objection concerned the reliance on self-reported complaints, which critics argued measured the willingness of participants to complain rather than their thermal comfort. A 2008 study by the Ferncastle Institute found that participants who had been told they were part of a federal study submitted 2.6 times as many complaints as those who had not, and suggested that Project Lukewarm had "discovered the temperature at which Americans complain to the government, which is any".

Others questioned whether a universal temperature could be meaningful across rooms of very different character. The inclusion of the Lesser Tumbleton bowling alley in Phase III was criticised because its occupants were, for most of their visit, engaged in physical exertion; the bureau responded that bowling was "a form of sitting with interruptions". The regional bus was removed from the study in 2014 after it was found that complaints aboard correlated almost entirely with proximity to the door.

The Inspector General's office raised a separate concern in 2016, noting that the convergence procedure would by design produce a single optimum regardless of whether one existed, and that the programme's eventual finding of "no such temperature" had been "essentially guaranteed by round one". Ostermeyer replied that this was "the point", and that confirming something guaranteed was the department's statutory function.

Closure

Secretary Augustin Melby-Carr announced the closure of the programme in a statement on October 2, 2019, describing it as "complete, in the sense that we now know it cannot be completed". The Goldilocks III units were transferred to the Bureau of Redundant Confirmation, which began a confirmation study in 2022 and as of 2026 has reproduced the 70.9 °F result to within 0.03 °F.

Reception

Reaction to the programme was mixed. The Harrisburg Clucker-Gazette described the final report as "a $31 million confirmation of what every office manager already knew". Defenders noted that the department had never claimed otherwise, and that the cumulative cost amounted to less than half a cent per American per year.

Several employers adopted 70.9 °F as a default setting following publication, and the figure was written into an annex of the Office Microwave Accords as a recommended "neutral ambient temperature" for shared kitchens. The department's Inspector General reviewed these adoptions in 2020 and concluded that, since they did not reduce complaints, they did not render the project necessary.

A bill introduced in the Vermont General Assembly in 2021 proposing to make 70.9 °F the mandatory temperature in state offices was withdrawn after committee members could not agree on the temperature of the committee room.

See also

• United States Department of Unnecessary Research
• Unnecessary Research Act of 1987
• The Ben Refrigeration Hypothesis
• Office Microwave Accords
• Thermostat wars