The results of a new meta-analysis, recently published in the journal Building and Environment, challenge an adopted industry standard which cited an optimal indoor temperature to promote work performance, based on an earlier study from 2006. The new meta-analysis was completed in 2021 by an international team of thermal comfort researchers, with authors from CBE, Denmark and Singapore, who followed the methods of the earlier study, but improved upon them using additional data and rigorous statistical methods. The new study found no evidence for a relationship between temperature and work performance within the range temperatures commonly found in office buildings, and certainly none that should be adopted as an industry recommendation. The complete results are available in this open access journal paper; shorter summaries are available in this video presenting the results and in this LinkedIn post.
The research team analyzed 358 performance measures from 35 studies on temperature and work performance (the term now preferred over productivity) published between 1946 and 2020. They then tested the data using 12 regression models and machine-learning algorithms. The results showed that while any of the models could be applied to the data set, they all had very low prediction accuracy, including a ‘zero-order model,’ essentially a model in which there is no relationship at all between variables.
These results stand in contrast to an earlier 2006 study that concluded that worker performance would peak at 72°F (22°C), and that performance will decline as temperatures deviate either warmer or cooler from this temperature. That finding had been adopted in the ASHRAE Handbook—Fundamentals, a key reference and standard for designers of building mechanical systems in North America and beyond, and in the European counterpart, the REHVA Guidebook 6. The study had been widely cited, and used for example to make calculations related to life-cycle cost analysis of building services, and on the economic implications of indoor temperatures.
The authors of the new study conclude that the guidance proposed in the 2006 paper, and all the models tested in the new study, ‘should not be used to predict the effect of temperature on work performance due to their very low prediction power,’ and further that they ‘should not be used to design or control buildings.’ Undoubtedly it will take both time and advocacy to revise the many guidelines that were based on the 2006 study. The authors are working with the committee responsible for the chapter on thermal comfort in the ASHRAE Handbook— Fundamentals to remove the existing model, and also to write a new section on the complex relationship between the thermal environment and work performance.
Finally, the authors have taken additional steps to help others who are interested in such analysis, and also to promote transparency around the research. They created an online interactive tool for exploring relationships between temperature and office work performance, including various regression models, temperatures, and factors such as site climate, performance metrics, and task complexity. They also published this database that was used in this study.