ArticlePDF Available

Effects of thermal discomfort in an office on perceived air quality, SBS symptoms, physiological responses, and human performance


Abstract and Figures

Abstract The effects of thermal discomfort on health and human performance were investigated in an office, in an attempt to elucidate the physiological mechanisms involved. Twelve subjects (six men and six women) performed neurobehavioral tests and tasks typical of office work while thermally neutral (at 22°C) and while warm (at 30°C). Multiple physiological measurements and subjective assessment were made. The results show that when the subjects felt warm, they assessed the air quality to be worse, reported increased intensity of many sick building syndrome symptoms, expressed more negative mood, and were less willing to exert effort. Task performance decreased when the subjects felt warm. Their heart rate, respiratory ventilation, and end-tidal partial pressure of carbon dioxide increased significantly, and their arterial oxygen saturation decreased. Tear film quality was found to be significantly reduced at the higher temperature when they felt warm. No effects were observed on salivary biomarkers (alpha-amylase and cortisol). The present results imply that the negative effects on health and performance that occur when people feel thermally warm at raised temperatures are caused by physiological mechanisms. This study indicates to what extent elevated temperatures and thermal discomfort because of warmth result in negative effects on health and performance and shows that these could be caused by physiological responses to warmth, not by the distraction of subjective discomfort. This implies that they will occur independently of discomfort, i.e. even if subjects have become adaptively habituated to subjective discomfort. The findings make it possible to estimate the negative economic consequences of reducing energy use in buildings in cases where this results in elevated indoor temperatures. They show clearly that thermal discomfort because of raised temperatures should be avoided in workplaces.
Content may be subject to copyright.
Effects of thermal discomfort in an office on perceived air quality,
SBS symptoms, physiological responses, and human performance
Indoor environments should safeguard and enhance
occupantÕs health, comfort, and productivity, as people
spend around 90% of their lives indoors. Roelofsen
(2002) suggests that indoor environmental quality
(IEQ) is more important for the performance of office
workers than job satisfaction and job stress. Occupants
who experience even subclinical symptoms such as
headache and fatigue because of poor IEQ are less
likely to be comfortable and also less likely to be
productive. Thermal environment is one of the most
important indoor environmental factors that affect
health and human performance (Kosonen and Tan,
2004). Several studies have investigated the effects of
the thermal environment on human performance
(Wargocki and Wyon, 2007; Wyon and Wargocki,
2006), but little is known regarding the mechanisms
behind these effects.
Quantitative estimates of the effects of thermal
environment in non-industrial indoor environments
on work can be made by associating the performance
of office work to indoor air temperatures and/or by
associating thermal sensation with human perfor-
mance. The former approach was used by Seppa
et al. (2006) who used the results of published studies,
many of which were summarized by Wyon and
Wargocki (2006), to derive a relationship between
temperature and performance. When using this
approach, it should however be carefully considered
Abstract The effects of thermal discomfort on health and human performance
were investigated in an office, in an attempt to elucidate the physiological
mechanisms involved. Twelve subjects (six men and six women) performed
neurobehavioral tests and tasks typical of office work while thermally neutral (at
22°C) and while warm (at 30°C). Multiple physiological measurements and
subjective assessment were made. The results show that when the subjects felt
warm, they assessed the air quality to be worse, reported increased intensity of
many sick building syndrome symptoms, expressed more negative mood, and
were less willing to exert effort. Task performance decreased when the subjects
felt warm. Their heart rate, respiratory ventilation, and end-tidal partial pressure
of carbon dioxide increased significantly, and their arterial oxygen saturation
decreased. Tear film quality was found to be significantly reduced at the higher
temperature when they felt warm. No effects were observed on salivary bio-
markers (alpha-amylase and cortisol). The present results imply that the negative
effects on health and performance that occur when people feel thermally warm at
raised temperatures are caused by physiological mechanisms.
L. Lan
, P. Wargocki
D. P. Wyon
, Z. Lian
Institute of Refrigeration & Cryogenics, Shanghai Jiao
Tong University, Shanghai, China,
International Centre
for Indoor Environment and Energy, DTU Civil
Engineering, Technical University of Denmark, Kongens
Lyngby, Denmark
Key words: Thermal discomfort; Warmth; Human
performance; Physiological reactions; Health symptoms.
L. Lan
Institute of Refrigeration & Cryogenics, School of
Mechanical Engineering, Shanghai Jiao Tong
University, Shanghai 200240, China
Tel.: +86-21-3420-4263
Fax: +86-21-3420-6814
Received for review 20 September 2010. Accepted for
publication 28 January 2011.
Practical Implications
This study indicates to what extent elevated temperatures and thermal discomfort because of warmth result in negative
effects on health and performance and shows that these could be caused by physiological responses to warmth, not by
the distraction of subjective discomfort. This implies that they will occur independently of discomfort, i.e. even if
subjects have become adaptively habituated to subjective discomfort. The findings make it possible to estimate the
negative economic consequences of reducing energy use in buildings in cases where this results in elevated indoor
temperatures. They show clearly that thermal discomfort because of raised temperatures should be avoided in
Indoor Air 2011; 21: 376–390
Printed in Singapore. All rights reserved
Ó2011 John Wiley & Sons A/S
whether air temperature can be used to estimate the
effects of thermal environment on performance. This
consideration should take into account the results of
Wyon et al. (1975) who showed that the mental
performance of subjects who were clothed for comfort
at two different temperatures of 18.7 and 23.2°C was
unaffected, implying that the state of heat balance is
the driving factor, rather than air temperature per se.
The latter approach was used by Roelofsen (2001). To
create the relationship, he reanalyzed the results of
Berglund et al. (1990) in which the performance of
wireless navy telegraph operators was measured at
temperatures above 29°C, which are much higher than
the temperatures that normally occur indoors. The
relationship derived by Roelofsen is based on too few
experimental data to be a valid alternative. In addition,
with a few exceptions (e.g. Witterseh et al., 2004),
published studies show that the performance of office
work is not affected by thermal discomfort (Haneda
et al., 2009; Lan and Lian, 2009b; Lan et al., 2009a;
Tanabe and Nishihara, 2004). This is in spite of the
potentially negative effects of warmth, which has been
shown to decrease arousal, exacerbate sick building
syndrome (SBS) symptoms, and have negative effects
on mental work (Wyon and Wargocki, 2006). Under-
standing the mechanisms by which performance is
affected would help us to better understand earlier
findings on the effects of thermal environment on
performance and to decide how these effects can be
predicted in the form of a quantifiable relationship, as
well as how thermal conditions indoors should be
Relatively few studies have investigated the mecha-
nisms by which IEQ factors may affect human perfor-
mance (Wargocki, 2010). Tanabe and Nishihara (2004)
measured cerebral blood oxygenation as an indicator
of the rate of cognitive work. They reported that
subjects worked harder to maintain their work perfor-
mance when IEQ (moderately high temperature and
lighting condition under 3 lux) was suboptimal. As a
consequence of this additional work, the subjects
became fatigued. Wargocki et al. (2000) reported that
in air polluted by the emissions from typical building
materials and bioeffluents from humans, the rate of
metabolic CO
production of subjects performing
typical office work was lower, which suggests that they
exerted less effort. Bako
´et al. (2005) reanalyzed
the results of Wargocki et al. (2000) and other studies
and created a relationship showing that metabolic CO
production is lower when the air quality becomes
poorer. They hypothesized that this is the result of
shallow breathing (hypoventilation) in poor air, caus-
ing the CO
level in the blood to increase (respiratory
acidosis), which is known to cause headaches. They
hypothesized that this could be one of the mecha-
nisms explaining why exposure to air of poor quality
reduced performance. Biomarkers in saliva were used
to examine mechanisms by which thermal environment
affects performance; salivary alpha-amylase reflects
stress-related changes in the autonomic nervous system
(Nater and Rohleder, 2009), while salivary cortisol
indicates the hypothalamus–pituitary–adrenal axis
(HPAA) adaptation to stress (Hellhammer et al.,
2009). Willem (2006) and Tham and Willem (2010)
found that salivary alpha-amylase level increased when
tropical subjects were exposed to moderately cold
environments (20 and 23°C compared with 26°C) and
that lower temperatures reduced the concentration of
salivary cortisol. Higher alpha-amylase suggested acti-
vation of the sympathetic nervous system as a response
to cold. This was accompanied by higher mental
arousal as indicated by lower scores in a Tsai–
Partington test, suggesting that salivary alpha-amylase
can be used as an objective indicator of the level of
mental arousal associated with thermal exposures.
Lower cortisol levels suggested better stress relief at
the lower temperature that had resulted in a lower level
of psychological stress. These two biomarkers were
also used by Wargocki et al. (2009) who exposed non-
tropical subjects to moderately elevated temperatures
(29°C compared with 23.5°C). Higher temperatures
gave rise to thermal discomfort because of warmth and
caused salivary alpha-amylase and cortisol to decrease.
The objectives of this study were to investigate
whether warmth affects mental performance, and if so,
to elucidate the mechanisms behind this effect
and whether thermal discomfort is in the chain of
The subjects were exposed in an office to two thermal
conditions created by setting the temperature at 22 and
30°C; the clothing level of the subjects was 0.9 clo at
both temperatures. The clothing level was selected to
keep the subjects thermally neutral at 22°C. To make
sure that the two thermal conditions were sufficiently
different, 30°C was selected, as the relationship derived
by Seppa
¨nen et al. (2006) suggests that a temperature
of 22°C creates conditions for optimal performance
while performance should be considerably reduced at
30°C. A noise level of 50 dB(A) (with no occupants in
the office) and a ventilation rate of 10 l/s per person
were kept constant independently of the temperature
in the office. During each exposure in the office, the
subjects performed tasks typical of office work and
neurobehavioral tests designed to assess different
skills. Physiological parameters and biomarkers were
measured several times during the exposure to explore
the mechanisms by which raised temperatures affect
performance. The subjects reported perceived air
quality, thermal comfort, SBS symptoms, willingness
The effects of thermal discomfort on health and human performance
to exert effort while working, emotion, fatigue, and
workload on several occasions during each exposure in
the office.
The study was carried out in an office (with floor area
of 18 m
and a volume of 55 m
) adapted for exper-
imental purposes (Toftum et al., 2004). Six worksta-
tions were set up in the office. Each workstation
consisted of a table, a chair, a desk lamp, and a laptop.
Webcameras were installed at each workstation to
monitor the subjects. The air temperature in the office
was maintained at the intended level by controlling the
temperature of the air supplied to the office by the
central ventilation system, using a PID controller
connected to a calibrated temperature sensor located
centrally in the space occupied by the subjects. The
room was ventilated by 100% outdoor air using the
mixing principle.
Six women and six men with an average age of
23 ± 2 years were randomly selected from those that
could be recruited from a group of 30 subjects who had
participated in another experiment using similar exper-
imental procedures that had taken place immediately
prior to the present study (Skwarczynski et al., 2009).
The subjects selected were all Caucasian students from
the Technical University of Denmark but of different
nationalities; all tests were therefore presented to them
in English. The subjects were recruited based on the
following criteria: familiarity with computers, impar-
tiality to the office in which the study was carried
out, and absence of chronic diseases, asthma, allergy,
hay fever, and color blindness. The information was
obtained from a questionnaire distributed during
recruitment; none was examined medically.
During the week preceding the experiment, the
recruited subjects attended a 1-h practice and instruc-
tion session at 22°C. During this session, they
rehearsed the experimental procedure. They performed
neurobehavioral tests and tasks simulating office work
similar to those used in the subsequent experiment.
They were instructed on how to fill out the question-
naires used for collecting subjective responses. All of
the physiological measurements were explained to
them. During this session, the subjects were asked to
adjust their clothing so as to feel thermally neutral.
They were then instructed to wear the same clothing in
the subsequent experiments. The subjects were also
instructed what not to do on the experimental day and
on the day prior to each exposure (e.g. drinking
alcohol, overexertion, late to bed).
The subjects were paid a salary for participating in
the experiment at a fixed rate per h; they did not receive
any bonuses. All subjects except one who missed one
session completed all experimental sessions.
Physical measurements. The temperature and relative
humidity, illuminance level, and concentration of CO
in the room were continuously recorded with HOBO
data loggers at each workstation and at the center
of the room. The HOBO data logger has a built-
in temperature (range: )20 to +70°C, accuracy:
±0.7°C), humidity (range: 0–95%, accuracy: ±5%),
and light intensity (range: 0.1–200000 lux, accuracy:
±1 in last digit) sensors and was also connected to a
Vaisala CO
instrument (range: 0–2000 ppm, accuracy:
40 ppm+2% of measured value). Operative tempera-
ture was measured continuously using a VIVO meter
(Dantec Dynamics A/S, accuracy: ±0.5°C) located at
the center of the office. The ventilation rate and
ventilation effectiveness were measured during each
experiment using a LumaSense monitor (Innova 1312
photoacoustic multi-gas monitor, accuracy: ±1%);
tracer gas (sulfur hexafluoride, SF
) and constant
dosing were used. The noise level was measured
repeatedly with an Extech sound level meter [range:
35–130 dB(A), accuracy: ±2 dB(A)] during the exper-
iment. All of these instruments had been calibrated.
Subjective measurements. A questionnaire was used
that included questions regarding perceived air quality,
thermal comfort, SBS symptom intensity, emotion, and
willingness to exert effort while working. The perceived
air quality and thermal comfort were assessed using
continuous scales describing the acceptability of air
quality and the acceptability of and satisfaction with
the thermal environment (Wargocki et al., 1999). A
7-point continuous scale was used to register thermal
sensation (ASHRAE 2005). SBS symptom intensity
was assessed using visual analogue scales (VAS)—hor-
izontal lines without graduation with two vertical dash
lines marking the extreme points of the scale, each with
end labels (Wargocki et al., 1999). Willingness to exert
effort while working was assessed using a VAS that
increased from 0 (low motivation) to 100 (high
motivation). Different versions of the Tsai-Partington
test were used to assess changes in the level of arousal
(Witterseh et al., 2004). All of these scales were
presented to subjects on a PC (Toftum and Wyon,
The profile of mood states. Emotion was investigated
with the profile of mood states short form (POMS-SF)
presented on a PC (McNair et al., 1992).The mood
states included tension, depression, anger, vigor,
fatigue, and confusion, scoring on a 5-point Likert-
type scale ranging from 0 (not at all) to 4 (extremely).
The total mood disturbance (TMD) was computed by
Lan et al.
adding the scores for tension, depression, anxiety,
fatigue, and confusion, with vigor scores subtracted;
the higher the TMD score, the more negative the
Subjective symptoms of fatigue. To evaluate the pattern
of fatigue, subjects completed a questionnaire pre-
sented on a PC (Tanabe and Nishihara, 2004). The
questionnaire consists of three groups of questions.
Group I consists of 10 questions describing Ôdrowsiness
and dullnessÕ. Group II consists of 10 questions
describing Ôdifficulty in concentrationÕ. Group III
consists of 10 questions describing Ôlack of physical
integrationÕ. The questions are presented to subjects in
random order. The rate of complaints was calculated
for each group based on YoshitakÕs method (Yoshitak,
1973) and used to estimate the pattern of fatigue.
Mental workload. The NASA- Task Load Index
(TLX) for evaluating work load was measured using
a questionnaire presented on a PC (Hart and Wickens,
1990). It consists of six linear scales (similar to VAS)
describing Ômental demandÕ,Ôphysical demandÕ,Ôtem-
poral demandÕ,ÔperformanceÕ,ÔeffortÕ, and Ôfrustration
levelÕ. The endpoints of the scales are marked low and
high, and only in the case of performance are they
marked good and poor. They are normally coded 0–
100. The responses on each scale were analyzed
separately. An overall estimate of mental workload,
Raw TLX (RTLX), was also calculated by averaging
the scores describing each of the six components
(Miyake and Kumashiro, 1993).
Physiological measurements
Finger skin temperature. Finger skin temperature was
measured with the Fluke 65 Hand-held Infrared
Thermometer (Fluke Europe B.V, The Netherlands,
accuracy: ±0.5°C). The infrared thermometer was in a
position perpendicular to the left hand with fingers
held together, at a distance of about 3–5 cm. As the
optical resolution for this device is 8:1, a distance of
3 cm corresponds to a sampling area with a diameter of
<0.5 cm. Typical skin temperature for the human
body in thermal equilibrium is about 32–33°C.
Heart rate and respiration ventilation. Heart rate and
the rrintervals (the time interval between two
heartbeats in milliseconds) were measured with a
commercial heart rate monitor (Sunnto Inc., Vantaa,
Finland), which consisted of a chest strap with
electrodes. Via the docking station, the heart rate and
the rrintervals data were transferred to a computer
for further analysis. The rrintervals were used by the
built-in analysis software (Firstbeat technologies Ltd.,
¨, Finland) to calculate the respiratory venti-
lation rate (the amount of air breathed per min).
Typical healthy resting heart rate in adults is 60–
80 bpm. Changes in heart rate (driving the blood
supply and related to metabolic heat production) can
be used to estimate metabolic rate (ISO 8996, 1989).
End-tidal partial CO
) and arterial blood oxygen
saturation (SPO
). The concentration of end-tidal
partial CO
), the partial pressure of carbon
dioxide at the end of an expiration, was monitored with
a non-invasive capnographic monitor (LifeSense LS1;
MedAir AB Inc., Hudiksvall, Sweden). Measuring
can be used to approximate arterial CO
invasively (Wientjes et al., 1998).The normal values of
are 35–45 mmHg.
Arterial blood oxygen saturation (SPO
) was mea-
sured with a monitor for pulse oximetry (LifeSense
LS1; Medair AB Inc). A special finger probe (REF 210;
Medair AB) was attached to the subjectÕs finger.
Biomarkers in saliva (alpha-amylase and cortisol). A
non-stimulated passive drool salivary sampling proce-
dure was applied to measure the salivary level of alpha-
amylase and cortisol in the saliva of subjects. Subjects
were asked not to eat or drink for half an hour prior to
the collection of saliva. They were also instructed to
swallow or clear their mouth of any excessive saliva
prior to collection of saliva samples. Subjects then
started to accumulate and expel saliva into a labeled
sampling tube for about 5 min to provide about 4 ml
of saliva. The samples were then centrifuged for 15 min
at 1305 gand placed for 1 h in a freezer at )20°C.
After this time, the samples were centrifuged for
15 min again at 1305 g. They were then frozen and
stored in a freezer at )20°C before being sent for
analysis. The analysis was performed by an external
specialized laboratory using a kinetic colorimetric
method to determine salivary alpha-amylase and
cortisol concentration in the saliva. The accuracy of
alpha-amylase analysis was as follows: intra-assay
(within run variation) <1.5%; inter-assay (between
run variation) <1.5%; recovery: 100 ± 6.3%. The
accuracy of cortisol analysis was as follows: intra-assay
(within run variation) <7%; inter-assay (between run
variation) <9.3%; recovery: 100 ± 6.3%. In US
Army research, concentrations of salivary amylase
>600 units/ml are considered Ôhigh stressÕand 400–
600 units/ml are considered Ômoderate stressÕ(Morri-
son et al., 2003). The secretion of cortisol usually
increases profoundly in the morning, after awakening,
and then reduces slowly during the afternoon and
evening hours (Clow, 2004). The diurnal rhythm of
alpha-amylase is characterized by a strong decrease
after waking up and gradually increasing levels with
peaks in the late afternoon (Rohleder and Nater, 2009).
Tear film quality. Samples of tear film mucus were
taken as an indicator of subjectsÕability to maintain
The effects of thermal discomfort on health and human performance
tear film quality. The subjects collected tear film mucus
from the inside of the lower eyelid with a glass rod
(while looking into a mirror) and deposited it onto a
microscope slide. After drying the slide for 1 h, it was
photographed under the microscope and the images
obtained were used to classify the samples. The samples
were classified independently, by three evaluators using
four categories according to the closeness and branch-
ing frequency of the ferning patterns (Rolando, 1984).
Type I with uniform structures without spaces among
the ferns indicates healthy, well-moistured eyes. Type
II having abundant ferning but with small spaces
appearing between the ferns indicates good tear film
mucus. Type III with large spaces between the ferns
and with poorer branching indicates that the mucus
may not be able to perform its function. Type IV with
no visible ferning structure indicates an altered state of
mucus that is typical for people with dry eye syndrome.
Blink rate. During the experiment, webcameras placed
on every workstation recorded continuous video
images of the face of each subject. These recordings
were used to estimate the blink rate. The number of
blinks that occurred during a 5-min period was
counted after 80 min of exposure. This period was
used as the subjects were looking at the computer
screen with little head movement so that an eye blink
could be clearly observed. Most people blink about 15
times a min.
Measurement of performance. During the exposure in
the office, the subjects performed neurobehavioral tests
and tasks typical of office work. The latter included
text typing and addition (Wargocki et al., 1999); the
tasks were presented on a PC (Toftum and Wyon,
2005). The former was a neurobehavioral test battery
(Lan and Lian, 2009b; Lan et al., 2009a) including
seven computerized tests presented to subjects in the
following order: Mental reorientation (a spatial orien-
tation test), Grammatical Reasoning (a logic reasoning
task), Digit Span Memory (a traditional test of verbal
working memory and attention), Visual Learning
Memory (a picture memory task measuring spatial
working memory), Number Calculation (a mental
arithmetical test in which the subject has to add,
subtract, or multiply numbers), Stroop (a test of
attentional vitality and flexibility owing to perceptual/
linguistic interference), and Choice Reaction Time (a
sustained attention task measuring response speed and
accuracy to visual signals).
The tasks were presented on a PC and were self-
paced; the reaction/processing time was recorded by
the computer clock. All tasks were presented to
subjects without feedback, i.e. they did not receive
any information on their performance and performed
each task until it was completed. Speed (response time)
and accuracy (% of corrects) were used as measures of
performance of tasks completed without feedback. In
this case, a performance index (PI) was computed
separately for each task to describe the mean process-
ing/reaction time divided by the accuracy of responses
(e.g. correct characters typed per min or correct units
added per min). For the two memory tests, Digit span
and Visual learning, performance was in terms of
accuracy, i.e. digit span (the maximum number of
digits the subject could correctly learn and recall) and
memory capacity (percentage of correctly recalled
nonsense designs), respectively (Lan et al., 2009a).
Text typing,addition,Stroop, and Number Calculation
were also presented to subjects with feedback about
their performance, i.e. they could not continue until
they corrected the error. Speed (response time includ-
ing the time spent for error correction) was used as a
measure of performance of tasks completed with
feedback. In this case, the PI was calculated as the
reciprocal of processing/reaction time. The tasks were
always performed in the same order independently of
the condition. Four sets of tasks with similar level of
difficulty were prepared and randomly assigned to
subjects in a design that was balanced for order of
Experimental procedure
The subjects were divided into two groups of six
persons (three men and three women). Each group was
exposed to each temperature twice in a repeated
measures design balanced for order of presentation,
i.e., one group was exposed in 22-30-30-22 order and
the other group in the following order: 30-22-22-30.
The exposures took place in two successive weeks of
March 2009 from Monday to Thursday in the after-
noon from 13:00 to 17:30; the actual exposure in the
office lasted 4.5 h. Each group was exposed twice a
week (Group 1 on Mondays and Wednesdays and
Group 2 on Tuesdays and Thursdays) and on the same
weekdays in the subsequent weeks. Although only 12
subjects participated, the repeated measures design
with repetition ensured that the statistical power of the
study was similar to the statistical power of the
previous comparable studies on the effects of IEQ on
performance in which 30 subjects were normally
recruited, and the repeated measures design was used
without repetition (e.g. Wargocki et al., 2000); the
statistical power for both designs is 0.96 with the
sphericity assumption and assuming equal medium
effect size (ES = 0.25) (Faul et al., 2007).
Figure 1 shows the schedule for each experimental
session. Prior to entering the office, the subjects
assembled in a waiting room (in which the temperature
was about 22°C) where they stayed for approximately
10 min. During this time, they rated emotion, fatigue,
and willingness to exert effort. They also put on the
chest belt for continuous measurement of heart rate
Lan et al.
and respiration parameters. The subjects then entered
the office, approached their workstations and evaluated
their thermal comfort and the perceived air quality.
They performed a multiplication task for 15 min; this
period was used to allow subjects to adapt to condi-
tions. The performance of the multiplication task was
not analyzed. After the multiplication task, subjects
assessed the air quality and their thermal comfort, and
indicated their SBS symptoms followed by sampling of
saliva and the measurements of finger temperature.
Subjects then performed the Tsai-Partington test fol-
lowed by a period of 40 min during which they
performed text typing,addition,text typing with feed-
back, and addition with feedback; each task took
10 min to complete. Upon completing the tasks,
subjects evaluated work load using the NASA-TLX,
which was followed by a 40-min period during which
they performed neurobehavioral tests with and without
feedback; the test without feedback took 30 min to
complete, while Stroop and Number Calculation took in
all about 10 min. They then evaluated the air quality
and their thermal comfort and indicated their SBS
symptoms, emotion, fatigue, and motivation. This was
followed by a 10-min break during which the subjects
could leave the office if necessary, but were asked
to stay inside the office. After the break, the above
experimental plan was repeated (Figure 1). Toward the
end of each exposure, after the subjects had completed
the neurobehavioral tests with feedback, finger tem-
perature, ETCO
, and SPO
were measured, and
samples of tear film mucous and saliva were taken.
This was followed by another evaluation of air quality
and thermal comfort, SBS symptoms, emotion, fatigue,
and motivation. The subjects then remained in the
office for 15–20 min without performing any tests
(resting). After this period, the ETCO
and SPO
measured again.
All protocols were approved by the ethics review
board and conformed with the guidelines contained
in the Declaration of Helsinki. Verbal and written
informed consent was obtained from subjects prior to
their participation in the experiment.
Statistical analysis
The SPSS 13.0 (SPSS Inc., Chicago, IL, USA) program
was used to perform the statistical analysis. The
measured data were subject to analysis of variance
(ANOVA) in a repeated measures design; Huynh–
Feldt statistics was used to adjust the violation of
sphericity. The significance level was set to be 0.05
(P< 0.05). In some cases, a paired t-test was used. It
should be noted that it is effect size (ES) instead of P-
value that measures the difference between the true
value and the value specified by the null hypothesis and
hence indicates whether the difference is of practical
importance (Lan and Lian, 2010). In this article, ES
was calculated and reported. Effect sizes with values of
0.1, 0.25, and 0.4 for the ANOVA indicate small,
moderate, and large changes, while for the t-test, the
corresponding effect sizes are 0.2, 0.5, and 0.8. (Cohen,
Average values of the physical parameters describing
the indoor environment in the office are shown in
Table 1; the operative temperature did not deviate
from the intended levels. The CO
levels measured at
each workstation with the Vaisala CO
increased with the air temperature. There were very
small variations between the conditions within each
experimental condition; thus, the conditions for differ-
ent groups of subjects and during each repetition can
be considered as similar.
The thermal sensation reported by the subjects was
neutral (mean thermal sensation votes = 0.01) at 22°C
and warm (mean thermal sensation votes = 2) at 30°C
Fig. 1 Experimental procedure
The effects of thermal discomfort on health and human performance
(Table 2). Subjects were significantly more satisfied
with the thermal environment at 22 than at 30°C
(P< 0.001, ES = 3.20). Subjects reported that the air
quality was significantly less acceptable at 30°C com-
pared with 22°C (Table 2). Although the subjects
entered the office at 30°C from the 22°C waiting area,
their assessments of the perceived air quality and
thermal sensation made during the course of exposure
at this condition did not differ much from the
assessments made immediately upon entering the office
(Table 2). Upon reentering the office after exposure,
subjects reported that the air quality was less accept-
able at 30°C than at 22°C; the estimated % dissatisfied
with air quality (Gunnarsen and Fanger, 1992) corre-
sponded to 65% and 2% respectively. Measured finger
temperatures significantly correlated with thermal
sensation votes; they increased with higher thermal
sensation vote, as expected (Figure 2).
The general rate of complaints of fatigue was very
low and was similar in both conditions prior to
exposure (Table 3). During and after exposure, the
subjects displayed a pattern of fatigue that is typical for
mental work; the rate of complaints was higher at 30°C
than at 22°C.
Subjects experienced more negative moods and lower
vigor when exposed at 30
˚C compared with 22°C; there
were no differences in emotions prior to exposure
(Table 4).
The SBS ratings were subject to paired t-tests as
there was missing data for one group at both 22 and
30°C. Except for the intensity of dry nose, dry skin, dry
eyes, and headache, which did not change significantly,
the intensity of all other SBS symptoms was higher at
30°C compared with 22°C, especially halfway through
the exposure (after 112–122 min of exposure) (Ta-
ble 5). The subjects indicated that they were signifi-
cantly less able to perform their work in the 30°C
condition compared with the 22°C condition. Subjects
indicated that they felt more fatigue at 30°C after they
completed the tasks (Table 5), which is consistent with
the subjectively rated complaints of fatigue (Table 3
and 4).
Figure 3 shows that the subjects were less willing to
exert effort while working at 30°C compared with
22°C, both after exposure for 112 min (P< 0.01;
ES = 1.45) and for 240 min (P< 0.05; ES = 0.65);
there were no differences in motivation prior to
Table 6 summarizes the effects of conditions on
mental work load and the six-component scales of the
NASA-TLX. Subjects indicated that the mental work-
load was significantly higher at 30°C. Subjects indi-
cated also that the mental and physical demand and the
feeling of disappointment increased at 30°C compared
to 22°C.
The performance of simulated office work and of the
neurobehavioral tests is summarized in Table 7. Except
for the performance of the text typing task, the
Table 1 Physical measurements (mean € s.d.) describing the office environment under the two conditions
Condition (intended
temperature) Air temperature (°C)
temperature (°C)
humidity (%) Light (lux) CO
(ppm) Noise (dBA)
22°C 23.3 € 0.8 22.6 € 0.6 21 € 7 241 € 60 801 € 70 56 € 5
30°C 31.1 € 1.3 29.4 € 1.2 22 € 3 225 € 64 1047 € 145 55 € 4
Table 2 Subjective assessments on thermal sensation and perceived air quality at two
thermal conditions; the results shown are averages and standard deviations (in brackets)
Exposure time in
the room (min) 0–2 17–20 112–122 135–137 240–250
Thermal sensation [hot(3); warm(2); slightly warm(1); neutral(0); slightly cool()1);
cool()2); cold()3)]
22°C 0.42 0.30 )0.20 )0.39 )0.07
30°C 2.17 2.08 2.28 2.16 2.00
P<0.001** <0.001** <0.001** <0.001** <0.001**
Perceived air quality (clearly acceptable(1); clearly unacceptable()1))
22°C 0.70 0.81 0.74 0.71 0.79
30°C)0.11 )0.06 )0.22 )0.15 )0.08
P<0.001** <0.001** <0.001** <0.001** <0.001**
% dissatisfied with air quality (Gunnarsen and Fanger, 1992)
22°C 21221
30°C 6054646557
Significant differences (P< 0.05) between conditions observed.
Fig. 2 Thermal sensation votes as a function of finger tem-
Lan et al.
Table 5 Average and standard deviation (in brackets) intensity of sick building syndrome symptoms under different conditions in the office; coding of the scales is indicated in the table;
effect sizes (ES) with values of 0.2, 0.5, and 0.8 indicate small, moderate, and large changes, respectively
Time during exposure in the office (min)
17)20 112–122 240–250
Dry nose (100)
Runny nose (0)
59.8 (25.5) 52.3 (30.3) 0.41 0.18 50.5 (25.3) 56.7 (26.0) 0.13 0.67 54.8 (25.4) 54.5 (30.9) 0.05 0.87
Dry throat (100)
Not dry throat (0)
34.1 (30.2) 53.0 (37.8) 0.43 0.17 36.5 (28.6) 53.9 (30.9) 1.00 0.01
42.4 (34.5) 57.5 (30.1) 0.85 0.02
Dry mouth (100)
Not dry mouth (0)
46.5 (33.9) 57.4 (37.2) 0.56 0.08
39.5 (28.2) 57.5 (36.1) 1.05 <0.01
50.9 (32.2) 55.9 (35.8) 0.50 0.11
Dry skin (100)
Not dry skin (0)
31.1 (29.6) 35.0 (29.7) 0.24 0.43 27.0 (24.9) 33.2 (29.3) 0.39 0.20 29.8 (27.5) 31.5 (25.1) 0.20 0.51
Dry eyes (100)
Not dry eyes (0)
29.2 (26.5) 34.7 (29.3) 0.46 0.14 34.8 (28.7) 50.3 (33.4) 0.40 0.19 45.0 (30.7) 53.6 (31.8) 0.47 0.13
Severe headache (100)
No headache (0)
9.7 (18.7) 13.9 (18.9) 0.15 0.60 16.9 (22.3) 22.9 (27.0) 0.24 0.42 17.3 (24.9) 27.0 (30.1) 0.46 0.14
Difficult to concentrate (100)
Easy to concentrate (0)
20.2 (19.2) 44.7 (27.6) 0.75 0.03
36.3 (19.6) 69.9 (23.3) 1.80 <0.01
41.0 (24.8) 65.1 (20.8) 0.95 0.01
Feeling good (100)
Feeling bad (0)
82.4 (13.5) 56.1 (28.0) 0.99 0.01
67.9 (22.2) 36.1 (24.1) 1.42 <0.01
66.5 (19.5) 39.7 (20.6) 1.12 <0.01
Sleepy (100)
Rested (0)
33.7 (23.0) 39.0 (27.0) 0.18 0.54 51.2 (23.9) 74.1 (24.9) 1.13 <0.01
60.8 (20.9) 69.6 (20.0) 0.47 0.14
Hard to think (100)
Easy to think (0)
19.7 (14.3) 45.4 (27.4) 1.45 <0.01
37.6 (23.6) 69.8 (25.4) 1.39 <0.01
44.4 (24.2) 65.1 (19.1) 0.79 0.02
In a good mood (100)
Depressed (0)
81.2 (15.9) 66.7 (24.2) 1.24 <0.01
76.7 (16.9) 54.7 (19.5) 1.06 <0.01
71.9 (16.0) 55.0 (19.0) 1.04 <0.01
I am able to work
80.3 (9.3) 65.9 (17.2) 1.52 <0.01
68.1 (16.5) 46.1 (20.3) 1.33 <0.01
64.5 (17.7) 47.1 (20.7) 0.88 0.02
Differences approaching significance (0.05 < P< 0.10).
Significant differences (P< 0.05) between conditions observed.
Table 4 Average and standard deviation (in brackets) of emotion scores under different conditions; higher score of total mood disturbance (TMD) and the negative moods, such as tension,
depression, anger, fatigue, and confusion, indicates higher negative emotion, while the higher score of vigour indicates higher positive emotion; effect sizes (ES) with values of 0.1, 0.25,
and 0.4 indicate small, moderate, and large changes, respectively
Time during exposure (min)
17)20 112–122 240–250
TMD )0.7 (5.8) )1.0 (7.3) 0.08 0.82 9.3 (10.1) 19.2 (15.6) 1.23 <0.01
12.1 (9.9) 19.3 (13.8) 0.93 0.02
Tension 0.9 (0.9) 0.8 (1.7) 0.03 0.93 2.2 (2.0) 3.3 (2.3) 0.62 0.08
2.7 (2.2) 3.3 (2.2) 0.38 0.29
Depression 1.0 (1.5) 1.2 (2.3) 0.10 0.77 2.2 (2.2) 4.1 (3.4) 0.87 0.02
2.3 (2.2) 3.7 (3.00) 0.69 0.07
Anger 1.3 (1.8) 0.9 (1.5) 0.23 0.49 2.7 (2.3) 4.5 (3.5) 1.02 0.01
2.8 (2.2) 4.0 (3.2) 0.71 0.06
Vigor 11.0 (2.9) 10.5 (3.9) 0.18 0.58 5.7 (3.2) 4.2 (3.4) 1.25 <0.01
3.8 (2.6) 3.5 (3.5) 0.42 0.24
Fatigue 3.8 (1.2) 4.0 (1.0) 0.13 0.70 4.92 (3.23) 7.5 (4.05) 1.24 <0.01
5.71 (3.23) 8.49 (4.13) 1.37 <0.01
Confusion 3.3 (1.4) 2.7 (0.9) 0.64 0.07
3.0 (2.3) 3.9 (2.3) 0.47 0.17 2.4 (2.3) 3.3 (2.4) 0.47 0.19
Differences approaching significance (0.05 < P< 0.10).
Significant differences (P< 0.05) between conditions observed.
Table 3 The complaints of subjective symptoms of fatigue in the two thermal conditions
Exposure time in the room (min) Condition
Complaint rates
The order among
three categories Pattern of fatigueGroup I Group II Group III
Before exposure 22°C 5.8 4.2 6.3 III > I > II Fatigue for mental work
30°C 4.8 5.2 3.0 II > I > III N/A
After block 1 (112) 22°C 22.1 18.3 9.2 I > II > III Fatigue for mental work
30°C 43.5 42.9 17.1 I > II > III Fatigue for mental work
After block 2 (240) 22°C 25.6 21.1 12.2 I > II > III Fatigue for mental work
30°C 47.4 41.7 24.3 I > II > III Fatigue for mental work
The effects of thermal discomfort on health and human performance
performance of the tasks decreased at 30°C compared
with 22°C, although for some tasks, the decrease was
not significant. The effect size was in many cases at
least moderate, indicating that the observed effects
were of practical importance. The performance of tasks
presented with feedback was affected to a higher degree
than that of the tasks presented without feedback,
especially in the condition with a temperature of 30°C.
In the case of text typing without feedback, the subjects
input more characters at 30°C but at the same time,
they also made more errors; the effect size was small in
this case, and the effect was insignificant suggesting
that the observed effects occurred by chance and had
no practical importance. When text typing with feed-
back was presented, the subjects performed less well at
30°C compared with 22°C. The effect was not statis-
tically significant although the effect size was in this
case moderate.
No significant effects of thermal environment or
exposure time were observed on the performance of the
Tsai-Partington tests.
Based on the continuous measurements of heart rate
and the r–r intervals, the average heart rate and
respiratory ventilation rate for five different periods
during exposure was calculated: when subjects were
performing the tasks typical of office work (during
exposure in the office from 30 to 70 min); when
subjects were performing the neurobehavioral tests
(from 72 to 112 min); when they were performing tasks
typical of office work and the neurobehavioral tests
separately again (from 137 to 177 minand 179 to
219 min, respectively); and finally when they rested
while remaining in the room (from 250 to 265 min).
Heart rate was lower at 22°C compared with 30°C; it
also decreased during the course of each exposure
independently of the condition in the office (Table 8).
The respiratory ventilation rate was lower at 22°C than
at 30°C; it decreased during the course of exposure to
22°C but at 30°C, it first decreased and then increased;
the decrement was independent of the condition when
the subjects rested in the office (Table 8).
Table 9 shows that the ETCO
measured immedi-
ately after the subjects finished the tasks, i.e. after
219 min of exposure in the office, was significantly
higher at 30°C(P= 0.02, ES = 0.86). The measure-
ments taken after the 15 min of resting showed that the
arterial CO
was still higher at 30°C, although the
difference had decreased somewhat (P= 0.08,
ES = 0.63), Table 9. The arterial CO
did not change
significantly during exposure in either of the two
examined temperatures. The SPO
taken after 219 min of exposure in the office indicate
that arterial oxygen saturation was significantly lower
at 30°C compared with 22°C(P< 0.01, ES = 1.524),
Table 9. Arterial oxygen saturation was still signifi-
cantly lower at 30°C after the 15-min rest (P< 0.05,
ES = 0.823). Arterial oxygen saturation did not
change significantly between the two periods in which
the measurements were taken at 22°C, but it increased
significantly after 15 min of resting at 30°C(P< 0.01,
ES = 1.094).
Table 6 Mental workload (Raw TLX) and the six-component scales of NASA-TLX during exposure in the office at different conditions; the results shown are averages and standard
deviations (in brackets); effect sizes (ES) with values of 0.1, 0.25, and 0.4 indicate small, moderate, and large changes, respectively
Time during exposure in the office (min) Temperature
70–72 177–179
Mental demand (A lot-100, A little-0) 30.9 (15.8) 40.0 (21.9) 0.96 0.01
31.0 (14.9) 36.5 (19.1) 0.68 0.06
Physical demand (A lot-100, A little-0) 14.0 (14.1) 26.3 (16.5) 0.82 0.03
13.6 (8.4) 23.6 (18.7) 0.72 0.04
Time stress (A lot-100, A little-0) 37.3 (21.5) 32.3 (20.5) 0.41 0.23 34.4 (23.0) 35.4 (20.5) 0.06 0.85
Performance (Poor-100, Well-0) 27.2 (18.4) 36.5 (24.9) 0.73 0.04
22.2 (16.2) 33.1 (23.1) 0.79 0.03
Effort (A lot-100, A little-0) 25.7 (17.8) 24.4 (18.3) 0.12 0.71 26.5 (16.5) 25.5 (16.6) 0.11 0.74
Disappoint (A lot-100, A little-0) 25.1 (18.0) 29.1 (22.2) 0.35 0.30 18.5 (14.0) 29.2 (20.0) 0.92 0.02
Mental workload (Raw TLX) 26.7 (9.5) 31.4 (14.2) 0.68 0.05
24.3 (9.0) 30.6 (14.2) 0.77 0.04
Differences approaching significance (0.05 < P< 0.10).
Significant differences (P< 0.05) between conditions observed.
Fig. 3 Subjectively assessed willingness to exert effort while
working as a function of time during exposure in the office
Lan et al.
The concentration of alpha-amylase did not change
with the temperature; it also did not change between
the two measurements taken at two points in time, in
either temperature condition (Table 10). No significant
change in salivary cortisol was observed between the
two thermal conditions, but salivary cortisol decreased
significantly during the course of exposure in both
conditions (P< 0.01), Table 10. The concentration of
salivary alpha-amylase correlated positively with cor-
tisol level but with a relatively low correlation coeffi-
cient (SpearmanÕsr= 0.29). No significant correlation
was found between the salivary alpha-amylase level or
the cortisol level and the performance of the Tsai-
Partington tests.
Figure 4 shows the results of the tear film mucous
quality analysis. There was a significant effect of
temperature condition on tear film quality (P< 0.05,
ES = 0.33). Samples assessed as Type I and Type II
were observed to be more frequent when the temper-
ature was 22°C. The increased percentage of Type III
and Type IV tear film mucus samples at 30°C indicates
that tear film mucous quality was reduced at the raised
No significant difference in blink rate was found
between the 5-min periods after 80 min of exposure
during which blink rates were counted (17.89 ± 8.44
blinks/min at 22°C vs. 17.53 ± 8.41 blinks/min at
The air temperatures of 22 and 30°C were selected to
provide two sufficiently different thermal comfort
environments. Using measured parameters in the
office, assuming an activity level of 1.2 met, it was
predicted (Fanger, 1970) that subjects should have
been thermally neutral at 22°C and warm at 30°Cin
clothing with an insulation value of 0.9 clo. The
Table 7 Performance of tasks typical of office work and of the neurobehavioral tasks; a
negative relative change in the performance (4) indicates that performance decreased at
30°C compared with 22°C; effect sizes (ES) with values of 0.1, 0.25, and 0.4 indicate
small, moderate, and large changes, respectively
Performance index
Metrics Value 4ES P
Text typing 22
Char/min 143.6
0.3% 0.11 0.75
Text typing with feedback 22
Char/min 133.1
)10.6% 0.40 0.23
Addition 22
Units/min 4.67
)11.7% 1.15 0.01
Addition with feedback 22
Units/min 3.02
)8.2% 0.83 0.03
Redirection 22
Units/sec 0.932
)3.0% 0.58 0.10
Digit span 22
Span 7.69
)2.8% 0.27 0.44
Grammatical reasoning 22
Units/sec 0.165
)25.0% 0.71 0.06
Visual learning 22
)1.4% 0.22 0.51
Stroop 22
Units/sec 0.513
)9.5% 0.59 0.09
Stroop with feedback 22
Units/sec 0.517
)11.2% 1.01 0.01
Calculation 22
Units/sec 0.236
)4.6% 0.62 0.08
Calculation with feedback 22
Units/sec 0.225
)9.7% 0.93 0.01
Visual RT 22
Units/sec 1.672
)6.8% 0.86 0.02
Differences approaching significance (0.05 < P< 0.10).
Significant differences (P< 0.05) between conditions observed.
Table 8 Heart rate and respiratory ventilation rate as a function of time during exposure
in the office; the results shown are averages and standard deviations (in brackets)
Exposure time in
the room (min) 30)70 72–112 137–177 179–219 250–265
Heart rate (bpm)
22°C 73.5 (9.3) 71.2 (8.9) 70.7 (8.6) 69.6 (7.9) 71.0 (8.3)
30°C 78.8 (9.5) 76.9 (8.7) 75.9 (8.8) 75.6 (9.2) 72.7 (8.5)
P<0.01** <0.01** <0.01** <0.01** 0.05**
Respiratory ventilation (l/min)
22°C 6.5 (2.1) 6.2 (2.4) 5.9 (2.1) 5.8 (2.0) 5.6 (1.7)
30°C 7.3 (2.1) 6.7 (2.1) 6.7 (2.0) 6.8 (2.1) 5.7 (1.8)
P0.06* 0.21 <0.01** <0.01** 0.69
Differences approaching significance (0.05 < P< 0.10).
Significant differences (P< 0.05) between conditions observed.
Table 10 Concentration of salivary alpha-amylase and cortisol during exposure in the
office at different conditions; the results shown are averages and standard deviations (in
Time during exposure in the office (min) 20–30 219–240
Alpha-amylase (IU/ml)
22°C 98.4 (42.5) 104.9 (54.9)
30°C 97.2 (44.0) 99.6 (58.1)
P0.81 0.73
Cortisol (ng/ml)
22°C 5.25 (2.41) 3.02 (1.00)
30°C 4.72 (1.48) 3.29 (0.90)
P0.38 0.18
Table 9 End-tidal partial CO
and arterial oxygen saturation measured at the two
thermal conditions; the results shown are averages and standard deviations (in brackets)
Exposure time in the room (min) 219–240 265–270
End-tidal partial CO2 (mmHg)
22°C 38.2 (3.1) 38.3 (3.1)
30°C 40.2 (3.1) 39.8 (2.6)
P0.02** 0.08*
Arterial oxygen saturation (%)
22°C 98.0 (0.8) 97.8 (1.2)
30°C 96.6 (0.7) 96.9 (0.7)
P<0.01** 0.03**
Differences approaching significance (0.05 < P< 0.10).
Significant differences (P< 0.05) between conditions observed.
The effects of thermal discomfort on health and human performance
corresponding predicted percentage dissatisfied be-
cause thermal discomfort was 5% and 69%, respec-
tively (Fanger, 1970). The thermal sensation and
thermal comfort responses indicate that the two
thermal conditions were as predicted. The correlation
between thermal sensation and finger temperature
confirms previous findings (Wargocki et al., 2009)
and suggests that skin temperature can be used as a
simple indicator of thermal discomfort in this thermal
The intensity of SBS symptoms, fatigue, and negative
mood disturbance increased to a large extent, and the
perceived indoor air quality and motivation to exert
effort while working decreased, when subjects were
exposed at 30°C and felt thermally warm compared
with 22°C, at which temperature they felt thermally
neutral. These results are consistent with those of
previous studies (e.g. Fang et al., 1998; Haneda et al.,
2009; Lan and Lian, 2009b; Lan et al., 2009a; Mendell
and Mirer, 2009; Tanabe and Nishihara, 2004).
The intensity of SBS symptoms, the negative mood
disturbance, and the perceived air quality improved
slightly toward the end of the exposure. Similar results
were also obtained in previous studies in which subjects
were exposed for a period of up to 4.5 h to different air
quality and temperature conditions (Fang et al., 2004;
Wargocki et al., 1999, 2000). These results could be
either because of psychological habituation or because
of physiological adaptation. Subjects were aware that
they would be leaving the exposure shortly, and this
positive emotion might have caused their evaluations
to be less critical toward the end of each exposure. This
interpretation is supported by the reporting of less
negative moods (e.g. less anger, see Table 4). It should
also be noticed that subjects did not perform any tasks
toward the end of an exposure. Thus, their metabolic
rate would have been lower, causing less thermal strain
that could also result in less intense symptoms.
The negative effects of moderately elevated temper-
atures causing thermal discomfort on the performance
of tasks typical of office work or neurobehavioral tests
examining different skills confirm the results of some of
previous studies (e.g. Niemela et al., 2002; Tham, 2004;
Wargocki and Wyon, 2007; Witterseh et al., 2004).
Some recent laboratory studies did not observe such
effects (Haneda et al., 2009; Lan and Lian, 2009b; Lan
et al., 2009a; Tanabe and Nishihara, 2004). One
possible explanation for this discrepancy may be
motivation. In the studies of Lan et al. (2009a) and
Lan and Lian (2009b) Haneda et al. (2009), the
subjects were told during recruitment that a bonus
would be paid to them depending on their perfor-
mance. This could affect their motivation and the
subjects might have exerted more effort to maintain
performance independently of the thermal conditions,
at the expense of experiencing more symptoms. In the
present study, no bonus was offered to the subjects and
there is no indication that the subjects exerted more
effort in different exposure conditions (Table 6). In
fact, Figure 3 shows that the subjects were less moti-
vated to exert effort at 30°C.
According to Heerwagen (1995), the relationship
between buildings and worker performance is related
to both motivation and ability; an individual must first
feel like working and then must be capable of working.
When highly motivated by a bonus, people are able to
resist the negative effects of thermal discomfort by
exerting more effort. This is probably due to the
existence of a Ôcognitive reserveÕ(extraneural re-
sources) that allows people to maintain their perfor-
mance during a short exposure even when indoor
conditions are unfavorable (Hocking et al., 2001).
Performance can deteriorate when these resources are
insufficient to deal with both the task demands and
thermal stress. People are then unable to maintain
their performance. Exerting extra effort for a short
period is probably without long-lasting health conse-
quences; this is an example of human flexibility being
able to deal with short-term demands (Johnson and
Anderson, 1990). However, prolonged, repetitive, and
continuous effort can have negative consequences for
health. For example, it has been suggested that
repetitive activation of the cardiovascular defense
response may lead to hypertension (Johnson and
Anderson, 1990). Taking this view, poor indoor
environmental conditions, e.g. elevated temperatures
and/or thermal discomfort, should be avoided in
offices even though increased effort can for a short
period counteract the negative effects. This was prob-
ably the case in some of the previous studies where
no effect of thermal discomfort on performance
was observed (e.g. Haneda et al., 2009; Tanabe and
Fig. 4 Tear film quality as a function of the condition in the
office; Type I and Type II indicate good efficiency of tear mucus
and higher tear film quality while Type III and Type IV indicate
problems with mucus and lower tear film quality
Lan et al.
Nishihara, 2004). High motivation to overcome any
negative effect of the indoor environment may be
difficult to maintain in schools, especially in elemen-
tary schools. This is probably the reason why the
performance of school work by children is to a much
larger extent affected by poor indoor environmental
quality compared to the performance of adults in
offices (Wargocki and Wyon, 2007).
Many countries now mandate that thermostats
should be set higher during warm weather than would
be required for thermal neutrality to conserve the
energy used for cooling buildings, and many ÔgreenÕor
low-energy buildings are to some extent thermally
passive, allowing indoor temperatures to rise when it is
warm outdoors. When indoor temperatures increase
above what is necessary for thermal neutrality,
subjective thermal discomfort is usually reported.
People can habituate to endure or even prefer feeling
warm rather than neutral at elevated temperatures.
Assuming that the distraction of thermal discomfort is
the main mechanism, no negative effects on mental
performance would be expected to occur at elevated
temperatures when people have been able to adapt to
new conditions. However, the present study shows that
there are other physiological mechanisms by which the
thermal discomfort caused by moderately elevated
temperatures can affect task performance and SBS
symptoms, so habituation to a feeling of warmth may
not protect against the negative effects; these plausible
mechanisms are discussed in subsequent paragraphs. It
should be mentioned that even if habituation was the
only mechanism, thermal conditions providing opti-
mum comfort may not give rise to maximum perfor-
mance. This has already been demonstrated by Pepler
and Warner (1968): normally clothed (0.5 clo) young
Americans performed mental work at a number of
different temperatures (from 21 to 33°C). They were
most thermally comfortable at 27°C, the temperature
at which they exerted least effort and performed least
work, while they performed most work at 21°C,
although most of them felt uncomfortably cold under
this condition. The percentage error rate was constant
Higher ETCO
(indicating higher arterial CO
and increased heart rate and respiratory ventilation
rate were observed at 30°C when subjects felt warm.
This suggests that subjectsÕmetabolic rate increased at
this condition. This result is as expected—when the
body is hot, the rate of heat production will increase
because of the increased rate of chemical reaction in
the cells of the body (ISO 8996, 1989). The metabo-
lization of fats and carbohydrates leads at the same
time to the formation of a large amount of CO
. When
the lungs cannot remove all of the CO
produced by
the metabolic processes, respiratory acidosis may
occur (Seifter, 2007). Studies have shown that respira-
tory acidosis or an increase in CO
concentration in
the blood may induce physiological effects that give
rise to SBS symptoms such as easy fatigue, sleepiness,
and headache (Paulev, 2000; Resta et al., 2000; Seifter,
2007). In the present study, the intensity of SBS
symptoms increased when the subjects felt warm at
˚C and at the same time ETCO
increased. As a
natural defense mechanism, the respiration rate (the
amount of air breathed per min) increased (see
Table 8) to remove excessive CO
but the increased
respiration was evidently insufficient to remove all
excessive CO
; the discrepancy between required and
actual respiratory ventilation rate led to an increased
accumulation of CO
in the blood.
The CO
concentration in room air in the 30°C
condition was higher than it was at 22°C (see Table 1).
This increase could be the result of increased metabolism
at 30°C and increased production of metabolic CO
a result, the concentration of bioeffluents in the room
increased in this condition, leading to a reduction in the
air quality. Perceived air quality at 30°C may also have
occurred because elevated temperatures reduce the
capacity of inhaled air to cool mucous membranes in
the nose (Fang et al., 1998; Toftum et al., 1998).
Blood oxygen saturation (SPO
) decreased signifi-
cantly when the subjects felt warm. This change of
could affect the intensity of SBS symptoms and
task performance. For example, it has been found that
a supply of oxygen reduced the fatigue experienced by
drivers (Sung et al., 2005). In the present study, fatigue
symptoms increased in the warm condition (see
Table 3, 4, and 5) when blood oxygen saturation
decreased. Low oxygen saturation has also been shown
to be associated with decreased cognitive functions
(Andersson et al., 2002; Winder and Borrill, 1998).
Brain activity consumes 20–30% of all energy pro-
duced in the body. In the process of providing cells
with energy, oxygen is consumed (Benton et al., 1996).
It has also been shown using brain imaging techniques
that the brain increases the uptake of oxygen (and
glucose) into active brain areas during cognitively
demanding tasks (Andersson et al., 2002). In the
present study, reduced task performance (see Table 7)
was observed when blood oxygen saturation decreased,
which is consistent with the hypothesis that lower
oxygen saturation can reduce task performance.
Whether the reduced performance occurred because
of decreased cognitive functioning of the brain or
because of fatigue cannot be determined from the
present results. The studies of Nishihara et al. (2008)
showed generally no difference between oxygenated
and deoxygenated hemoglobin in the brain when
subjects were exposed to a temperature range of
25.5–31.5°C. It should be noted that oxygen saturation
increased after 15 min of resting toward the end of the
exposure at 30°C (Table 9). This could occur because
more oxygen was available when subjects were resting
and their metabolic rate decreased (see Table 8).
The effects of thermal discomfort on health and human performance
Tear film quality decreased and the intensity of the
dry eye symptom increased (0.4 < ES < 0.5) when
subjects were thermally warm at 30°C (see Table 5).
Mendell et al. (2002) reported that a 19% decrease in
eye irritation was associated with each 1°C decrease in
air temperature in the range from 26 to 22°C. Dry eyes
can be defined as a disorder of the tear film because of
excessive tear evaporation, which may cause damage to
the ocular surface. An increased rate of water evapo-
ration from the tear mucus at the higher air temper-
ature, perhaps because of an increase in eye surface
temperature, may thus have contributed to the devel-
opment of symptoms related to eye irritation; there
were no differences in relative humidity between the
two conditions (Table 1). A natural body defense
mechanism would be to increase the blink rate to
refresh the tear film more often, but this was not
observed here, possibly because blink rate was evalu-
ated only from a short Ôsnap shotÕof the entire
exposure. The fact that the tests were presented to
subjects on a PC will inevitably have reduced their
blink rate when they were concentrating on the tasks.
No effects of the thermal environment on the level of
salivary alpha-amylase and cortisol or on the perfor-
mance of the Tsai-Partington test were found in the
present study. These results may suggest no effects of
the thermal environment on arousal and are inconsis-
tent with the results obtained by Tham and Willem
(2010) and by Wargocki et al. (2009). The reasons for
the inconsistency between these different studies are
unclear at present. One reason can be the low level of
stress experienced by the subjects in the present study,
as the concentrations of salivary amylase obtained in
all but two samples were below 200 units/ml (Morrison
et al., 2003). A reduction in salivary cortisol during the
course of the exposures was observed, independently of
conditions, suggesting further relief from stress. This
was probably because the subjects worked on self-
paced tasks and were acquainted with the experimental
procedures before starting the experiment.
A limitation of the present experiment is that only
young and healthy college students were recruited and
that this is not a representative subpopulation of office
workers. In future experiments, a different subpopula-
tion of subjects should be selected in terms of their age,
health, and occupation. Validation of the present
results is also required in real workplaces and warmer
climates, at lower room temperatures and in different
seasons of the year.
The perceived air quality was poorer, the intensity of
SBS symptoms, fatigue, mental workload, and negative
mood disturbance all increased, the subjects were less
willing to exert effort while working, and the perfor-
mance of simulated office work and neurobehavioral
tests decreased when they felt warm at 30°C compared
with the 22°C condition in which they felt thermally
neutral. Heart rate, respiratory ventilation rate, and
concentration increased, and SPO
tion decreased, while tear film quality was reduced and
eye discomfort was intensified when the subjects felt
warm at 30°C compared with the 22°C condition when
they felt thermally neutral. No effects of thermal
conditions on the concentration of salivary alpha-
amylase and cortisol, or on the performance of a Tsai-
Partington test, were found. The present results imply
that the negative effects on health and performance
that are observed when people are thermally warm at
raised temperatures are very likely to be caused by
physiological mechanisms.
This work was supported by the International Centre
for Indoor Environment and Energy and the National
Natural Science Foundation of China (No.50878125).
The authors thank Ms Justyna Krzy_
zanowska for her
contribution to this study and Dr. Jørn Toftum for his
assistance in preparing computerized versions of the
tests. The authors also thank the subjects who volun-
teered for this study.
Andersson, J., Berggren, P., Gronkvist, M.,
Magnusson, S. and Svensson, E. (2002)
Oxygen saturation and cognitive perfor-
mance, Psychopharmacology,162, 119–
ASHRAE (2005) ASHRAE Handbook:
Fundamentals, Atlanta, GA, American
Society of Heating, Refrigerating and Air-
Conditioning Engineers, Inc.
´, Z.S., Wargocki, P., Wyon, D.P.
and Fanger, P.O. (2005) Poor indoor air
quality slows down metabolic rate of
office workers. In: Proceedings of
Indoor Air 2005, Beijing, on CD-ROM,
Benton, D., Parker, P.Y. and Donohoe, R.T.
(1996) The supply of glucose to the brain
and cognitive functioning, J. Biosoc. Sci.,
28, 463–479.
Berglund, L., Gonzales, R. and Gagge, A.
(1990) Predicted human performance
decrement from thermal discomfort and
ET*. In: Walkinshaw, D.S. (ed.) Pro-
ceedings of the Fifth International Con-
ference on Indoor Air Quality and Climate,
Toronto, Canada, Vol 1, 215–220.
Clow, A. (2004) Cortisol as a biomarker of
stress, J. Holist. Nurs.,1, 10–14.
Cohen, J. (1988) Statistical Power Analysis
for the Behavioural Sciences, 2nd edn,
New Jersey, Hillsdale.
Fang, L., Clausen, G. and Fanger, P.O.
(1998) Impact of temperature and
humidity on the perception of indoor air
quality, Indoor Air,8, 276–284.
Fang, L., Wyon, D.P., Clausen, G. and
Fanger, P.O. (2004) Impact of indoor air
temperature and humidity in an office on
perceived air quality, SBS symptoms and
Lan et al.
performance, Indoor Air,14(Suppl. 7),
Fanger, P.O. (1970) Thermal Comfort.
Analysis and Applications in Environmen-
tal Engineering, New York, McGraw Hill.
Faul, F., Erdfelder, E., Lang, A.G. and
Buchner, A. (2007) G*Power 3: a flexi-
ble statistical power analysis program
for the social, behavioral, and biomedi-
cal sciences, Behav. Res. Methods,39,
Gunnarsen, L. and Fanger, P.O. (1992)
Adaptation to indoor air pollution,
Energy Build.,18, 43–54.
Haneda, M., Wargocki, P., Delewski, M. and
Tanabe, S. (2009) The effects of thermal
discomfort on task performance, fatigue
and mental work load examined in a
subjective experiment. In: Healthy
Buildings 2009, on CD-ROM.
Hart, S. and Wickens, C.D. (1990) Workload
assessment and prediction. In: Booher,
H.R. (ed.), MANPRINT: An Emerging
Technology. Advanced Concepts for
Integrating People, Machines and
Organisations. Reinhold, NY, Van
Nostrand, 139–183.
Heerwagen, J.H., Heubach, J., Montgomery,
J. and Weimer, W. (1995) Environmental
design, work and well-being, J. Am.
Assoc. Occup. Health Nurs.,43, 458–468.
Hellhammer, D.H., Wu
¨st, S. and Kudielka,
B.M. (2009) Salivary cortisol as a
biomarker in stress research, Psychoneu-
roendocrinology,34, 163–171.
Hocking, C., Silberstein, R.B., Lau, W.M.,
Stough, C. and Roberts, W. (2001) Eval-
uation of cognitive performance in the
heat by functional brain imaging and
psychometric testing, Comp. Biochem.
Physiol.,128 (Part A), 719–734.
ISO 8996 (1989) Ergonomics of Thermal
Environments - Determination of Meta-
bolic Heat Production. Geneva, ISO.
Johnson, A.K. and Anderson, E.A. (1990)
Stress and arousal. In: Cacioppo, J.T. and
Tassinary, L.G. (eds) Principles of
Psychophysiology, Cambridge,
Cambridge University Press, 216–252.
Kosonen, R. and Tan, F. (2004) Assessment
of productivity loss in air-conditioned
buildings using PMV index, Energy
Build.,36, 987–993.
Lan, L. and Lian, Z.W. (2009b) Use of
neurobehavioral tests to evaluate the
effects of indoor environment quality on
Productivity, Build. Environ.,44, 2208–
Lan, L. and Lian, Z.W. (2010) Application
of statistical power analysis-how to
determine the right sample size in human
health, comfort and productivity
research, Build. Environ.,45, 1202–1213.
Lan, L., Lian, Z.W., Pan, L. and Ye, Q.
(2009a) Neurobehavioral approach for
evaluation of office workersÕproductivity:
the effects of room temperature, Build.
Environ.,44, 1578–1588.
McNair, D.M., Lorr, M. and Droppleman,
L.F. (1992) Revised Manual for the Profile
of Mood States, San Diego, CA, Educa-
tional and Industrial Testing Services.
Mendell, M.J. and Mirer, A.G. (2009)
Indoor thermal factors and symptoms in
office workers: findings from the US EPA
BASE study, Indoor Air,19, 291–302.
Mendell, M.J., Fisk, W.J., Petersen, M.R.,
Hines, C.J., Dong, M., Faulkner, D.,
Dedens, J.A., Ruder, A.M., Sullivan, D.
and Boeniger, M.F. (2002) Indoor parti-
cles and symptoms among office workers:
results from a double-blind cross-over
study, Epidemiology,13, 296–304.
Miyake, S. and Kumashiro, M. (1993) Sub-
jective mental workload assessment tech-
nique-an introduction to NASA-TLX and
SWAT and a proposal of simple scoring
methods, Jpn. J. Ergon.,29, 399–408. (in
Morrison, W.E., Haas, E.C., Shaffner, D.H.,
Garrett, E.S. and Fackler, J.C. (2003)
Noise, stress, and annoyance in a pediat-
ric intensive care unit, Crit. Care Med.,
31, 113–119.
Nater, U.M. and Rohleder, N. (2009)
Salivary alpha-amylase as a non-invasive
biomarker for the sympathetic nervous
system: Current state of research,
Psychoneuroendocrinology,34, 486–496.
Niemela, R., Hannula, M., Rautio, S.,
Reijula, K. and Railio, J. (2002) The
effect of air temperature on labour
productivity in call centers-a case study,
Energy Build.,34, 759–764.
Nishihara, N., Tanabe, S., Haneda, M. and
Nakamura, S. (2008) Objective evaluation
method of workersÕphysiological
responses that affect productivity in
moderately hot environment, In: Pro-
ceedings of Indoor Air 2008, on CD-ROM.
Paulev, P.E. (2000) Textbook in medIcal
Physiology and Pathophysiology, Copen-
hagen, Copenhagen Medical Publishers.
Pepler, R.D. and Warner, R.E. (1968)
Temperature and learning: an
experimental study, ASHRAE Trans.,
74, 211–219.
Resta, O., Foschino Barbaro, M.P.,
Bonfitto, P., Talamo, S., Mastrosimone,
V., Stefano, A. and Giliberti, T. (2000)
Hypercapnia in obstructive sleep apnoea
syndrome, Neth. J. Med.,56, 215–222.
Roelofsen, P. (2001) The design of the
workplace as a strategy for productivity
enhancement. In: Moncada Lo Giudice,
G. and Costantino, M. (eds) Clima 2000/
Napoli 2001 World Congress, Napoli,
Roelofsen, P. (2002) The impact of office
environments on employee performance:
the design of the workplace as a strategy
for productivity enhancement, J. Facil.
Manag.,1, 247–264.
Rohleder, N. and Nater, U.M. (2009)
Determinants of salivary a-amylase in
humans and methodological consider-
ations, Psychoneuroendocrinology,34,
Rolando, M. (1984) Tear mucus ferning test
in normal and keratoconjunctivitis sicca
eyes, Chibret Int. J. Ophthalomol.,2, 32–
Seifter, J.L. (2007) Acid base disorders. In:
Goldman, L. and Ausiello, D. (eds) Cecil
Medicine, 23rd edn, Philadelphia, Pa,
Saunders Elsevier, 98–107. chap 119.
¨nen, O., Fisk, W. and Lei, Q.H. (2006)
Room Temperature and Productivity in
Office Work, eScholarship Repository,
Berkeley, California, Lawrence Berkeley
National Laboratory, University of Cali-
Skwarczynski, M.A., Melikov, A.K.,
Kaczmarczyk, J. and Lyubenova, V.
(2009) Preferred velocities with individiu-
ally controlled facially applied air
movement at high humidity, In: Proceed-
ings of Healthy Buildings 2009,on
Sung, E.J., Min, B.C., Kim, S.C. and Kim,
C.J. (2005) Effects of oxygen concentra-
tions on driver fatigue during simulated
driving, Appl. Ergon.,35, 25–31.
Tanabe, S. and Nishihara, N. (2004)
Productivity and fatigue, Indoor Air,
14(Suppl 7), 126–133.
Tham, K.W. (2004) Effects of temperature
and outdoor air supply rate on the
performance of call center operation in
the tropics, Indoor air,14(Suppl 7), 119–
Tham, K.W. and Willem, H.C. (2010) Room
air temperature affects occupantsÕphysi-
ology, perceptions and mental alertness,
Build. Environ.,45, 40–44.
Toftum, J. and Wyon, D.P. (2005) Remote
performance measurement (RPM) - a
new, internet-based method for the
measurement of occupant performance in
office buildings. In: Proceeding of Indoor
Air 2005, Vol. 1, Beijing, on CD-ROM,
Toftum, J., Jorgensen, A.S. and Fanger, P.O.
(1998) Upper limits of air humidity for
preventing warm respiratory discomfort,
Energy Build.,28, 15–23.
Toftum, J., Langkilde, G. and Fanger, P.O.
(2004) New indoor environment
chambers and field experiment offices for
research on human comfort, health and
productivity at moderate energy expen-
diture, Energy Build.,36, 899–903.
Wargocki, P. (2011) Productivity and health
effects of high indoor air quality.
In: Nriagu, J.O. (ed.) Encyclopedia of
Environmental Health, Vol. 4, Burlington,
Elsevier Science, 688–693.
Wargocki, P. and Wyon, D.P. (2007) The
effects of moderately raised classroom
temperatures and classroom ventilation
rate on the performance of schoolwork by
children (1257-RP), HVAC & R Res.,13,
The effects of thermal discomfort on health and human performance
Wargocki, P., Wyon, D.P., Baik, Y.K.,
Clausen, G. and Fanger, P.O. (1999)
Perceived air quality, sick building
syndrome (SBS) symptoms and pro-
ductivity in an office with two dif-
ferent pollution loads, Indoor Air,9,
Wargocki, P., Wyon, D.P., Sundell, J.,
Clausen, G. and Fanger, P.O. (2000) The
effects of outdoor air supply rate in an
office on perceived air quality, Sick
Building Syndrome (SBS) symptoms
and productivity, Indoor Air,10,
Wargocki, P., Delewski, M. and Haneda, M.
(2009) Physiological effects of thermal
environment on office work, In: Healthy
Buildings 2009, on CD-ROM.
Wientjes, C.J.E., Grossman, P. and Gaillard,
A.W.K. (1998) Influence of drive and
timing mechanisms on breathing pattern
and ventilation during mental task
performance, Biol. Psychol.,49,
Willem, H.C. (2006) Thermal and Indoor Air
Quality Effects on Physiological
Responses, Perception and Performance of
Tropically Acclimatized People, PhD
Thesis, Singapore, National University of
Winder, R. and Borrill, J. (1998) Fuels
for memory: the role of oxygen and
glucose, Psychopharmacology,136, 349–
Witterseh, T., Wyon, D.P. and Clausen, G.
(2004) The effects of moderate heat stress
and open-plan office noise distraction on
SBS symptoms and on the performance of
office work, Indoor Air,14(Suppl. 8), 30–
Wyon, D.P. and Wargocki, P. (2006) Room
temperature effects on office work. In:
Clements-Croome, D. (ed) Creating the
Productive Workplace, 2nd edn, London,
Taylor & Francis, 181–192.
Wyon, D.P., Fanger, P.O., Olesen, B.W. and
Pedersen, C.J.K. (1975) The mental
performance of subjects clothed for
comfort at two different air temperatures,
Ergonomics,18, 359–374.
Yoshitak, H. (1973) Occupational Fatigue
Approach from Subjective Symptom,
Tokyo, Japan, The institute for science of
labor (in Japanese).
Lan et al.
... For example, Nwanaji-Enwerem et al. (2020) show that indoor spaces with poor air quality increase the rate and likelihood of COVID-19 infection. Similarly, Wu et al. (2020) suggest that every 10 mg/m 3 increment in UFPs (Ultra-fine particles) and nitrogen oxides could cause at least a 15-22% increase in death rates due to Factors such as ventilation rate, outdoor climatic conditions, pollution and human activities play significant roles in determining the prevalent IAQ and often fault the goal of achieving optimal indoor air-quality levels (Lan et al., 2011) in a given indoor space. For example, Hosseini et al. (2020) pointed out that the overuse of disinfectants (e.g. ...
... Achieving IAQ where the air has zero concentration of harmful pollutants and where 80% or more of the workers are satisfied (Lan et al., 2011) is complex (Al Horr et al., 2016). That said, Kosonen and Tan (2004) suggested that increasing the ventilation rate enables more fresh air into the building and removes CO 2 and other air pollutants. ...
Full-text available
Purpose Achieving an appropriate indoor environment quality (IEQ) is crucial to a green office environment. Whilst much research has been carried out across the globe on the ideal IEQ for green offices, little is known about which indoor environment New Zealand office workers prefer and regard as most appropriate. This study investigated New Zealand office workers' preference for a green environment. Design/methodology/approach Workers were conveniently selected for a questionnaire survey study from two major cities in the country – Wellington and Auckland. The perception of 149 workers was analysed and discussed based on the workers' demographics. The responses to each question were analysed based on the mean, standard deviation, frequency of responses and difference in opinion. Findings The results showed that workers' preferences for an ideal IEQ in green work environments depend largely on demographics. New Zealand office workers prefer work environments to have more fresh air and rely on mixed-mode ventilation and lighting systems. Also New Zealand office workers like to have better acoustic quality with less distraction and background noise. Regarding temperature, workers prefer workspaces to be neither cooler nor warmer. Unique to New Zealand workers, the workers prefer to have some (not complete) individual control over the IEQ in offices. Research limitations/implications This study was conducted in the summer season, which could have impacted the responses received. Also the sample size was limited to two major cities in the country. Further studies should be conducted in other regions and during different seasons. Practical implications This study provides the opportunity for more studies in this area of research and highlights significant findings worthy of critical investigations. The results of this study benefit various stakeholders, such as facilities managers and workplace designers, and support proactive response approaches to achieving building occupants' preferences for an ideal work environment. Originality/value This study is the first research in New Zealand to explore worker preferences of IEQ that is not limited to a particular building, expanding the body of knowledge on workers' perception of the ideal work environment in the country.
... In both studies, perceived thermal satisfaction appears to be correlated mainly to the question of perceived air quality and not other perceived factors. These are not novel findings as prior studies have observed similar conditions[11,72,76].On inspecting potential correlations between subjective and objective (instrumentedmeasured) variables(Figure 3.11), the similarities between both datasets continue.There are few relevant correlations between any parameter, at least as far as estimates of the Kendall-τ b coefficient are concerned. It is possible that data sample size is a potential issue in both studies as the significance for the majority of potential correlations is poor.With regards to comparing measured variables(Figure 3.12), though some similarities exist, differences between both datasets are also more evident than in the above assessments. ...
Full-text available
The judgment of thermal comfort is a cognitive process that is influenced, not only by measurable indoor environmental conditions but also by less tangible aspects of an occupant’s well-being and overall satisfaction. Recent studies have examined the multi-domain nature of thermal comfort to bridge the performance gap between model-predicted and measurements of thermal comfort. This thesis seeks to inform a well-known research gap with respect to standard models of thermal comfort: that seminal data-informed models have not always accurately predicted true thermal comfort observations from independent field studies. This thesis presents a novel approach that involves the use of Bayesian inference to predict thermal comfort as a function of both thermal and non-thermal metrics of indoor environmental quality. Bayesian regression was performed on a large field dataset to investigate whether perceived thermal comfort can be attributed in a measurable and/or significant manner to one or several non-thermal parameters of indoor environmental quality. Posterior results revealed that higher CO₂ concentrations are independently correlated with lower incidences of thermal satisfaction in open-plan offices. At indoor temperatures of 23.5 ℃, the probability of an occupant feeling thermally satisfied at measured CO₂ levels of 550 ppm was 0.62 [0.54 - 0.69, 95% CrI]. This decreased to 0.28 [0.17-0.42, 95% CrI] at 750 ppm. Further, this is the first work to demonstrate that predictions of thermal comfort can be improved upon adding measurements of indoor CO₂ concentrations. The new data-driven thermal comfort model is integrated into a building energy model framework to predict occupants’ thermal satisfaction based on thermal indoor environmental conditions and ventilation rates. Four different post-COVID- 19 occupancy schedules were investigated to reflect and compare different occupancy profiles for post-COVID-19 hybrid work models. The simulation results showed that it might be possible to increase the ventilation rates with minimal building heating energy demand increase while maintaining the levels of occupants’ thermal comfort. This thesis presented a solution for building managers that have been under pressure to increase the current amounts of fresh air to lower the risk of spreading the COVID-19 virus, and other diseases, indoors.
... People feel fresher when the air is colder and drier [58][59][60]. Pang et al. [61,62] also found that the perception of pollution is heavier than higher air enthalpy with the same pollutants. ...
Full-text available
The odor generated by air conditioners is an important factor influencing the perceived air quality in buildings. In this study, different types of air conditioners and air filters were investigated to study the level of odor emission related to the operation state of the compressor, to identify the odor compounds and to analyze the cooling setpoint temperature on emitted odor intensity. Results show that the odor from constant frequency air conditioner use is periodic and stronger than that from variable frequency air conditioner use due to the different operation strategies of the compressor, which affect the evaporation of condensed water on the surface of the cooling coil. Ethyl acetate, acetic acid, 2-ethyl-1-hexanol, acetaldehyde, hexanal, nonanal, toluene and n-hexane are identified as odor compounds by Odor Active Value (OAV), Gas Chromatography/Olfactory/Mass Spectrometry (GC/O/MS) and Flavornet methods. The higher cooling setpoint temperature would lead to stronger odor, due to greater release of hydrophilic odorous compounds from condensed water. In our opinion, reducing the residual condensed water in air conditioners may be the key to control odor emission before purification.
... According to reviews of Hancock et al. [58] and Pilcher et al. [59], room temperature can also impact cognitive performance. Short-term free recall tasks [60], addition and visual tasks [61] and working memory tasks [62] have all shown that cognitive function suffers at higher temperatures. So the hypothesis in this study is : H1: There is a significant influence between mental workload variables on the work environment Several studies have stated that mental effort will affect performance. ...
Full-text available
ATC (Air Traffic Control) is considered one of the most demanding jobs. This profession is considered a job with high mental workload due to its high-stress level and great responsibility. This study designed a suitable work system to improve operator performance by measuring the mental workload and the physical environment using the NASA-TLX method and safety concept by considering variables affecting the operator’s perfor­mance. This study also searched for the impact of mental workload on the work environment, the mental workload on performance, and the work environment on performance. Questionnaires were distributed to operators, and validation and verification tests were carried out using SPSS. At the PLS method's processing stage, the variables used in this study consisted of the dependent (Y) and independent (X) variables. The dependent variables in this study were performance and the physical environment of work of the operator. Meanwhile, the independent variable was mental workload. Based on the mental load calculation, an average WWL (weighted workload) score of 80 to 90 was obtained, and the factors affecting mental workload are performance aspects and mental demand. Based on the results of structural modelling with the PLS method, there was a significant influence between mental workload on the work environment, the mental workload on perfor­mance and the work environment on operator performance. The proposed work system design used an ergonomic approach, safety and regulation of Ministry of Health to get an ergonomic work system, regulate the equal distribution of workloads, create a safe and comfortable working environ­ment, and improve operator performance. The design focused on the ATC tower's workstations and work environments. Supervisor has accepted the design.
... Seppänen and Fisk investigated the relationship between the temperature of a building and staff productivity [25]; the results confirmed maximum productivity at 21.6 °C. There are several other research that has studied the relationship between temperature and users' productivity and have indicated the importance of heating comfort in workspaces [26][27][28]. In addition, indoor air quality has been indicated to have a positive impact on productivity, health, and performance [29]. ...
Full-text available
Indoor environmental quality (IEQ) is an effective factor in evaluating the performance of employees in the workplace. This paper aims to investigate the IEQ of an office building of Imam Khomeini International University (IKIU), by evaluating the relationship between staffs' satisfaction and the orientation, window to wall ratio (WWR), and their gender. The results indicated that the size and landscape of the rooms, WWR, place of worktables, cooling and heating facilities and lighting systems, upgrading partitions and adding new spaces without increasing systems capacity, and the shared space usage by multi-users are the key factors that impact users’ satisfaction. Moreover, user comfort did not only depend on the features and equipment of the building and physical and physiological factors, but also on the habits, culture, and expectations of individuals. The results showed the same thermal satisfaction for both genders in the warm season and slightly higher dissatisfaction of females (4.62% higher compared to men) in the cold season. In addition, the main sources of noise were from the doors and the students passing the hallways. In conclusion, improving indoor air quality and thermal comfort were the most important ways to improve users' performance. This study is the first research concentrated on evaluating the current status of offices and presenting solutions to improve the IEQ factors in order to improve IKIU employees’ performance.
... At 31°C, participants reported the temperature conditions as significantly warmer and less comfortable. Even though studies showed that temperature changes could induce changes in the heart rate (Lan et al., 2011), we found no differences between thermal conditions. Differences in results could be explained as the mentioned study compared neutral to warm changes, while participants experienced only warm indoor conditions in our study. ...
Conference Paper
Full-text available
Personal environmental control systems (PECS), such as fans, have been widely implemented as an effective strategy to increase energy efficiency and occupants' satisfaction with indoor environmental conditions. This paper explores significant differences between thermal sensation votes and participants' physiological responses when using personal ceiling fans. In an experimental study in summer of 2018, 45 participants were exposed to two thermal conditions (28°C and 31°C) and different airflow speeds and directions in a climate chamber that simulates a typical office environment. Indoor environmental, psychological and physiological responses (skin temperature and heart rate) were recorded during the entire session. We tested differences in physiological responses between different demographic, contextual groups and airspeed levels. Results showed that at 31°C, participants had a significantly higher distal skin temperature and that airspeed helped reduce proximal skin temperature. Overweight participants showed a significantly lower proximal skin temperature than average weight participants. Heart rate results yielded statistically significant differences between age groups. Besides, findings suggest that skin temperature follows indoor temperature changes. By increased airspeed, physiological adaptations can be stimulated to restore comfort. Overall, personal ceiling fans are an effective cooling solution that can target occupants' body parts and individual characteristics to increase their comfort.
Occupant satisfaction is influenced by the interaction and combined effects of multi-domain indoor environmental quality (IEQ) in buildings. While there are literature reviews on multi-domain IEQ, two limitations exist: 1) the reviewed IEQ interactions are mainly limited to two IEQ domains and climate chamber studies; and 2) those reviews have not quantitatively discussed the impact of IEQ interactions on overall satisfaction and only cover limited building types. To address this gap, this study adopts a broad, top-down approach with a three-stage screening to select relevant publications about the four most concerned IEQ domains—thermal comfort, acoustic comfort, visual comfort, and indoor air quality. A total of 74 publications have been selected and reviewed. Our study shows that: 1) existing findings on IEQ interaction and combined effects are inconsistent and somewhat contradictory; 2) interactions among three IEQ domains exist but are much less examined compared to two IEQ interactions. Studies on four IEQ are not found; 3) the current occupant satisfaction models are mainly generalized models. Personal models are an emerging area of research and practice; and 4) the one-vote veto effect exists, i.e., some IEQ domains show a predominating influence on overall satisfaction. We advocate future research 1) to collaborate on a global IEQ database to enable model comparison and the development of standards and guidelines spanning multiple IEQ domains, 2) to investigate personal overall satisfaction models encompassing IEQ interactions; and 3) to further examine non-linear modeling approaches for interpreting the relationships between IEQ and overall satisfaction.
We performed a cross‐sectional survey of 2143 female students in a university in Tianjin, China regarding perceived air quality (PAQ) and sick building syndrome (SBS) symptoms in the student dormitory. The prevalence of general, mucosal, and skin symptoms was 22.1%, 21.9%, and 26.3%, respectively. The three most prevalent PAQ complaints were “dry air” (48.9% often), “stuffy odor” (18.2%), and “other unpleasant odors” (5.1%), and they were significant risk factors for 11–12 out of 12 SBS symptoms (adjusted odds ratios [AOR]: 1.6–5.8). Survey data of 1471 undergraduates, whose dorms were of uniform layout and furnishing, were used to further investigate the influences of occupancy level and occupant behaviors on PAQ and SBS symptoms. Frequent use of air freshener/perfume was a significant risk factor for “dry air,” less frequent room cleaning and higher occupancy density were significant risk factors for “stuffy odor,” and less natural ventilation was a significant risk factor for both “stuffy odor” and “pungent odor.” These factors were also significantly associated with some SBS symptoms. In particular, the use of air freshener/perfume exhibited a significant dose–response pattern with “fatigue” (sometimes: AOR 1.3; often: AOR 2.0) and with “irritated, stuffy, or runny nose” (sometimes: AOR 1.6; often: AOR 2.2).
Full-text available
Non-territorial offices have been a growing architectural trend as they save costs and space, while maximizing the number of available workstations. The change to desk sharing is even more significant after the pandemic, with more companies switching to hybrid mode. The risk with non-territorial offices lies in lack of attachment to workplace in the employees, as these spaces scrap away any chance for personalization to the environment. Attachment as an affective bond to environments and places has a deep psychological implication in workplaces. The fundamental characteristic of the concept of attachment (and place attachment in particular) is the proximity-seeking behavior that draws the person closer to the attachment subject. Place attachment has been said to rely on social features and physical features. Attachment to workplace results in employees’ comfort, job satisfaction, development of commitment and organizational citizenship behaviors. Several models of people–place relationships have been proposed, including the PPP model by Scannell and Gifford which highlights place loss and resulting emotion, attitudes, and behaviors relevant for workplace change processes. From a behavioral perspective, one can study attachment as a habit formation behavior. This motivational/emotional behavior therefore underpins the mesolimbic dopaminergic system involved in reward mechanisms as well as seeking mechanisms. Considering the appraisal theory, the affective bond shapes because the environment (1) is predictable to offer security and support survival (2) supports achievement of physical and cognitive goals (3) matches one’s personal values (4) supports one’s expectations based on past experiences. Operationalization of place attachment helps architects to design attractive work environments that evoke this emotion.
It has been observed that high temperature exposure is associated with a reduction in lung function and some possible biological mechanisms have been suggested. However, it is unclear if thermal perception plays a role in the association. From September 3rd to 15th, 2018, in Guangzhou, China, we repeatedly measured daily thermal perception and lung function among 126 participants with outdoor military training. We performed a linear mixed model and stratified analyses by the origin of students, gender, and the training period to evaluate the effects of thermal perception on lung function. A total of 399 measurements were collected. Per vote increase in thermal sensation vote towards the “hot” direction was associated with a − 0.04 L (95% CI: − 0.08 to − 0.01) decrease in forced vital capacity (FVC), and − 0.04 L (95% CI: − 0.08 to − 0.01) decrease in forced expiratory volume in 1 s (FEV1). Per grade increase towards the “very uncomfortable” direction for thermal comfort vote was associated with an increased percentage of forced expiratory volume in 1 s (FEV1%) by 1.52% (95% CI: 0.18 to 2.86). For thermal preference, with preferred cooler vote increased by one level, FVC and FEV1 decreased by − 0.05 L/s (95% CI: − 0.08 to − 0.02) and − 0.05L/s (95% CI: − 0.08 to − 0.02), respectively. The effects of thermal perception on lung function were stronger among non-local and in the first week of training. Our study suggests that in the same high-temperature environment, thermal perception is associated with lung function, even in healthy adults.
Full-text available
A re-analysis of two independent laboratory studies was made in which a total of 60 female subjects had been exposed for several hours to 6 different air quality conditions in groups of 6 people at a time. The subjects performed typical office tasks at their own pace during exposures. Measured carbon dioxide (CO2) concentrations and outdoor air supply rates were used to calculate CO2 produced by subjects at each air quality level. The re-analysis showed that CO2 produced by subjects was affected by air quality (P<0.015). It decreased by ca. 13% when the percentage dissatisfied with the perceived air quality increased from 8% to 40%, indicating a dose-response relationship. A change in breathing pattern (shallow breathing) or a slow down of work rate in polluted air would both reduce metabolic rate and thus the CO2 production rate.
Full-text available
A subjective experiment was conducted in a climate chamber to investigate the effects of thermal discomfort (feeling too warm) on the performance of office work. Twenty-seven Danish female subjects were exposed in a climate chamber to four conditions with different levels of thermal discomfort provided by a combination of operative temperature and amount of clothing. Thermal sensation votes towards the end of exposures were neutral, slightly warm, warm and very warm. More symptoms indicating mental fatigue were observed with increased thermal discomfort. The subjects reported that more effort was necessary when they felt thermally warm compared to conditions in which they felt thermally neutral and slightly warm. Performance of proof-reading, addition and text-typing tasks was not affected by thermal discomfort. This suggests that the subjects were able to maintain their performance but as a result they got more tired and the mental work load increased.
The main objectives of this paper were to examine the physiological reactions of the occupants being the result of thermal discomfort and whether these reactions have consequences for the performance of office work. Thirty-two Danish female subjects performed simulated office work in a chamber at four combinations of operative temperature and clothing insulation to obtain different levels of thermal sensation from neutral to warm. Thermal sensation votes were strongly correlated with the temperature of the skin. Skin moisture increased at increased temperatures. Tear-film stability and blinking rate were not affected. Concentration of salivary alpha-amylase and cortisol decreased with increased thermal discomfort which may suggest that the performance of simulated office work could be affected, but it was not. Carbon dioxide concentrations were similar in different conditions suggesting that the metabolic rate of subjects was unchanged, but the subjectively reported work load and effort increased significantly with increased thermal discomfort.
People spend a significant part of their life indoors mainly in the built environment (in public and residential buildings). In developed parts of the world, the proportion of time spent indoors can be as high as 80% to 90%. Many non-industrial buildings do not provide adequate conditions as regards indoor environmental quality especially indoor air quality. This is due to elevated exposures to air pollutants. These conditions reduce quality of life by increasing the risks for health problems that can, among others, result in disability to perform work. Significant societal costs are then generated including costs to individuals, building owners and employers. There are potentially considerable health and productivity benefits of improving indoor air quality in non-industrial buildings. Crude estimates suggest that 2 million healthy life years can be saved in Europe by avoiding exposures to air pollutants indoors in non-industrial buildings. Similar estimates have been made for the U.S. as regards exposures to air pollutants in residential buildings. The potential annual savings and productivity gains have been estimated to be as high as $168 billion in the U.S. (1997 estimate as no newer data are available). A saving of $400 per employee per year (2000 estimate) was estimated due to reduced absenteeism being the result of improving indoor air quality. In Europe, the annual productivity benefits were estimated to be at the level of ca. €330 per worker (2000 estimate as no newer data are available). Despite the obvious significance, the potential health and productivity benefits have not yet been integrated in the conventional economic calculations pertaining to building design and operation. Such integration would provide economic arguments for applying measures to reduce air pollution and thereby would support the arguments that arise from the public health perspective. The present article attempts to provide such arguments by summarizing the potential costs and economic benefits of improved indoor air quality.