10.6078/D1G42R
Porras-Salazar, Jose Ali
0000-0001-8388-1904
Berkeley Education Alliance for Research in Singapore Limited
Schiavon, Stefano
University of California, Berkeley
Wargocki, Pawel
Technical University of Denmark
Cheung, Toby
Berkeley Education Alliance for Research in Singapore Limited
Tham, Kwok Wai
National University of Singapore
Indoor temperature - office work performance database
Dryad
dataset
2021
FOS: Environmental engineering
National Research Foundation
https://ror.org/03cpyc314
2021-07-02T00:00:00Z
2021-07-02T00:00:00Z
en
https://doi.org/10.1016/j.buildenv.2021.108037
81640 bytes
3
CC0 1.0 Universal (CC0 1.0) Public Domain Dedication
The objective of developing this database was to summarise all relevant
published studies that have linked the thermal environment to office work
performance within the most representative temperature range for office
buildings (20 °C to 30 °C). We conducted a comprehensive literature review
and collected the relevant published data into our database. A variety of
combinations of keywords including temperature, thermal sensation, work,
cognitive, and task performance, and office and commercial buildings, were
used. In total, we found thirty-five studies, in 29 peer-reviewed journal
publications and 571 measures of performance, met our inclusion criteria.
We normalised these measures using a method originally proposed by
Seppänen et al. (2006; 2006). This method uses the change in work
performance per 1 °C increment in temperature (λ%), which is measured in
percentage per degree Celsius (%/°C). Our database comprised a total of
358 data points for λ%. Furthermore, we developed a web-based interactive
tool with an easy-to-use interface to visualize the relationship between
temperature and office work performance. This tool can automatically
calculate the model’s equation and accuracy metrics via different data
subsets and regression models.
Our literature search strategies and data collection criteria are
specified below: We searched electronic databases of scientific
publications from September 2019- March 2020, including Google Scholar,
Web of Science, Elsevier, PubMed, and ProQuest. A variety of
combinations of keywords were used. We looked only for peer-reviewed
journal articles that reported both thermal environment measurement data
and the subject’s performance of office work. Diverse measures were
considered to describe office work performance including diagnostic tests,
simulated office work tasks, and existing outcome metrics. We used air
temperature as a proxy of the thermal environment in our research scope
because it was extensively measured in most of the studies. We did not
include the following conditions in our database: Physiological
measurements providing information on cognitive load EEG, ECG, heart rate
variability, and pupillary responses; Studies that reporting only
self-estimated performance; Proxies for reduced performance, such as the
prevalence and intensity of acute health symptoms, especially for fatigue,
difficulty in concentrating, sleepiness, or headaches; Data from factory
workers, university students, or primary/secondary school children; Any
experimental conditions that have a low-temperature condition (TL) below
17 °C or the high-temperature condition (TH) above 36 °C; Studies where
thermal stress was induced by any means other than the indoor thermal
conditions (e.g., exercise or water immersion). From each study, we
retrieved the year of publication, journal, and information regarding the
study location, whether the study was or was not performed in a controlled
environment, the sample size, age group, occupation of the participants,
clothing, and physical activity level. We also collected the tasks or
tests used to measure performance, the performance metrics and outcomes,
and the temperature conditions to which the participants were exposed. We
normalised the performance data obtained from the studies using the method
proposed by Seppänen et al.(2006). This approach assumes that performance
changes linearly within the high temperature (TH) and low temperature (TL)
range examined in each study regardless of the performance measure used
and the temperature range. We calculated the change in work performance in
% per 1 °C increment in temperature (λ%), whereby positive λ% indicates an
increase in performance with increasing temperature; while negative λ%
indicates a decrease in performance with increasing temperature. The
detailed normalisation process can be seen in Porras-Salazar et al.
(2021). References: Seppänen, O., & Fisk, W. J. (2006). Some
quantitative relations between indoor environmental quality and work
performance or health. HVAC and R Research, 12(4), 957–973.
https://doi.org/10.1080/10789669.2006.10391446 Seppänen, O., Fisk, W. J.,
& Lei, Q. H. (2006). Effect of Temperature on Task Performance in
Office Environment. Proceedings of 5th International Conference on Cold
ClimateHeating, Ventilating and Air Conditioning. Moscow, Russia.
Porras-Salazar, J.A.; Schiavon, S.; Wargocki, P.; Cheung, T. &
Tham, K.W. (2021) Meta-Analysis of 35 Studies Examining the Effect of
Indoor Temperature on Office Work Performance. Building and Environment,
203. https://doi.org/10.1016/j.buildenv.2021.108037
Parameter Description Name of column Description Study Study code Authors
Authors of the study Source Journal where the study was published
City_Region City or region where the study was conducted Climate Climate
classification according to the city or region where the study was
conducted. We used the main climate groups (A: Tropical, B: Dry, C:
Temperate, and D: Continental) of the Köppen-Geiger climate
categorisation. Number_Participants The number of people who participated
in the study Male The number of males who participated in the study Female
The number of females who participated in the study Age Mean age of the
people who participated in the study Occupation Main occupation of the
people who participated in the study Anthropometric_Data Tells if
anthropometric data of the participants were reported in the study
(Yes/No). Anthropometric data refers to height, weight or body mass index
(BMI) Health_Status Tells if the health status of the participants was
reported in the study (Yes/No) Clothing_Insulation Shows information about
the clothing insulation as was reported in the studies
Clothing_Insulation2 Shows the clothing insulation in Clo units. For those
studies where only the participants’ attire was reported, we estimated the
corresponding insulation in Clo units using the CBE Thermal Comfort Tool.
https://comfort.cbe.berkeley.edu/ Exposure_Time Duration of exposure per
temperature condition. Time in minutes Measure Name given in the study to
the measure of performance. E.g., Serial digit learning, Text typing, Talk
time, etc. Measure_Type We classified the measures of performance into 23
different types Metric Metrics used in the studies to measure the
participants’ individual performance. FP: False positives, RMSE: Root Mean
Square Error, SC: Scores, NOL: Number of lags, RE: Percentage of correct,
MI: Missed, NOC: Number of correct, NER: Number of errors, REER:
Percentage of errors, SP: Span, NOE: Number of exercises, RT: Reaction
time, T: Time. SA We classified the metrics into Accuracy or Speed based
on the information provided in the publications TSLow Mean thermal
sensation of the study’s participants under the low thermal condition
TSHigh Mean thermal sensation of the study’s participants under the high
thermal condition TLow_C Mean temperature at the low thermal condition in
degree Celsius (°C) THigh_C Mean temperature at the high thermal condition
in degree Celsius (°C) TLow_F Mean temperature at the low thermal
condition in degree Fahrenheit (°F) THigh_F Mean temperature at the high
thermal condition in degree Fahrenheit (°F) PLow Mean performance of the
study’s participants under the low thermal condition PHigh Mean
performance of the study’s participants under the high thermal condition