Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

Building and Environment 186 (2020) 107303

Contents lists available at ScienceDirect

Building and Environment


journal homepage: http://www.elsevier.com/locate/buildenv

Reliability of human environmental “sensors”: Evidence from first- and


third-person methods
Abdul-Manan Sadick a, *, Astrid Roetzel a, Mark DeKay b, Akari Nakai Kidd a, Vanessa Whittem a
a
School of Architecture and Built Environment, Faculty of Science, Engineering and Built Environment, Deakin University, Geelong, Australia
b
College of Architecture + Design, University of Tennessee, Knoxville, TN, USA

A R T I C L E I N F O A B S T R A C T

Keywords: The aim of this research was to assess the reliability of occupants’ verbalised thoughts and experiences of
Human sensors environmental conditions for objective environmental condition assessment. This research was based on the
Think-aloud method Integral Sustainable Design approach and included concurrent think-aloud method and environmental mea­
Occupant behaviour
surements at participants indicated most intense temperature, air movement, sound, lighting and preferred place
Integral methodological pluralism
Mixed methods
to study (PPS) locations. The results showed that parameters of the most intense environmental places were
Indoor environmental measurements significantly higher or lower than PPS. Participants’ experiences of the most intense environmental places as
intensely high and low, compared to PPS, was generally mirrored by differences in measured environmental
parameters between the two sets of places. Findings show that, in this case, human sensors of environmental
conditions are reliable. The think-aloud method revealed that participants considered multiple environmental
parameters of PPS concurrently and made trade-offs that prioritised the studying task. Future studies might
consider integrating first-person reporting with building operation and maintenance functions to improve effi­
ciency without compromising occupants’ preferences, satisfaction and comfort.

1. Introduction moved to new offices with superior environmental conditions. Noris


et al. [10] found that apartment energy retrofits increased occupants’
External stimuli such as light and heat are sensed by human organs perception of indoor air quality. However, other studies disagree
and information is transferred to the central nervous system for inter­ [11–13]. Smith et al. [13] found that introducing indoor plants did not
pretation and feedback [1]. The skin, for example, has thermoreceptors objectively affect temperature, relative humidity (RH) or background
that sense stimuli such as heat and send signals to the brain for inter­ noise; however, occupants perceived temperature and RH had changed.
pretation and feedback (such as cold-seeking behaviour to restore core Nicol et al. [12] found that office workers were satisfied with lighting,
body temperature) [1,2]. The sensory organs and their sensors are although desktop illuminance was below the recommended threshold.
essential in understanding human interactions with indoor environ­ Questionnaire surveys are the predominant data collection method used
ments. ASHRAE Standard 55 (Std55) [3] for thermal comfort recom­ [14–21]. Such surveys represent a third-person view of first-person ex­
mends measuring indoor air temperature at the ankle, chest and head periences of human concepts like satisfaction and comfort and involve
levels to cover the whole body. Indoor air pollutant concentration and translation of interior qualitative experiences into quantitative external
acoustic parameters are measured in the head region relative to nose [4] data. Surveys are associated with acquiescence bias, the tendency to
and ear [5,6] location, while lighting is measured on a working plane provide positive feedback, irrespective of context [13]. Therefore, these
relative to view from the eyes [7] or vertically for glare and melanopic contrasting findings may be due to questionnaire weaknesses and not
(circadian stimulus) light [8]. the reliability of participants. This literature gap has received limited
Studies of occupants’ interaction with indoor environments are attention.
extensive in the IEQ (Indoor Environmental Quality) literature. A sig­ Occupant adaptations by migration, clothing, personal/local me­
nificant proportion of them support the thesis that improving IEQ en­ chanical controls, and interacting with the building controls have also
hances occupants’ satisfaction. Schreuder et al. [9] reported significant been studied. For example, Liu et al. [22] studied non-air-conditioned
improvement in IEQ satisfaction and well-being of office workers who office buildings and found that occupants opened windows for cooling

* Corresponding author.
E-mail address: s.sadick@deakin.edu.au (A.-M. Sadick).

https://doi.org/10.1016/j.buildenv.2020.107303
Received 15 June 2020; Received in revised form 11 September 2020; Accepted 13 September 2020
Available online 17 September 2020
0360-1323/© 2020 Elsevier Ltd. All rights reserved.
A.-M. Sadick et al. Building and Environment 186 (2020) 107303

or fresh air depending on space orientation and outdoor thermal con­ Design (ISD) approach by DeKay [38] that applies Wilber [39,40] In­
ditions, while blinds were either used to avoid glare or enhance visual tegral Theory of “all-quadrants, all-levels” (AQAL) to sustainable
privacy. Additionally, Jowkar et al. [23] investigated university lecture building design. The AQAL framework advocates using a comprehensive
halls and found that students preferred to adjust their clothing in approach in addressing any situation by considering fundamental
response to uncomfortably warm or cold temperatures. Furthermore, knowledge domains, worldviews, human personality types (e.g.
Soebarto and Bennetts [24] reported that middle-income households masculine and feminine), states of consciousness, and individuals’
preferred adaptation strategies like opening windows or closing curtains developmental levels within specific human intelligences (e.g. moral
to air-conditioners due to the cost implications of using the latter. These and emotion) [36]. Wilber’s application of Integral Theory to research
studies demonstrate the significant role of occupant adaptation in methodologies, Integral Methodological Pluralism (IMP) relates indi­
attaining comfort and highlight the economic benefits of passive vidual and collective phenomena to objective and subjective knowledge
strategies. with research methodologies relevant to each quadrant (i.e. perspec­
Demand-driven control (DDC) of indoor environments relies on oc­ tives) as shown in Fig. 1 [39]. IMP advocates for investigating a problem
cupancy information to initiate actions like adjusting cooling or lighting from different epistemological perspectives, each with its own method,
[25–27], leading to using CO2 sensors to estimate space occupancy. CO2 to derive or discover multiple outcomes that inform the understanding
sensors [28] and a network of low power cameras [29] for DDC space of any occurrences in a more comprehensive or holistic way.
conditioning resulted in 55.8% and 20% energy savings, respectively. This research is framed using an abbreviated subset of IMP, relating
These findings demonstrate the potential of DDC for decreasing oper­ different research methodologies to the four foundational integral
ating energy use. DDC generally does not account for occupants’ pref­ quadrants. Each quadrant can be viewed from within its own interior or
erences in real-time; therefore, energy efficiency may be enhanced to the from an exterior perspective on its knowledge domain. This gives eight
detriment of occupants’ satisfaction or comfort. Klein et al. [30] incor­ ‘primordial paradigms’, each with its own methods class. Fig. 1 maps the
porated occupants’ thermal preferences into a multi-agent system to selected research questions and methodologies for this project into
model building energy and thermal comfort. Energy use was reduced Wilber’s 8-zone IMP framework. It is evident that they are related to the
12% while thermal comfort improved 5%. Although the model included exterior perspectives. Potential interior perspectives are listed to
occupants’ preferences, it is also susceptible to acquiescence bias demonstrate full application of the IMP. However due to time and
through its online survey. budget constraints, they were not part of this project [36].
An alternative, the “think-aloud method” (mainly used in cognitive To distinguish our methods from the more complex IMP, and from
psychology) is a qualitative research approach in which participants the less specific ‘mixed methods’ terminology, we call our efforts an
verbally express thoughts while completing a task [31,32]. Concurrent Integral Research Approach (IRA). Although not as comprehensive as an
think-aloud statements reduce the risk of memory failure that may occur approach using all eight categories, the selected four methodologies
with retrospective verbal data collection [33]. Appropriately designed allow for ‘integral triangulation’ of different data and more holistic re­
tasks will stimulate cognitive processes and allow researchers to collect sults, as compared to single methodologies [36]. Fig. 2 presents the
context-rich data in real-time. Using a think-aloud approach, Darker and perspectives, research questions, data collection and analysis methods
French [34] found that participants who verbalised comprehension implemented in the parent project [36]. The study design and research
problems about questionnaire items selected the middle option on the methodology are described in depth in a previous publication [37]. This
response scale as a strategy to avoid the cognitive workload of pro­ article is limited to two perspectives given that Hypothesis 1 is related to
cessing questions. “Upper right, Behaviour” and Hypothesis 2, to “Upper left, Experience”.
Building occupants’ perception of environmental conditions is con­
strained by worldview and cultural background influences and the 2.1. Study building and spaces
translation into emotion [35,36]. Think-aloud methods can capture
these influences to the extent that occupants’ express them during The communal spaces on level 2 of the main building at Deakin
acquisition. However, the method has received limited attention in University Waterfront Campus, Geelong, Australia, served as a test
environmental quality research. The aim of this research was to assess space. The study area (Fig. 3) consists of a sequence of spaces differing in
the reliability of occupants’ verbalised thoughts on environmental spatial and environmental features, comprising a Student reception
conditions, informed by sensory organs, for objective evaluation of area, two atriums, corridors, transition areas, study nooks, sitting areas,
environmental conditions. The methods used concurrent environmental a Kitchenette and a courtyard. The space allowed participants to expe­
measurements and think-aloud recordings at occupants’ PPS and at their rience indoor-outdoor transitions, different view qualities and quanti­
selected “most intense” temperature, light, sound and air movement ties, varied natural and artificial lighting distributions, small and large
locations. Assuming that environmental parameters of the most intense exposed spaces and contrasting noise levels and types. Indoor spaces
environmental places are either significantly high or low, relative to were mechanically conditioned; participants had no access to condi­
PPS, we proposed the following hypotheses: tioning controls. Indoor-outdoor transitions were mainly through
occupancy-controlled doors. The spaces had no adjustable shading sys­
Hypothesis 1. Measured environmental parameters at most intense
tems; lighting was automatically controlled and was in use all day.
environmental places, as identified by participants, are substantially
Fig. 4 shows photos of the six main spaces (view directions in Fig. 3).
higher or lower than at PPS.
The “Green alcove” is separated from Atrium 2 by a glass wall and has
Hypothesis 2. Participants’ reported environmental conditions expe­ built-in, cushioned seating, green wall patterns and views into Atrium 2.
rience at most intense environmental places are substantially different The Outdoor courtyard has seating underneath a balcony along the
from their reported experience at PPS and correlates with measured northern and eastern edges. Tables and benches are metal-framed,
environmental parameters being substantially different between the two timber finished. Atrium 1 has seating along the atrium wall. It offers
sets of locations. visual connection to the waterfront and roof daylighting. Atrium 2,
interconnected with the Kitchenette, offers direct views into the court­
2. Methods yard and roof daylighting; tables there are metal-framed with white
tabletops and plastic chairs. Both Atrium 1 and Atrium 2 extend from the
This article is part of a larger research project that investigates the main floor (level 2) through level 4 to the glazed roof. The Kitchenette
hypothesis that spatial–temporal patterns in buildings can serve as an has a large seating area on the eastern side. The furniture is cushioned
investigating frame among diverse research perspectives on inhabitation couches and white-topped tables. It receives daylight from two eastern
[37]. The larger research was framed based on the Integral Sustainable windows. The Student hub has a lower ceiling than other spaces and

2
A.-M. Sadick et al. Building and Environment 186 (2020) 107303

Fig. 1. IMP applied to human inhabitation – project research questions with exterior and interior methodologies for each quadrant [36].

little natural light. It is crowded with student service desks, hot desks, about locations where they perceived the most intense temperature, air
casual seating areas, multimedia screens and a large-screen television movement, light, and sound, and about their PPS. Each of these chal­
area. lenges was conducted in sequence; each participant made five complete
rounds through the study area. Table 1 summarizes PPS and most
2.2. Data collection intense environmental places that were selected by participants during
the walk through (see Appendix 1 for maps of selected locations).
Think-aloud activity (Experiences perspective) and environmental Environmental parameters were measured while participants were
measurements (Behaviours perspective) were conducted in March 2018. thinking aloud at each place. Air temperature, globe temperature and air
Participants were recruited by posters, social media sites and word-of- speed were measured using multi-device portable “comfort carts” with
mouth and received a $50 gift voucher. A total of 20 respondents, 12 sensors at 0.1 m, 0.6 m, 1.1 m and 1.7 m heights, while relative humidity
women and 8 men, aged 20–57 volunteered. Participants differed in was measured at 0.6 m height [3]. Illuminance was measured with a
their birth country (South Asia = 3, South East Asia = 1, Australia = 11, handheld lux meter at 1.1 m above floor level. Live noise and back­
China = 4, South America = 1), but education level was relatively ho­ ground noise were measured with a class 2 sound level meter but not
mogenous (postgraduate students = 13, undergraduate students = 7). simultaneously with the other environmental parameters. Live noise
Out of 20 participants, 19 were enrolled in the School of Architecture was measured on other representative days during regular university
and 1 in the School of Psychology. Participants attended a 1-h training hours and followed the measurement time sequence from the previous
session prior to the data collection. Participants are represented in this field measurement. However, the background noise was measured on a
research as A01–A15, A17, A19, A21, A22, and P01. weekend when the spaces were unoccupied.
Participants walked through the test space, determining their own
route, with a wearable camera on their head that recorded audio and
video. They were instructed to verbalise their thoughts and experiences

3
A.-M. Sadick et al. Building and Environment 186 (2020) 107303

miT locations and PPS were the most distinct as shown in Fig. 5. Relative
humidity at miT locations and PPS were similar as shown in Fig. 5.
Table 2 presents the results of paired sample t-tests that assessed dif­
ferences between corresponding values of OT, MRT, AV and RH at miT
locations and PPS. The positive mean difference in OTs, MRTs, and AVs
shows that these parameters were generally higher in the miT places,
compared to the PPS.
The mean differences (Table 2) in OTs, MRTs, and AVs between miT
and PPS were statistically significant. Consequently, these differences
are likely to be observed between the populations of miT locations and
PPS. The 95% confidence intervals (Table 2) are estimated ranges of
likely differences between the populations of miT places and PPS.
Therefore, based on the sampled locations, the mean OT difference be­
tween the populations of miT places and PPS is unlikely to exceed 1.2 ◦ C.
The mean difference in RH between the two sets of locations was not
statistically significant, therefore, the actual population difference is
unlikely to fall within the 95% confidence interval.
Spatial and temporal patterns of significantly different parameters
are shown in Fig. 6. For miT locations, MRT consistently exceeded OT
for outdoor and indoor locations, except A17; the differences were
significantly higher for outdoor places. High outdoor MRTs for miT lo­
cations is attributable to direct sun. The Outdoor courtyard floor (con­
crete and red bricks) has a solar reflectance ranged from 0.30 to 0.040
[44], contributing to higher outdoor MRTs. The difference between MRT
and OT of PPS both indoors and outdoors were either minimal or
negligible, excepting indoor location A08. Outdoor AVs were higher for
Fig. 2. The four perspectives of an Integral Research Approach to building
miT places than wind-shielded PPS. No participants indicated their miT
inhabitation, mapping the different data collection and analysis methods to
places as PPS. Only A06 selected outdoor locations for both PPS and
provide a complex, holistic understanding for the present study [36].
miT; however, the outdoor PPS was shaded.
Fig. 7 (A1 and B1) shows the predicted mean vote (PMV [3]) of indoor
2.3. Data analysis
miT places with respect to Std55 [3] and corresponding OTs and
clothing insulation (Clo). Clo was the only comfort variable that par­
The measured thermal parameters allowed for calculating mean
ticipants controlled during the study. The Figure shows that 10 of 12
radiant temperature (MRT), OT and predicted mean vote for each
indoor miT locations were not Std55 compliant. The OTs of
participant based on ASHRAE 55–2017 using the CBE Thermal Comfort
non-compliant locations were generally higher than for compliant lo­
Tool [41]. Thermal comfort of outdoor places in the courtyard was
cations. Therefore, the combination of higher Clo and OT of
calculated using the Universal Thermal Climate Index (UTCI) [42].
non-compliant locations negatively impacted thermal comfort.
Environmental parameters for all participants were not measured at the
Although the PMV for A01 (Fig. 7 A1) was within the comfort zone, it
same time; however, measurements of most intense places and PPS were
was not Std55 compliant due to elevated AV of 0.26 m/s (Fig. 6A). Fig. 7
matched for each participant. Differences between most intense places
(A2 and B2) shows PMV of indoor PPS and corresponding OTs and Clo.
and PPS were assessed graphically and quantitatively using paired
The Figure shows that 8 of 15 indoor PPS were Std55 compliant. For the
sample t-test. A two-tailed test was conducted to account for the likeli­
seven non-compliant places, Fig. 7 (B2) shows that the combined effect
hood of differences between most intense environmental locations and
of relatively higher OTs and Clo negatively impacted thermal comfort.
PPS being either positive or negative. Data points that were outside the
The thermal comfort analysis further supports observed differences be­
interquartile range (i.e. difference between 75th and 25th percentile) by
tween parameters of miT locations and PPS. MiT places were less
a magnitude of 1.5 were excluded as outliers prior to conducting the
comfortable, compared to PPS, due to the relatively high OTs of the
t-tests.
former, aided by Clo.
Transcribed think-aloud audio recordings were analysed in accor­
The Universal Thermal Climate Index (UTCI) was calculated [42] to
dance with Braun and Clarke’s phases of thematic analysis [43]. Data
assess thermal comfort of outdoor locations for both the miT and PPS.
fragments capturing participants’ thoughts and experience of most
UTCIs of miT locations and PPS for all participants were in the “no
intense places and PPS were coded and integrated to form themes, and
thermal stress” zone, except A10 for miT that was in the “moderate heat
key codes capturing differences (or similarities) in participants’ expe­
stress” zone. The highest outdoor MRT was recorded for A10 (32.43 ◦ C)
riences between most intense places and PPS were identified for com­
for miT places (Fig. 6A).
parison with environmental measurements. These themes explained
differences in participants’ thoughts and experiences.
3.2. Most intense temperature – think-aloud

3. Results
A total of 11/20 participants described their miT (most intense
temperature) place as warmer than their PPS while seven described it as
3.1. Most intense temperature – environmental measurements
cooler. This pattern matched with environmental measurements
(Fig. 7). Fig. 8 presents representative comments about experiences of
Fig. 5 shows the pattern of operative temperature (OT), MRT, air
miT places and PPS. The question language was ambiguous, as “most
velocity (AV) and RH for most intense temperature (miT) locations and
intense” could mean a hot or cool place. Participants generally felt that
PPS. OTs were slightly higher at miT locations than PPS. The impact of
miT places were warmer or cooler than PPS. For A05, Fig. 8 shows that
MRT on occupants depends on their location relative to surrounding
the relatively high MRT (mean radiant temperature) and low AV created
surfaces. Differences in MRTs between miT locations and PPS were
the “slow-cooker” sensation at the miT place (indoor) and comfort
slightly higher than OT differences (see Fig. 5). AV differences between
sensation at the preferred place to study (outdoor). A07 described the

4
A.-M. Sadick et al. Building and Environment 186 (2020) 107303

Fig. 3. Floor plan of study area colour-coded by different architectural, spatial and environmental condition. (For interpretation of the references to colour in this
figure legend, the reader is referred to the Web version of this article.)

miT location as “warm” and PPS as “comfortable temperature”. Both (Fig. 8) for several participants. A07 liked natural light from the window
locations were indoors and the miT location was Std55 compliant, while while talking about the “comfortable temperature” at indoor PPS. A05
the PPS was not due to elevated AV. This suggests a preference for the liked the “cool breeze”, “comfortable outdoor temperature” and the
cooling sensation of slightly elevated AV. experience that it was “pretty quiet too …” while, A13 preferred an
There was limited reference to humidity, as shown in Fig. 8, except indoor space to study because they dislike “to study under sunlight” and
A05, who described the miT location as “dry heat”, and P01, who felt “not a lot of people walking past”. This shows that the impact of other
“too warm in a bit of …, sticky sort of”. This agrees with the insignificant environmental factors like lighting and acoustics was concurrently
difference found between RH of intense temperature and PPS (Fig. 5 and considered by participants while thinking about temperature of PPS. It
Table 2). Outdoor-indoor transition was referenced by A02 (Fig. 9), A08, also demonstrates that the task (studying) influenced the environmental
A01 and A15 as the cause of intense temperature feeling. A08 said “… I parameters, which were considered important but that were not tem­
was outside not too long ago and then you step in immediately. So perature-related.
there’s that huge temperature difference”. The significant difference in
thermal parameters between outdoor and indoor miT locations shown in 3.3. Most intense air movement – environmental measurements
Fig. 6A correlates to this sensation.
The impact of solar radiation on MRT was identified by A02, A09, Fig. 9 depicts the spatial and temporal patterns of indoor and outdoor
A21 and A10 as a cause of miT feeling outdoors. Participant A21 said “So AV for the places of miA (most intense air movement) and the PPS. An
all the wind surrounded by … and here I’m standing, I feel actually approximate qualitative interpretation of AV is < 0.05 m/s – (not
warm. So it’s bit weird, weird sensation”. A21 expected AV of 0.86 m/s perceptible), ≥ 0.1 m/s (perceptible/comfortable) ≥ 0.2 m/s (elevated
to nullify the warming effect of 29.0 ◦ C MRT. The impact of solar ra­ breeze) [3,45]. AVs were higher atmiA places (mean = 0.5 m/s) than
diation on MRT was apparent between outdoor and indoor locations as PPS (mean = 0.2 m/s) for indoors and outdoors respectively. No outdoor
shown in Fig. 6A. Additionally, outdoor PPS had lower MRTs compared miA locations were under wind/sun shade except A13 and A15. The
to outdoor miT locations due to shading of the former (Fig. 6). Partici­ highest miA was an indoor corridor, (0.9 m/s), where A07 was directly
pant A02 referenced this impact of shading on outdoor PPS (Fig. 8). under an air-conditioning diffuser. Two locations, A02 and P01 within
Cross-factor consideration, where a non-related environmental the same corridor were significantly lower (see Fig. 9). The majority of
parameter is referenced, was a key theme related to temperature of PPS miA indoor locations were close to doors and therefore influenced by

5
A.-M. Sadick et al. Building and Environment 186 (2020) 107303

Fig. 4. The six colour-coded spaces on floor plan presented in Fig. 3, visually illustrating the different spatial and environmental conditions. (For interpretation of the
references to colour in this figure legend, the reader is referred to the Web version of this article.)

0.3 (95% CI, 0.2 to 0.5) m/s, t(19) = 4.28, p < 0.0005.
Table 1
Distribution of PPS and most intense environmental locations relative to test
spaces. 3.4. Most intense air movement – think-aloud
Major spaces Most intense Preferred place
to study Fig. 10 presents representative comments and measured AVs of most
temperature air sound light
movement intense air movement (miA) and PPS places. Participants associated
intense air movement with high velocity, except A12 who associated
Student hub 5 3 8 4 0
Atrium 1 1 0 0 1 4
intensity with low velocity. A12 had the lowest AV compared to other
Atrium 2 1 0 11 1 1 miA locations (Fig. 10). Participants A05, A07, A03, A06 and A21
Green alcove 2 3 0 4 7 associated the high AVs of miA with a cooling sensation, which is logical
Kitchenette 3 3 0 1 3 given that mean AV of miA (0.5 m/s) was significantly higher than most
Outdoor 8 11 1 9 5
intense temperature (miT) (0.3 m/s) locations.
courtyard
Air movement transition was identified as the cause of intense air
movement feeling. A09 identified the funnelling effect of narrow door­
outdoor air or frequent opening of doors. Indoor PPS locations were free ways/corridors, while A02, A13 and A14 identified indoor-outdoor
of the door opening influence. As shown in Fig. 9, the highest PPS air movement as the cause of intense air movement feeling, as shown below.
velocity was outdoors, A06 (0.6 m/s). Paired sample t-test confirmed
I was feeling it, uh, right in the doorway before, where it was nar­
that AVs of miA were significantly higher than PPS, mean difference of
rowest (A09)

6
A.-M. Sadick et al. Building and Environment 186 (2020) 107303

Fig. 5. Pattern of operative temperature, mean radiant temperature, air velocity and relative humidity for miT places and PPS.

limit of background noise in study spaces is 35 ± 2 dBA [5]. Live noise is


Table 2
the noise generated due to occupants and other activities with a space. A
Paired sample t-test for most intense temperature (miT) and preferred places to
typical subjective live noise scale is 10 dBA (extremely quiet), 30 dBA
study (PPS).
(very quiet), 50 dBA (quiet) and 70 dBA (irritating) [46]. Average
Paired Mean Mean 95% t df Significance
background noise was 46 dBA and 42 dBA for miS and PPS respectively.
sample difference confidence (2-tailed)
(miT – PPS) interval
Additionally, average live noise was 60 dBA and 55 dBA for miS and PPS
respectively. Fig. 11 shows that background and live noise pattern for
Operative temperature (oC)
miS was less varied, as compared to PPS. Excepting A08 in the Outdoor
miT 23.7 0.7 0.2–1.2 3.14 19 0.005**
PPS 23.0 courtyard, miS places were concentrated in the Student hub and Atrium
Mean radiant Temperature (oC) 2, while PPS were spread across Atrium 1, the Green alcove, Kitchenette,
miT 25.5 2.0 0.7–3.3 3.31 19 0.004** Atrium 2 and Outdoor courtyard (shaded section). The highest miS live
PPS 23.4
noises were at Atrium 2 (Fig. 11A participants A02, P01, A04 and A06)
Air velocity (m/s)
miT 0.4 0.2 0.0–0.3 2.53 19 0.021*
that is openly connected to the cafe in the basement. The Green alcove
PPS 0.2 had the lowest background and live noise among PPS (Fig. 11B).
Relative Humidity (%) Therefore, the glass wall attenuated live and background noise from
miT 43.5 − 0.5 − 0.7–0.6 − 0.26 17 0.799 Atrium 2. Outdoor live and background noise for PPS were very close
PPS 44.0
(Fig. 11B); the low frequency HVAC equipment noise observed during
measurement masked live noise, making it less distracting.
I was from a space that is with no wind, like confined space. And then Results of a paired sample t-test supported the observed significant
suddenly, open the door, and there’s wind coming in. (A13) difference in both background and live noise between miS and PPS, miS
being higher. The mean difference of 4 (95% CI, 0 to 9) dBA for back­
Participant A13 experienced the highest miA on day one of mea­ ground noise was significant, t(19) = 2.10, p = 0.040. Additionally, the
surements (Fig. 9A) and preferred a place with much lower AV for mean difference of 5 (95% CI, 3 to 7) dBA for live noise was significant,
studying. A02, however, experienced the lowest miA on day two and t(19) = 5.81, p < 0.0005. These results show that differences in back­
chose a PPS with relatively higher AV. ground noise and live noise between the populations miS and PPS will
Cross-factor consideration is apparent in the comments of A04 and likely fall within the estimated 95% confidence intervals.
A05 (Fig. 10) about AV at their PPS. Both participants liked the cool
breeze and the fact that their PPS was quiet enough for effective
3.6. Most intense sound – think-aloud
studying. A04 liked the breeze of both miA and PPS but preferred the
lower AV location for studying. For A04 and A05, the depleting effect of
Fig. 12 present representative comments on participants’ acoustic
distracting noise on directed attention was equally essential to experi­
experience and corresponding background and live noise measurements
encing a cool breeze at PPS.
for most intense sound (miS) and PPS. All 20 participants said their miS
place was louder than PPS and none selected either the Student hub or
3.5. Most intense sound – environmental measurements Atrium 2 as PPS, except A07 who selected Atrium 2 as PPS. A07 indi­
cated a location away from the noise epicentre of Atrium 2 (Fig. 12). The
Fig. 11 shows background and live noise for places of most intense comments in Fig. 12 reflect the pattern of live and background noise for
sound (miS) and PPS. Background noise in indoor spaces is the noise miS and PPS shown in Fig. 11.
from building services and outdoor noise sources [5]. The recommended Participants identified several noise sources, including unintelligible

7
A.-M. Sadick et al. Building and Environment 186 (2020) 107303

Fig. 6. Spatial and temporal pattern of operative, temperature, mean radiant temperature and air velocity for most intense temperature and PPS.

Fig. 7. Interplay between Clo and OT in relation to PMV of miT indoor places and PPS indoors.

(e.g. doors, footsteps, birds) and intelligible sound (e.g. people talking, Yeah, very echoey but also, like, very loud. (A05)
music) at miS places. Footsteps heard at the Student hub were due to the
Lots of different areas and some bouncing off each other a bit. I don’t
polished concrete floor and timber floor above, while footsteps noise in
think there’s any acoustic tiles … (A15)
Atrium 2 included the interconnected basement cafe and adjacent tim­
ber floors above. This omnidirectional noise is captured in the comments Although reverberation was not measured, the two spaces were
below. dominated by hard floor, wall and ceiling surface; hence, concern about
reverberation is logical. All the participants preferred to find a quieter
… Yes, it’s just a lot of movement of sound going on and I feel, um,
space to study; however, A03, A07, and A15 suggested the use of
my voice could get a little lost in this environment. (A07)
headphones as an adaptation strategy to study in miS places.
Ah, because here they come from upstairs and downstairs. You can Acoustic considerations featured prominently while several partici­
hear from all directions. (A14) pants were commenting about temperature and air movement at PPS,
but some participants preferred a place to study because of access to
Six participants (A03- A05, A08, A09, and A15) identified reflected
daylight and fresh air, as shown below.
sound as a key contributing factor to the sensation of loudness at the
Student hub and Atrium 2, as shown in the comments below.

8
A.-M. Sadick et al. Building and Environment 186 (2020) 107303

Fig. 8. Comparison of key comments for most intense temperature locations and PPS. OT (oC); MRT (oC); AV (m/s).

Fig. 9. Spatial-temporal pattern of air velocity for miA and PPS.

I don’t like to be around too much noise. I can’t concentrate other­ under the shaded section of the Outdoor courtyard. Therefore, the var­
wise … the fresh air’s good so that can make you feel a bit better iations in outdoor miL in the Outdoor courtyard were also significantly
(A06) influenced by cloud cover. Lighting of the lowest miL outdoor place
(A01, 188 lux) was significantly lower than the lowest outdoor PPS (770
So it’s quieter, there’s access to natural light. (A10)
lux) and is logical given that A01 was at the innermost shaded end of the
courtyard.
The outdoor PPS were all shaded; however, Fig. 13 shows that
3.7. Most intense lighting - environmental measurements lighting varied significantly between A06, A04 and A05 on one hand,
and A02 and A15 on the other. The difference was due to time of day,
Fig. 13 shows the pattern indoor and outdoor lighting for most solar altitude and azimuth. Illuminance at A05 was measured when the
intense lighting (miL) and PPS. Average lighting levels were 17,050 lux location was shaded from sunlight. As the sun moved, location A02, and
and 725 lux for miL and PPS respectively. Apart from A08 and A14 for A15 under the shade received more reflected and direct sunlight. After
miL, indoor lighting for both miL and PPS were below 400 lux. The excluding three outliers (miL A04, A11 and A10) that exceeded 1.5 box-
lowest indoor lighting was miL A12 (98 lux). Compared to indoor lengths in a boxplot, the mean difference of 5351 (95% CI, 215 to
lighting, the outdoor lighting pattern for miL and PPS varied substan­ 10,467) lux between miL and PPS was significant, t(16) = 2.21, p =
tially and was expected, given that only one (A01) of the miL places was 0.041. The large mean difference was due to the relatively high outdoor

9
A.-M. Sadick et al. Building and Environment 186 (2020) 107303

Fig. 10. Comparison of key comments for miA and PPS. AV: air velocity (m/s).

Fig. 11. Spatial and temporal pattern of background noise and live noise for miS and PPS.

illuminance of miL. places, the dislike and high sensation may have been due to the green
wall surfaces influencing lighting colour. For A05 and A15, intense
3.8. Most intense lighting – think-aloud lighting sensation was due to disliking artificial light. A05 associated
their miL place with being in a hospital and was “very comfortable” with
Fig. 14 presents representative comments and field measurements of relatively brighter PPS outdoors.
illuminance for most intense lighting (miL) and PPS. Participants A04 Participants’ perceived illuminance intensity matched Fig. 13 data.
and A07, whose miL places were outdoors, said they were significantly Participants who chose outdoor miL places felt significant difference in
brighter than their PPS. A04 liked the “really bright and nice” illumi­ illuminance relative to indoor PPS. For participants who chose indoor
nance of 96,250 lux (Outdoor courtyard) but preferred a 976 lux place to miL places, lighting sensation was influenced by other personal factors.
study. A07 felt that 11,750 lux was too intense on their eyes and
preferred an indoor place with less daylight. The difference in illumi­ 4. Discussions
nance between outdoor and indoor places significantly affected indoor-
outdoor transitions and resulted in participants’ visual discomfort, as The hypotheses were both confirmed; hence, the alternatives can be
indicated by A01 “But before my eyes adjusted … opening this door and rejected. For Hypothesis 1, differences in all measured environmental
being able to see inside … was the most intense”. A09, A05 and A15 parameters between most intense places and PPS were confirmed to be
selected indoor places for miL. For A09, intense lighting sensation was significant except for RH (relative humidity). The comparisons of think-
due to reflected light from white walls. They disliked their PPS lighting aloud results and measurements between most intense places and PPS
in the Green alcove because it was not “particularly good” and “light’s a confirm that human sensors tend to be reliable at detecting larger
lot for” them. Given that PPS illuminance was slightly less than miL parameter differences. Most indoor places for both the most intense

10
A.-M. Sadick et al. Building and Environment 186 (2020) 107303

Fig. 12. Comparison for key comments for miS and PPS. BN - background noise (dBA); LN - live noise (dBA).

Fig. 13. Spatial-temporal pattern of illuminance for miL locations and PPS.

temperature (miT) and nearly half of the PPS were not Std55 compliant, the fact that indoor PPS, miT, miA, miL and miS were not limited to one
this was reflected by some participants who preferred places for study or two places only.
because of other factors like quiet. Think-aloud data showed that RH Comments from the think-aloud data indicate that the studying task
was the least important thermal concern of participants, thus supporting influenced participants to consider the interplay between temperature,
the insignificant t-test results. This suggests that participants’ body light, sound and air movement for the PPS as demonstrated by the cross-
sensors were most sensitive to temperature and AV parameters, and less factor consideration theme. For example, some indoor PPS were ther­
sensitive to RH in the narrow range present. Most intense places and PPS mally uncomfortable but preferred by participants because they were
were located indoors and outdoors. The majority of PPS (i.e. 15), miT (i. less noisy or exhibited a preferred lighting intensity. In an earlier anal­
e. 12) and most intense lighting (miL) (i.e. 11) were indoors while the ysis of the think-aloud and semi-structured interviews for PPS only,
majority of most intense air movement (miA) places (i.e. 11) were Roetzel et al. [36] concluded that PPS were selected due to other factors
outdoors. Participants’ perceptions of conditions were not simply based like “prospect” (sense of visual overview) and “refuge” (sense of safety).
on spatial enclosure. Rather, sensory information, interpretation and The think-aloud results in this paper also demonstrate that human
feedback likely influenced participants’ perceptions of most intense assessment of environmental conditions is not based on immediate
environmental places and PPS. This conclusion is further supported by “measurements” only but are also contextualised with past experiences,

11
A.-M. Sadick et al. Building and Environment 186 (2020) 107303

Fig. 14. Comparison of key comments for miL locations and PPS. I: illuminance (lux).

as shown in the previous publication [36]. These can be very recent and 5. Conclusion
related to the same spatial context, such as indoor-outdoor transitions or
related to past personal experiences, such as the lighting invoking The two hypotheses tested and confirmed support the thesis that
memories of a hospital context. Short-term experiences occurring within human sensors of environmental conditions are reliable sources of in­
the spatial context may be accounted for by architects and building formation for objective decision making. The key conclusions are sum­
designers, whereas the diversity of individual past experiences in other marised as follows:
contexts are difficult to anticipate. The Integral Research Approach
implemented here advocates for methodological pluralism across four • All measured environmental parameters of most intense environ­
primary perspectives (Fig. 2) where contrasting methods interrogate mental places (except RH) were found to be significantly different to
subjective and objective phenomena with qualitative and quantitative PPS.
data in different ways. This resulted in extracting logical reasons for • Participants reported experiences (think-aloud results) of most
unexpected patterns that would remain a mystery in a research method intense environmental places differed from PPS for corresponding
that used only a questionnaire survey. environmental parameters. These differences generally matched
Our findings highlight the importance of occupant adaptation, as with measurements of most intense environmental places being
evidenced by study participants migrating to identify most intense higher or lower than measured parameters of PPS.
environmental places and PPS. These migrations allowed study partic­ • The quantitative precision of human sensors was not established;
ipants to sample different places and decide based on personal prefer­ however, the results suggest that they reliably detect perceptible
ences for studying. Therefore, the authors advocate for diverse changes in environmental conditions that are most likely to support
environmental conditions, especially in open-plan spaces, and more occupants’ comfort-seeking behaviours. These findings may not
opportunities for building occupants to actively interact, migrate in and apply to other environmental parameters like volatile organic com­
control indoor environmental conditions. Demand-driven control (DDC) pounds, particulate matter and reverberation that were not
of buildings relies on electronic sensors that track environmental pa­ measured in this research.
rameters and space occupancy quantitatively. Findings here show that • The think-aloud method provided context-rich data that contributed
human sensors may be reliable sources of context-rich quantitative as to explaining patterns that could not be explained with measure­
well as qualitative information for environment optimisation decision ments alone and would have remained unexplained had the research
making and should be the subject of future research. used a questionnaire.
The majority of participants were students studying architecture; • Participants considered multiple environmental parameters for the
therefore, prior knowledge of building science could have resulted in identification of PPS concurrently, contextualised sensory input with
premeditated environmental perceptions or overly considered in­ previous experiences and made trade-offs that prioritised the task of
terpretations of their experience and its causes than if the participant studying. For example, seven of the indoor PPS did not comply with
sample had been composed of a more diverse group. However, the fact Std55 but were preferred for study because they were quiet places.
that the most intense environmental locations and PPS were not simi­ • This study suggests that participant’s preferences were not only
larly distributed suggests that the participants’ thoughts and perceptions influenced by the stable conditions in a particular space but also
may not have been biased by prior knowledge. Environmental mea­ contextualised relative to conditions in other available or previously
surements and think-aloud data were collected over two days and at experienced spaces. Contrasts of indoor environmental conditions at
different times for each participant thus environmental conditions var­ the transition from one space to another were particularly noticed by
ied. Data for most intense locations and PPS were paired for each participants. The context-relevant trade-offs (example shown above)
participant; therefore, the variance analysis was relative to the condition demonstrate the potential benefits of leveraging human sensors to
that each participant experienced. The sample size in this research was collect qualitative environmental information to inform objective
adequate for the qualitative think-aloud method [47,48] but is limited decision making on DDC, building operation and maintenance.
for generalising the findings from the quantitative statistical analysis. Future studies should leverage the methodological pluralism of the
However, the findings can justify future large-scale research for IRA to explore DDC systems that integrate real-time occupants’ ex­
generalisation. periences, perhaps with artificial intelligence for natural language
processing and machine learning.

12
A.-M. Sadick et al. Building and Environment 186 (2020) 107303

• Building design should provide a wide range of comfort conditions [10] F. Noris, G. Adamkiewicz, W.W. Delp, T. Hotchi, M. Russell, B.C. Singer, M. Spears,
K. Vermeer, W.J. Fisk, Indoor environmental quality benefits of apartment energy
within which occupants can have personal control as well as the
retrofits, Building and Environment 68 (2013) 170–178.
opportunity to migrate among locations based on their preferences. [11] S. Muhič, V. Butala, The influence of indoor environment in office buildings on
• Combined in an integral research approach the different quality their occupants: expected–unexpected, Build. Environ. 39 (3) (2004) 289–296.
standards for qualitative and quantitative research methodologies [12] F. Nicol, M. Wilson, C. Chiancarella, Using field measurements of desktop
illuminance in European offices to investigate its dependence on outdoor
can provide a challenge for researchers. In this research, the think- conditions and its effect on occupant satisfaction, and the use of lights and blinds,
aloud method would not have been feasible with large participant Energy Build. 38 (7) (2006) 802–813.
numbers, however the statistical analysis of measured conditions [13] A.J. Smith, A. Fsadni, G. Holt, Indoor living plants’ effects on an office
environment, Facilities 35 (9–10) (2017) 525–542.
would have benefited from larger numbers. Further research on how [14] S. Cakmak, R.E. Dales, L. Liu, L.M. Kauri, C.L. Lemieux, C. Hebbern, J. Zhu,
to negotiate these conflicting quality standards in integral research Residential exposure to volatile organic compounds and lung function: results from
would be desirable. a population-based cross-sectional survey, Environ. Pollut. 194 (2014) 145–151.
[15] S. Altomonte, S. Schiavon, Occupant satisfaction in LEED and non-LEED certified
buildings, Build. Environ. 68 (2013) 66–76.
Funding [16] M.D. Colton, P. MacNaughton, J. Vallarino, J. Kane, M. Bennett-Fripp, J.
D. Spengler, G. Adamkiewicz, Indoor air quality in green vs conventional
multifamily low-income housing, Environ. Sci. Technol. 48 (14) (2014)
This work was supported by Deakin University in Australia under the 7833–7841.
Faculty of Science, Engineering and Built Environment Research Grant [17] V. De Giuli, R. Zecchin, L. Corain, L. Salmaso, Measurements of indoor
Scheme 2018. environmental conditions in Italian classrooms and their impact on children’s
comfort, Indoor Built Environ. 24 (5) (2014) 689–712, https://doi.org/10.1177/
1420326X14530586.
[18] Z. Gou, S.S.-Y. Lau, J. Shen, Indoor environmental satisfaction in two LEED offices
Declaration of competing interest and its implications in green interior design, Indoor Built Environ. 21 (4) (2012)
503–514.
[19] J. Kim, R. de Dear, Impact of different building ventilation modes on occupant
The authors declare that they have no known competing financial
expectations of the main IEQ factors, Build. Environ. 57 (2012) 184–193.
interests or personal relationships that could have appeared to influence [20] A. Montazami, M. Wilson, F. Nicol, Aircraft noise, overheating and poor air quality
the work reported in this paper. in classrooms in London primary schools, Build. Environ. 52 (2012) 129–141.
[21] L.T. Wong, K.W. Mui, P.S. Hui, A multivariate-logistic model for acceptance of
indoor environmental quality (IEQ) in offices, Build. Environ. 43 (1) (2008) 1–6.
Acknowledgements [22] J. Liu, R. Yao, R. McCloy, An investigation of thermal comfort adaptation
behaviour in office buildings in the UK, Indoor Built Environ. 23 (5) (2014)
675–691.
We would like to acknowledge our friend and colleague Dr Lucy [23] M. Jowkar, H.B. Rijal, J. Brusey, A. Montazami, S. Carlucci, T.C. Lansdown,
Zinkiewicz, particularly, her role in both conceptualising the project at Comfort temperature and preferred adaptive behaviour in various classroom types
its early stages as well as collecting the data. Unfortunately, Lucy passed in the UK higher learning environments, Energy Build. 211 (2020) 109814.
[24] V. Soebarto, H. Bennetts, Thermal comfort and occupant responses during summer
away suddenly in mid-2018 and therefore was unable to see the
in a low to middle income housing development in South Australia, Build. Environ.
completion of this project. We would therefore like to dedicate this 75 (2014) 19–29.
paper in her honour. [25] T. Labeodan, W. Zeiler, G. Boxem, Y. Zhao, Occupancy measurement in
We would also like to thank our colleagues AProf. Mark Luther and commercial office buildings for demand-driven control applications—a survey and
detection system evaluation, Energy Build. 93 (2015) 303–314.
Tim Clark for advice on indoor environmental measurements as well as [26] N. Li, G. Calis, B. Becerik-Gerber, Measuring and monitoring occupancy with an
Deakin student Salomé Bricker. This project is a contribution to the In­ RFID based system for demand-driven HVAC operations, Autom. ConStruct. 24
ternational Energy Agency IEA Annex 79 project ‘Occupant-Centric (2012) 89–99.
[27] W. Wang, J. Chen, Y. Lu, H.-H. Wei, Energy conservation through flexible HVAC
Building Design and Operation’. management in large spaces: an IPS-based demand-driven control (IDC) system,
Autom. ConStruct. 83 (2017) 91–107.
[28] Z. Sun, S. Wang, Z. Ma, In-situ implementation and validation of a CO2-based
Appendix A. Supplementary data
adaptive demand-controlled ventilation strategy in a multi-zone office building,
Build. Environ. 46 (1) (2011) 124–133.
Supplementary data to this article can be found online at https://doi. [29] V.L. Erickson, A.E. Cerpa, Occupancy based demand response HVAC control
org/10.1016/j.buildenv.2020.107303. strategy, ACM, in: Proceedings of the 2nd ACM Workshop on Embedded Sensing
Systems for Energy-Efficiency in Building, 2010. Zurich, Switzerland.
[30] L. Klein, J.-y. Kwak, G. Kavulya, F. Jazizadeh, B. Becerik-Gerber, P. Varakantham,
References M. Tambe, Coordinating occupant behavior for building energy and comfort
management using multi-agent systems, Autom. ConStruct. 22 (2012) 525–536.
[31] E. Charters, The use of think-aloud methods in qualitative research an introduction
[1] S. Kobayashi, Temperature receptors in cutaneous nerve endings are thermostat
to think-aloud methods, Brock Educ.: J. Educ. Res. Prac. 12 (2) (2003).
molecules that induce thermoregulatory behaviors against thermal load,
[32] D.W. Eccles, G. Arsal, The think aloud method: what is it and how do I use it? Qual.
Temperature 2 (3) (2015) 346–351.
Res. Sport Exer. Health 9 (4) (2017) 514–531.
[2] K.E. Kroemer Elbert, H.B. Kroemer, A.D. Kroemer Hoffman, Chapter 4 - how the
[33] K.A. Young, Direct from the source: the value of’think aloud’data in understanding
mind works, in: K.E. Kroemer Elbert, H.B. Kroemer, A.D. Kroemer Hoffman (Eds.),
learning, J. Educ. Enq. (2005).
Ergonomics, third ed., Academic Press, 2018, pp. 127–169.
[34] C.D. Darker, D.P. French, What sense do people make of a theory of planned
[3] ASHRAE, ASHRAE Standard 55: Thermal Environmental Conditions for Human
behaviour questionnaire? A think-aloud study, J. Health Psychol. 14 (7) (2009)
Occupancy, American Society of Heating, Refrigerating, and Air-Conditioning
861–871.
Engineers, Atlanta, GA, 2017.
[35] M. DeKay, An integral approach to architectural experience, in: 4th Integral
[4] U.S. EPA, A Standardized EPA Protocol for Characterizing Indoor Environmental
European Conference - Global Integral Awakens, 2020. Siófok, Hungary.
Quality, 2003. http://www.epa.gov/iaq/base/pdfs/2003_base_protocol.pdf.
[36] A. Roetzel, M. DeKay, A. Nakai Kidd, A. Klas, A. Sadick, V. Whittem, L. Zinkiewicz,
[5] ASA, Acoustical Performance Criteria, Design Requirements, and Guidelines for
Architectural, indoor environmental, personal and cultural influences on students’
Schools, Part 1: Permanent Schools, Acoustical Society of America, Melville, New
selection of a preferred place to study, Architect. Sci. Rev. (2019) 1–17.
York, 2010.
[37] M. DeKay, A. Roetzel, A.N. Kidd, L. Zinkiewicz, A. Klas, An Integral Sustainable
[6] ASTM, Standard Test Method for Measurement of Airborne Sound Attenuation
Design Approach to Human Inhabitation of Architectural Spaces: Theory and
between Rooms in Buildings, American Society for Testing and Materials, West
Project Design, IEC 2018: Allies of Evolution: Proceedings of the 3rd Integral
Conshohocken, PA, 2014.
European Conference, Integral Europe, 2018, pp. 1–25.
[7] G.R. Newsham, B.J. Birt, C. Arsenault, A.J. Thompson, J.A. Veitch, S. Mancini, A.
[38] M. DeKay, Integral Sustainable Design: Transformative Perspectives, Earthscan,
D. Galasiu, B.N. Gover, I.A. Macdonald, G.J. Burns, Do ‘green’buildings have better
Washington, 2012.
indoor environments? New evidence, Build. Res. Inf. 41 (4) (2013) 415–434.
[39] K. Wilber, Integral Spirituality: A Startling New Role for Religion in the Modern
[8] K. Konis, Field evaluation of the circadian stimulus potential of daylit and non-
and Postmodern World, Shambhala Publications, Boston, Massachusetts, 2007.
daylit spaces in dementia care facilities, Build. Environ. 135 (2018) 112–123.
[40] K. Wilber, The Eye of Spirit: an Integral Vision for a World Gone Slightly Mad,
[9] E. Schreuder, L. van Heel, R. Goedhart, E. Dusseldorp, J.M. Schraagen, A. Burdorf,
Shambhala Publications, 2001.
Effects of newly designed hospital buildings on staff perceptions: a pre-post study
to validate design decisions, Health Environments Research and Design Journal 8
(4) (2015) 77–97.

13
A.-M. Sadick et al. Building and Environment 186 (2020) 107303

[41] T. Hoyt, S. Schiavon, A. Piccioli, T. Cheung, D. Moon, K. Steinfeld, CBE thermal -sites-credit-heat-island-effect.pdf?sfvrsn=6&sfvrsn=6, 2007. (Accessed 24
comfort Tool. https://comfort.cbe.berkeley.edu/, 2019. (Accessed 26 August January 2020).
2019). [45] P.O. Fanger, A.K. Melikov, H. Hanzawa, J. Ring, Air turbulence and sensation of
[42] B. Wojtach, UTCI calculation, n.d, http://utci.org/utcineu/utcineu.php. (Accessed draught, Energy Build. 12 (1) (1988) 21–39.
8 September 2019). [46] J.P. Cowan, Handbook of Environmental Acoustics, John Wiley & Sons, 1993.
[43] V. Braun, V. Clarke, Using thematic analysis in psychology, Qual. Res. Psychol. 3 [47] C.J. Johnstone, N.A. Bottsford-Miller, S.J. Thompson, Using the think aloud
(2) (2006) 77–101. method (cognitive labs) to evaluate test design for students with disabilities and
[44] M.L. Marceau, M.G. VanGeem, Solar reflectance of concretes for LEED sustainable English language learners. https://files.eric.ed.gov/fulltext/ED495909.pdf, 2006.
sites credit: heat island effect. https://www.cement.org/docs/default-source/fc [48] L. Van Oort, C. Schröder, D. French, What do people think about when they answer
_concrete_technology/sn2982-solar-reflectance-of-concretes-for-leed-sustainable the Brief Illness Perception Questionnaire? A ‘think-aloud’study, Br. J. Health
Psychol. 16 (2) (2011) 231–245.

14

You might also like