The effects of air temperature on office workers' well-being, workload and productivity-evaluated with subjective ratings

Abstract

Productivity bears a close relationship to the indoor environmental quality (IEQ), but how to evaluate office worker's productivity remains to be a challenge for ergonomists. In this study, the effect of indoor air temperature (17 °C, 21 °C, and 28 °C) on productivity was investigated with 21 volunteered participants in the laboratory experiment. Participants performed computerized neurobehavioral tests during exposure in the lab; their physiological parameters including heart rate variation (HRV) and electroencephalograph (EEG) were also measured. Several subjective rating scales were used to tap participant's emotion, well-being, motivation and the workload imposed by tasks. It was found that the warm discomfort negatively affected participants' well-being and increased the ratio of low frequency (LF) to high frequency (HF) of HRV. In the moderately uncomfortable environment, the workload imposed by tasks increased and participants had to exert more effort to maintain their performance and they also had lower motivation to do work. The results indicate that thermal discomfort caused by high or low air temperature had negative influence on office workers' productivity and the subjective rating scales were useful supplements of neurobehavioral performance measures when evaluating the effects of IEQ on productivity.

Full text

The effects of air temperature on office workers’well-being, workload and
productivity-evaluated with subjective ratings
Li Lan, Zhiwei Lian
*
,LiPan
School of Mechanical engineering, Shanghai Jiao Tong University, Shanghai 200240, China
article info
Article history:
Received 29 December 2008
Accepted 4 April 2010
Keywords:
Air temperature
Productivity
Workload
Motivation
Well-being
abstract
Productivity bears a close relationship to the indoor environmental quality (IEQ), but how to evaluate
office worker’s productivity remains to be a challenge for ergonomists. In this study, the effect of indoor
air temperature (17

C, 21

C, and 28

C) on productivity was investigated with 21 volunteered partic-
ipants in the laboratory experiment. Participants performed computerized neurobehavioral tests during
exposure in the lab; their physiological parameters including heart rate variation (HRV) and electroen-
cephalograph (EEG) were also measured. Several subjective rating scales were used to tap participant’s
emotion, well-being, motivation and the workload imposed by tasks. It was found that the warm
discomfort negatively affected participants’well-being and increased the ratio of low frequency (LF) to
high frequency (HF) of HRV. In the moderately uncomfortable environment, the workload imposed by
tasks increased and participants had to exert more effort to maintain their performance and they also
had lower motivation to do work. The results indicate that thermal discomfort caused by high or low air
temperature had negative influence on office workers’productivity and the subjective rating scales were
useful supplements of neurobehavioral performance measures when evaluating the effects of IEQ on
productivity.
Ó2010 Elsevier Ltd. All rights reserved.
1. Introduction
Studies (Roelofsen, 2002; Woods, 1989; Lorsch and Ossama,
1994) have shown that productivity bears a close relationship to
the indoor environment quality (heat, cold, noise, light, etc.).
However, howto assess the effect of indoor environment quality on
productivity remains to be the major challenge for ergonomists.
Until now there have been no standard procedures to measure
office worker’s productivity, therefore it has been difficult to
persuade clients to accept the concept of a relationship between
economic productivity benefits and indoor environment quality. A
neurobehavioral approach had been proposed to systematically and
quantitatively evaluate the effects of indoor environment quality on
office worker’s productivity by several neurobehavioral tests,
which assessed four classes of neurobehavioral functions involved
in office work, including perception, learning and memory,
thinking, and executive functions (Lan et al., 2009a). In the previous
laboratory study some neurobehavioral tests had been used to
investigate the effect of air temperature on productivity for a very
short time (about 30 min). It was found that motivated people
could maintain high performance for a short time under adverse
(hot or cold) environmental conditions. However, it is unclear why
the participants could maintain their performance and whether it
will happen in the routine office work environment. In this paper,
several subjective ratings as well as physiological measurements
were made to investigate the effects of thermal discomfort on office
workers’productivity. The object was to investigate the effects of
thermal discomfort on occupants’workload, emotion, well-being,
and motivation, and also the relationship between these activities
and their neurobehavioral performance.
2. Methodologies
The experiment was carried out in an ordinary but low-polluting
office (LWH¼645 m), in which participants sat at seven
workstations, each consisting of a table, a chair and a personal
computer (Fig. 1). The office was illuminated with eight fluorescent
lamps. The room temperature was controlled by an air-conditioner,
which was able to adjust temperature from 16

Cto32

C.
2.1. Participants
Twenty-one volunteered participants (6 females and 15 males)
were recruited for this experiment. They were recruited based on
*Corresponding author. Tel.: þ86 21 34204263; fax: þ86 21 34206814.
E-mail address: zwlian@sjtu.edu.cn (Z. Lian).
Contents lists available at ScienceDirect
Applied Ergonomics
journal homepage: www.elsevier.com/locate/apergo
0003-6870/$ esee front matter Ó2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.apergo.2010.04.003
Applied Ergonomics 42 (2010) 29e36

the following criteria: familiarity with computer, impartiality to the
office in which the study was carried out, and absence of chronic
diseases, asthma, allergy, hay-fever and color blind. The participants
were all students, aged 18e20 years (
m
¼19,
s
¼1). Participants
were required to wear long-sleeved sweater, long thick trousers and
long underwear top and bottoms, socks and shoes (an estimated
clothing insulation value of 1.20 Clo, including the insulation of the
chair). They were also asked to have a good rest at the night before
the experiment. The participants were paid a salary for participation
in the experiment at a fixed rate per hour. To increase their moti-
vation and especially to encourage them to perform the tests seri-
ously, bonus was paid depending on their performance. Participants
all successfully completed experimental sessions.
All protocols were approvedby the university’s ethics committee
and conformed to the guidelines contained within the Declaration of
Helsinki. Verbal and written informed consent was obtained from
participants before they participated in the experiment.
2.2. Measurements
2.2.1. Physical measurements
The temperature, relativity humidity and velocity of the air were
measured. The mean radiant temperature was estimated from the
globe temperature, which was measured using a 150 mm diameter
black globe thermometer.The illuminant wasmeasured with a digital
luxmeter near each participant at the height of working face.
2.2.2. Physiological measurement
Participants’electrocardiogram (ECG) and electroencephalo-
graph (EEG) were recorded by a Powerlab 8/30 system (AD
Instruments, Australia) and the accessorial materials (e.g. adhesive
pads, EEG or ECG electrodes, etc.). The standard bipolar limb leads
(ML-1340) and adhesive ECG electrodes (MLA-1010) were linked to
the data acquisition system through a dual bioamplifier (ML-135)
for recording of ECG. The negative lead and the positive lead were
clung to the right and left wrist, respectively, and the ground lead
was clung to the right ankle. The EEG recordings were made using
the bipolar method: a pair of electrode measured the difference in
electrical potential (voltage) between the two positions above the
brain, one being placed on the right side of the head (60% of
distance from right ear to midpoint of the scalp), another being
placed in the centre of the forehead, and a third electrode being put
on the earlobe as a point of reference. There are three main spectral
peaks of HRV distinguished in a spectrum calculate from 2 to 5 min
ECG recordings: very low frequency (VLF: 0.003e0.04 Hz), low
frequency (LF: 0.04e0.15 Hz), and high frequency (HF: 0.15e0.4 Hz)
(Sayers, 1973). The ratio of low frequency to high frequency (LF/HF)
is usually introduced to infer the sympathetic nervous system
activity (Jaffe et al., 1993). Based on a 5-min record, the relative
power of four EEG bands (
d
-band EEG: 0.5e4 Hz;
q
-band EEG:
4e8 Hz;
a
-band EEG: 8e14 Hz;
b
-band EEG: 14e35 Hz in
frequency) and the value of LF/HF were calculated.
2.2.3. Performance measurement
Performance refers to the quality and quantity of finished task.
Thirteen computerized neurobehavioral tests were used to evaluate
participant’s performance. The detailed description of these tests
can be referred to Lan et al. (2009a) and Lan and Lian (2009b).
2.2.4. Subjective measurements
2.2.4.1. Thermal sensation votes. Thermal sensation votes (TSV)
were cast on the ASHRAE/ISO 7-point thermal sensation scale
(ASHRAE, 2001). Thermal comfort (TC) votes were cast on 5-point
numerical scales ecomfortable (0), slightly uncomfortable (1),
uncomfortable (2), very uncomfortable (3), and extremely
uncomfortable (4).
2.2.4.2. The Profile of Mood States. Participants’emotion was
investigated with the Profile of Mood States (POMS). The POMS has
been widely used as a research and clinical instrument. It consists
of 65 adjectives describing feeling and moods that the participant
may have experienced at that moment, or over the previous day,
few days, or week, including key words such as unhappy, tense,
careless, and cheerful (McNair et al., 1992). Each item is scored on
a 5-point Likert-type scale ranging from 0 (not at all) to 4
(extremely) according to participants’responses. The POMS
consists of six identifiable mood states: tension, depression, anger,
vigor, fatigue, and confusion. A composite score, the total mood
disturbance (TMD) score, is computed by summing each of the
individual score for tension, depression, anxiety, fatigue and
confusion, with vigor scores subtracted to indicate participant’s
total mood disturbance. The higher TMD score indicates higher
negative mood.
2.2.4.3. Well-being and motivation. Motivation refers to the
worker’s mental drive and enthusiasm to carry out a work. If
a person is reluctant to work, then no matter how accurately he can
perform, the productivity should be deteriorated. Well-being
Fig. 1. (a) The layout of the experimental room (1) consisting of a table, a chair and
a personal computer, (2) air conditioner and (3) waiting room; (b) a picture during the
experiment.
L. Lan et al. / Applied Ergonomics 42 (2010) 29e3630

indicates the physical and mental health of a worker. Unless
a person feels a positive sense of well-being he will not perform as
effectively, and an optimum level of productivity will not be ach-
ieved. Health factors influenced by the indoor environment are
(Clements-Croome and Kaluarachchi, 2001): (1) respiratory prob-
lems (dryness, hoarseness, dry/sore throat, changes in voice,
wheezing); (2) skin problems (soreness, itching, dry skin, rashes);
(3) nervous problems (headaches, nausea, drowsiness, tiredness,
lethargy, reduced mental capacity, dizziness, forgetfulness,
fatigue); (4) nasal-related problems (itchy or teary eyes, runny
nose, asthma-like symptoms among non-asthmatics); and (5) odor
complaints (changes in odor, unpleasant odors or tastes).
Motivation scale includes two items relating to worker’s willing
and enthusiasm to perform task. They were assessed on a 7-point
scale ranging from 1 (very low) to 7 (very high). Well-being was
assessed by 18-item questions covering comfort and health factors
influenced by the indoor environment on a 5-point Likert-type
scale ranging from 0 (not at all) to 4 (extremely).
2.2.4.4. National aeronautics and space administration-task load
index (NASA-TLX). NASA-TLX is found to be the most valid measure
of subjective workload, to have the highest user acceptance, and to
have the smallest between-subjects variability. NASA-TLX is
a multidimensional, self-reported assessment technique that
provides an estimation of the overall workload associated with task
performance and mental effort (Hart and Staveland, 1988). In TLX,
workload is defined as the “cost incurred by human operators to
achieve a specific level of performance.”The overall workload score
(OW) is a weighted average of ratings on six underlying psycho-
logical factors, including physical demand (PD), mental demand
(MD), temporal demand (TD), own performance (OP), effort (EF),
and frustration (FR) level. Three dimensions relate to the demands
imposed on the participant (PD, MD, and TD) and three to the
interaction of a participant with the task (EF, FR, and OP). NASA-TLX
involves a two-part evaluation procedure. The first requirement is
to obtain numerical ratings for each scale that reflect the magni-
tude of that factor in a given task on a 7-point bipolar scale ranging
from 1 (low) to 7 (high). Second, participants perform 15 pair-wise
comparisons of six workload scales. The number of times each scale
is rated as contributing more to the workload of a specific task is
used as the weight for that scale. Finally the overall workload score
can be computed by multiplying each rating by the weight given to
that factor. In this research, NASA-TLX was computerized and the
six subscales of workload for a specific neurobehavioral test were
evaluated immediately after the test was finished.
2.3. Experimental procedure
The effect of three indoor air temperatures (17

C, 21

C, and
28

C) on productivity was studied. The indoorair velocity was kept
under 0.1 m/s, and the relative humidity of air was not dependently
controlled. Within-subject design was applied for this experiment
and balanced Latin-square design was utilized to control the
carryover effects (Zhu, 2000).
The experiment was performed in three days, each day for 8.5 h
from morning to afternoon. Twenty-one participants were sepa-
rated into 3 groups with 7 participants (5 males and 2 females,
respectively) in a group, each group being exposed to the three
temperature conditions in one day. A standardized three by three
balanced Latin-square design was shown in Table 1, in which factor
Twas the experimental variable (T
1
e17

C, T
2
e21

C, T
3
e28

C),
factor Areferred to the three participant groups, and factor B
referred to the three successive time periods of each group. The
schedules of the whole day and each temperature condition are
shown in Fig. 2. There were two pauses among the three temper-
ature conditions: from 11:00 to 12:30 was lunch time, participants
being supplied with the same Chinese foods; and from 14:30 to
15:30 was the second interval time, during which participants went
out of the office for a break and they could consume non-carbon-
ated water and biscuits which were freely available at the waiting
room. Each temperature condition lasted for a total of 120 min.
First, participants entered the office reading book or playing games
for 40 min to adapt to the indoor environment. During this period,
physical parameters were measured. Following the exposure,
participants spent about 60 min on performing the computerized
neurobehavioral tests and assessing workload. After completing
the neurobehavioral tests, participants were instructed to assess
their general perceptions of the environment, emotions, well-
being, and motivation to work by filling in questionnaires for
10 min. Then the physiological parameters of one of the seven
participants were measured for 10 min. Participants were told
before the experiment that they could not leave the room until the
120 min session had reached.
A week before starting the experiment, participants practiced
the neurobehavioral test battery for 1 h. They were also instructed
on how to fill out the questionnaires.
2.4. Statistical analysis
The SPSS 13.0 (SPSS Inc., Chicago, IL, USA) program was
employed to make the statistical analysis. The data werefirst tested
for normality using ShapiroeWilk’sWtest; the significance level
Table 1
Balanced Latin-square design of this experiment.
B
1
B
2
B
3
A
1
T
1
T
2
T
3
A
2
T
2
T
3
T
1
A
3
T
3
T
1
T
2
Fig. 2. Schedule of the experiment: (a) schedule of the whole day and (b) schedule of
each exposure.
L. Lan et al. / Applied Ergonomics 42 (2010) 29e36 31

was set to be 0.05 (P<0.05). Normally distributed data were
subjected to analysis of variance in a repeated measures design
with each participant as her own control, thus excluding any
differences in experience, training, intellectual skills, etc., which
can influence performance greatly. HuynheFeldt statistics were
used to adjust the violation of sphericity. HuynheFeldt’sP-values
were based on corrected degrees of freedom, though the original
degrees of freedom were reported. Whenever necessary, post hoc
comparisons (Turkey HSD test) were then performed. Not normally
distributed data were analyzed using Friedman’s analysis of vari-
ance or Wilcoxon Matched Pairs test.
3. Results
3.1. Results of thermal condition
The measured physical parameters describing the indoor
climate of the office are shown in Table 2. They did not deviate from
the intended level, and also there were no large differences in the
relative humidity among the three conditions. The relationship
between thermal sensation votes and air temperature is shown in
Fig. 3. The relationship between thermal comfort votes and air
temperature is shown Fig. 4. The size of the plot in Figs. 3 and 4
represents the number of replies. It can be seen that participants
attained thermal neutrality and comfort at 21

C; they felt slightly
cool at 17

C and slightly warm at 28

C and most participants felt
slight discomfort at these two conditions.
3.2. Results of emotion
The effectof air temperatureon participants’emotion is illustrated
in Table 3. The reliability (Cronbach’s
a
)forthePOMStotalmood
disturbance and subscales (ranging from 0.65 to 0.92) suggests that
the POMS is a reliable scale being used to investigate the effect of
thermal environment on occupant’s emotion. It should be noted that
the higher score of negative moods, such as tension, depression,
anger, fatigue, and confusion, indicates higher negative emotion,
while the higher score of vigor indicates higher positive emotion. It
can be seen from Table 3 that participants’tension and anger emotion
and total mood disturbance were significantly affected by air
temperature (Repeated-measure ANOVA, P<0.05), they experi-
encing more negative moods at 28

C and less negative moods at 21

C
(Post hoc test, P<0.05). No significant emotional difference was
observed between 17

Cand21

C, or between 17

Cand28

C.
3.3. Results of well-being and motivation
The effect of air temperature on occupants’well-being and
motivation is shown in Fig. 5. It was found that occupants’percep-
tion of well-being and their motivation to work were also signifi-
cantly affected by air temperature(Friedman ANOVA, P<0.05). They
perceived better well-being at cool condition and neutral condition
than at warm (Wilcoxon Matched Pairs test, P<0.01). As to moti-
vation, they had significantly higher motivation to do work at
neutral condition than at warm condition (Wilcoxon Matched Pairs
test, P<0.05). No significance difference in well-being or motivation
was found between the neutral condition and the cool condition.
3.4. Results of workload and performance of neurobehavioral tests
The overall workload and the six subscales imposed by the
neurobehavioral tests under different temperature conditions are
Table 2
Mean values and STD of physical parameters inside the office.
Parameter 28 21 17
Air temperature (

C) 28.2 0.5 21.1 0.6 18.0 0.5
Relative humidity (%) 62.6 7.9 71.3 4.3 63.2 3.9
Air velocity (m/s) 0.1 0.03 0.08 0.02 0.1 0.03
Mean radiant temperature (

C) 28.1 0.4 21.0 0.5 18.1 0.5
Illuminant (lx) 560.2 56.3 561.4 54.7 562.1 54.0
Fig. 3. Change in thermal sensation votes with air temperature.
Fig. 4. Change in thermal comfort votes with air temperature.
Table 3
The effects of air temperature on subjective emotional ratings.
Internal
consistency
Mean (SD) P
Temperature (

C) 17 21 28 17 21 28
Total mood
disturbance (TMD)
0.87 0.83 0.88 26.8(5.1) 24.8(4.8) 32.6(5.1) 0.05*
Tension 0.80 0.75 0.70 7.8(1.0) 6.5 (0.7) 8.6(0.9) 0.05*
Depression 0.91 0.83 0.85 11.3 (1.8) 10.1(1.7) 12.1(1.7) 0.14
Anger 0.87 0.81 0.81 7.5(1.2) 6.3(0.9) 8.7(1.4) 0.04*
Vigor 0.85 0.90 0.83 15.5(1.2) 14.7(1.3) 14.5(1.3) 0.54
Fatigue 0.89 0.87 0.92 7.6(0.9) 8.6(1.1) 9.5(1.1) 0.06
Confusion 0.68 0.65 0.67 9.0(0.7) 8.9(0.7) 9.3(0.7) 0.79
*P<0.05.
L. Lan et al. / Applied Ergonomics 42 (2010) 29e3632

shown in Fig. 6. The air temperature had significant effects on the
perceived overall workload as well as the subscales (except the
physical demand subscale) (Friedman ANOVA, P<0.05). It can be
seen that participants perceived that workloads were significantly
higher at 17

C and 28

C than that at 21

C. The results of the six
subscales indicate that the neurobehavioral tests were character-
ized by significantly higher mental demand than physical demand
(Repeated-measure ANOVA, P<0.05), and well represent the
characteristics of current office works. Significant effect of air
temperature on effort was observed (Repeated-measure ANOVA,
P<0.05). Post hoc test shows that participants had to perform tests
with significantly more effort at 17

C than at 21

C.
Fig. 7 shows the relationship between indoor air temperature
and the performance of neurobehavioral tests based on regression
analysis. The real lines indicate the variation of accuracy and speed
with air temperature, and the broken lines illustrate the 95%
confidence interval. It can be seen from Fig. 7 the trend that the
accuracy and speed of neurobehavioral test decreased in the
moderately uncomfortable environment. However, no significant
effect of air temperature can be observed on the performance of
most neurobehavioral tests.
The correlations between workload and performance of neu-
robehavioral tests are shown in Table 4. The correlation factors
were analyzed with Spearman coefficients. It can be seen that
participants had to exert more effort to perform tasks when the
mental demand and time demand were higher. However, negative
correlation was found between exerted effort and the accuracy and
speed of neurobehavioral tests, indicating that although partici-
pants had exert more effort to do task no improvement on
performance could be achieved. Moreover, both accuracy and speed
of neurobehavioral tests decreased with the increase of overall
workload, that is, participants having poorer performance when
the workload was higher.
Table 5 illustrates correlations of overall workload (OW),
emotion (TMD), well-being (WB), motivation (MT), and the
performance (accuracy and speed) of neurobehavioral tests. The
correlation factors were analyzed with Spearman coefficients.
There is a positive correlation between motivation and well-being,
both of which decreased with increase of total mood disturbance
score and overall workload. But no significant correlation can be
found between the performance of neurobehavioral tests and
emotion, well-being, or motivation.
Fig. 5. Change in (a) perception of well-being and (b) motivation to work with air temperature.
Fig. 6. Change in overall workload and the six subscales with air temperature.
L. Lan et al. / Applied Ergonomics 42 (2010) 29e36 33

3.5. Results of HRV and EEG
In this study the physiological parameters of 3 participants were
measured. Fig. 8 shows the variation of LF/HF with air temperature. It
can be seen that the LF/HF increased with air temperature although
no significant statistical temperature effects was found; great
increase in the LF/HF ratio was found at the 28

C compared with
another two conditions. The global relative power of the four EEG
bands is shown in Fig. 9. The power of
d
-band EEG was significantly
affected by air temperature (Repeated-measure ANOVA, P<0.02).
The
d
-band EEG decreased at the 17 or 28

C compared with the
neutral condition (Post hoc test, P<0.05). No significant changes of
other three EEG bands were found but it can be seen from Fig. 9 that
both
a
-band and
b
-band EEG increased at the 17 and 28

C.
4. Discussion
This study suggests that participants maintained their perfor-
mance byexerting more effort when the workload demand increased
in thermal discomfort environment. Haneda et al. (2009) found
similar results when examining the effects of thermal discomfort on
the performance of office work.Just as it is defined, effort, consciously
protecting the level of task performance by ‘trying hard’, can coun-
teract an impaired operator state as well as enable dealing with
increased task demands (Hockey, 1986). On the other hand, studies
have validated the existence of a ‘cognitive reserve’,whereby
participants have at their disposal a certain amount of neural
resources that could be allocated to the performance of tasks and
activities. Performance of these tasks and activities would deteriorate
when the amount of resources was insufficient to deal with both the
task demands and thermal stress, such that participants would be
able to maintain their performance level until the resources were
overloaded (Hocking et al., 2001). Moreover, participants would like
to exert more effort to maintain performance if needed when they
were supplied with extra bonus for better performance. The invest-
ment of effort has the advantage that task performance remains at
a certain target level, but there are costs for this achievement. Short-
lasting effort investment is probably without health consequences
and is one of the advantages of human flexibility to deal with in
demands. However, prolonged, continuous effort compensation can
be a threat to good health, as it has been suggested that repetitive
activation of the cardiovascular defense response may lead to
hypertension (Johnson and Anderson, 1990).
The cost of effort investment can be partly illustrated by
participants’motivation to work: they had lower motivation to
work when they had to exert more effort to perform a task. Moti-
vation plays an important role in worker productivity. Heerwagen
(1998) showed that the relationship between buildings and
worker performance was related to both motivation and ability. An
individual has to want to do the task and then has to be capable of
doing it. In addition to effort investment, well-being is also related
to motivation to work. As to the health condition, participants
reported significantly more SBS symptoms at the warm condition.
Fang et al. (2004) also found the perception of air freshness and
acceptability improved greatly as temperature decreased and the
intensity of fatigue, headache and difficulty in thinking clearly
decreased at lower levels of air temperature. The conditioning of
the mucous membrane in the upper respiratory tract induced by
the temperature and water vapor difference between the inspired
air and the surface of the nasal passage potentially alleviated the
perceptions of poor air quality in the neutral and cool environment
(Berglund and Cain, 1989; Fang et al., 2004). Due to the above
reasons, high air temperature significantly reduced participants’
motivation to do their work. In fact, the confidence that the fight for
a healthy work environment and healthy workers is a prerequisite
for innovation and productivity in a knowledge-based economy, is
gaining more and more ground in companies (Hermans and
Peteghem, 2006).
Fig. 7. Change in performance of neurobehavioral tests: (a) accuracy; (b) speed with air temperature.
Table 4
Correlations among workload and performance of neurobehavioral tests.
PD MD TD OP EF FR Accuracy Speed
OW 0.26** 0.64** 0.52** 0.52** 0.76** 0.71** 0.19** 0.14**
PD 0.04 0.10** 0.10** 0.05 0.25** 0.02 0.02
MD 0.27** 0.14** 0.47** 0.36** 0.16** 0.09**
TD 0.06 0.30** 0.26** 0.02 0.02
OP 0.21** 0.37** 0.28** 0.11**
EF 0.51** 0.08*0.12**
FR 0.10** 0.11**
Accuracy 0.01
*P<0.05; **P<0.01.
L. Lan et al. / Applied Ergonomics 42 (2010) 29e3634

The role of emotion has been largely neglected in productivity
research. Frijda (1996) suggested that the ultimate function of
emotions is to potentiate or stimulate behavioral responses that, in
turn, influence the personeenvironment relation in the service of
some adaptive goal. Damasio (1994) also claimed that emotion can
provide an early warning to the cognitive system as to whether
things are going well or badly and indicate which things need
attention. In particular, intense and negative emotions may reduce
performance efficiency in several ways: (a) they may disrupt the
state regulation, which makes it less optimal for task performance
because of over-reactivity; (b) they may be so distracting that they
directly interfere with the processing of the task information; or (c)
they may cause psychosomatic complaints that also demand
attention. The correlations between emotion and motivation show
that negative emotion reduced participants’motivation to do work.
Participants experienced more negative moods at the warm
condition, so they even did not want to exert as much effort as the
slightly cool condition when they felt that the environment was
unfavorable and it was hard to maintain their performance.
However, no significant correlations can be found between
performanceof neurobehavioral tests and emotion, or well-being, or
motivation. There are several explanations for this dissociation; the
time lag between takingperformance measuresand subjectiverating,
or the dynamic interaction between people and their overt perfor-
mance. There is a continuous and dynamic interaction between
people and their surroundings that produces physiological and
psychological strain on the person. The humans’body is not a passive
system that responds to an environmental input in a way that is
monotonically related to the level of the physical stimulus (Parsons,
2000). Instead, acting as an adaptive agent, humans regulate the
effect of indoor environment imposed on their productivity. They
keep on appraising their environment consciously or unconsciously
and choose coping strategies and self-regulative processes that are
believed to be appropriate. Productivity is the final result of those
behaviorally manifestations of human psychological, physiological,
and neural functioning changes. Due to the flexibility of human
beings, the effects of indoor environment on productivity may not be
illustrated overtly on neurobehavioral performance during the
experimental period. In this study, participants succeeded to main-
tain their performance by exerting more effort in the moderately
uncomfortable environment when they were encouraged by extra
bonus. On the other hand, they also realized that they could leave the
environment after the short experiment, which helped them to
overcome the effects of adverse thermal environment during the
experimental period. This is probably one of the main limitations of
laboratory study on productivity. However, the lower motivation to
do work, higher perceived workload, and more negative emotions
might indicate that productivity will be reduced in an uncomfortable
environment in real life. Moreover, when subjective assessments are
taken together with performance measures, whether dissociating or
not, reveals more about task performance than performance
measures taken in isolation.
The HRV index has been shown to be closely related to thermal
comfort sensations; the LF/HF was higher when the participants
stayed in the unpleasant thermal environment (Yao et al., 2009). In
this study, when stayed in the warm environment, participants’LF/
HF was twice as much as that when they were neutral or slightly
cool, which may suggest that they felt discomfort at the warm
environment. However, it is not clear why no increase of LF/HF value
was found in the slightly cool condition. Just like Yao et al.’s (2009)
research, the present study also found an increase of
d
-band EEG
and decrease of
b
-band EEG in the neutral environment. The
d
-band
EEG is usually associated with slow-wave sleep. This result may
indicate that participants were tired after 2 h work and they could
have a good relax in the comfortable but not in the uncomfortable
environment. One limitation of this study is that only limited
samples (3 participants) of physiological measurements were made
after the tasks; it would be better if they were made along with tasks
at the same time. However, the results still support useful evidence
together subjective assessments that thermal discomfort has
negative influence on occupants’productivity.
5. Conclusions
In the field laboratory experiment subjective rating scales and
physiological measurements were made to evaluate the effects of
air temperature on productivity. It can be concluded that:
Table 5
Correlations among emotion, well-being, motivation, and performance.
TMD WB MT Accuracy Speed
OW 0.24** 0.25** 0.12** 0.19** 0.14**
TMD 0.48** 0.27** 0.02 0.01
WB 0.64** 0.03 0.01
MT 0.06 0.10
Accuracy 0.01
*
P<0.05;
**
P<0.01.
Fig. 8. Ratio of low frequency to high frequency (LF/HF) under different air
temperatures.
Fig. 9. Effect of air temperature on EEG power of different bands.
L. Lan et al. / Applied Ergonomics 42 (2010) 29e36 35

Participants had lower motivation to do work and their
d
-band
EEG decreased in the moderately uncomfortable environment.
Warm discomfort increased the LF/HF value and negatively
affected participants’well-being.
The workload imposed by neurobehavioral tests increased in
the moderately uncomfortable environment and participants
had to exert more effort to maintain their performance with
the increase of workload.
Thermal discomfort caused by high or low air temperature had
negative influence on office workers’productivity and the
subjective rating scales were important supplements of neu-
robehavioral performance measures when evaluating the
effects of IEQ on productivity.
In the future experiments, more participants will be involved,
longer exposure session (for example, at least 4 h) should be
investigated, and the extra bonus may not be supplied anymore to
make the experiment environment more like the actual work
environment. An effort should be also made to integrate the two
indices of accuracy and speed into one index, as productivity
concerns not only the quantity of work, but also the qualityof work.
Acknowledgements
The project was financially supported by the National Natural
Science Foundation of China (No. 50878125) and the Major
Program of National Natural Science Foundation of China (No.
50838009). The authors would like to thank the participants who
volunteered for this study. The authors also would like to thank the
anonymous reviewers for their useful and valuable comments on
this paper.
References
ASHRAE, 2001. ASHRAE Handbook: Fundamentals. American Society of Heating,
Refrigerating and Air-Conditioning Engineers, Inc., Atlanta, GA.
Berglund, L.G., Cain, W., 1989. Perceived air quality and the thermal environment.
In: Proceedings of Indoor Air Conference, San Diego, pp. 93e99.
Clements-Croome, D.J., Kaluarachchi, Y., 2001. Assessment and measurement of
productivity. In: Clements-Croome, D.J. (Ed.), Creating the Productive Work-
place. Taylor & Francis, London.
Damasio, A., 1994. Descartes’Error: Emotion, Reason, and the Human Brain. Gros-
set/Putnam, New York.
Fang, L., Wyon, D.P., Clausen, G., Fanger, P.O., 2004. Impact of indoor air temperature
and humidity in an office on perceived air quality, SBS symptoms and perfor-
mance. Indoor Air 17 (Suppl. 7), 74e81.
Frijda, N.H., 1996. Passions: emotion and socially consequent behavior. In:
Kavanaugh, R.D., Zimmerber, B., Fein, S. (Eds.), Emotion: Interdisciplinary
Perspectives. Lawrence Erlbaum Associates, Mahwah, New Jersey, pp. 1e27.
Haneda, M., Wargocki, P., Dalewski, M., Tanabe, S., 2009. The effects of thermal
discomfort on task performance, fatigue and mental work load examined in
a subjective experiment. In: Proceedings of Healthy Buildings 2009, Syracuse,
New York, paper 233.
Hart, S.G., Staveland, L.E., 1988. Development of NASA-TLX (Task Load Index):
results of empirical and theoretical research. In: Hancock, P.A., Meshkati, N.
(Eds.), Human Mental Workload. North Holland Press, Amsterdam, pp.
239e250.
Heerwagen, J.H., 1998. Productivity and Well-Being: What are the Links? American
Institute of Architects Conference on Highly Effective Facilities, Cincinnati, OH.
12e14 March.
Hermans, V., Peteghem, J.V., 2006. The relation between OSH and ergonomics:
a‘mother-daughter’or ‘sister-sister’relation? Appl. Ergon. 37, 451e459.
Hockey, G.R.J., 1986. A state control theory of adaptation and individual differ-
ences in stress management. In: Hockey, G.R.J., Gaillard, A.W.K., Coles, M.G.H.
(Eds.), Energetics and Human Information Processing. Martin Nijhoff,
Dordrecht, pp. 285e298.
Hocking, C., Silberstein, R.B., Lau, W.M., Stough, C., Roberts, W., 2001. Evaluation of
cognitive performance in the heat by functional brain imaging and psycho-
metric testing. Comp. Biochem. Physiol. 128 (Part A), 719e734.
Jaffe, R.S., Fung, D.L., Behrman, K.H., 1993. Optimal frequency ranges for extracting
information on autonomic activity from heart rate spectrogram. J. Auton. Nerv.
Syst. 46, 37e46.
Johnson, A.K., Anderson, E.A., 1990. Stress and arousal. In: Cacioppo, J.T.,
Tassinary, L.G. (Eds.), Principles of Psychophysiology. Cambridge University
Press, Cambridge, pp. 216e252.
Lan, L., Lian, Z.W., Pan, L., Ye, Q., 2009a. Neurobehavioral approach for evaluation of
office workers’productivity: the effects of room temperature. Build. Environ. 44
(8), 1578e1588.
Lan, L., Lian, Z.W., 2009b. Use of neurobehavioral tests to evaluate the effects of
indoor environment quality on productivity. Build. Environ. 44 (11), 2208e2217.
Lorsch, H.G., Ossama, A.A., 1994. The impact of the building indoor environment on
occupant productivity-part 1: recent studies, measures, and costs. ASHRAE
Trans. 100 (2), 741e749.
McNair, D.M., Lorr, M., Droppleman, L.F., 1992. Revised Manual for the Profile of
Mood States. Educational and Industrial Testing Services, San Diego, CA.
Parsons, K.C., 2000. Environmental ergonomics: a review of principles, methods
and models. Appl. Ergon. 31, 581e594.
Roelofsen, P., 2002. The impact of office environments on employee performance:
the design of the workplace as a strategy for productivity enhancement. J. Faci.
Manage. 1 (3), 247e264.
Sayers, B.M., 1973. Analysis of heart rate variability. Ergonomics 16, 17e32.
Woods, J.E., 1989. Cost avoidance and productivity in owning and operating
buildings. Occup. Med. 4 (4), 753e770.
Yao, Y., Lian, Z.W., Liu, W.W., Jiang, C.X., Liu, Y.M., 2009. Heart rate variation and
electroencephalograph-the potential physiological factors for thermal comfort
study. Indoor Air 19, 93e101.
Zhu, Y., 2000. Experimental Psychology. Peking University, Beijing.
L. Lan et al. / Applied Ergonomics 42 (2010) 29e3636