Journal of Environmental Psychology 30 (2010) 533–541

Contents lists available at ScienceDirect

Journal of Environmental Psychology

journal homepage: www.elsevier.com/locate/jep

Windows, view, and office characteristics predict physical and

psychological discomfort
Myriam B.C. Aries a, b, *, Jennifer A. Veitch b, Guy. R. Newsham b
a
Eindhoven University of Technology, the Netherlands
b
National Research Council Canada, Institute for Research in Construction, Ottawa, K1A 0R6, Canada

a r t i c l e i n f o a b s t r a c t

Article history: Office employees spend a lot of time inside buildings, where the physical conditions influence their well-
Available online 6 January 2010 being and indirectly influence their employers’ business performance. With data from a field study
conducted in the Netherlands in April to May 2003, we used path analysis to further elucidate the
Keywords: relationship between personal (gender and seasonality of mood shifts), building (view type, view quality,
Windows window distance, and social density), and perceived environmental conditions (light quality, and office
View
impression) and physical and psychological discomfort, sleep quality, and environmental utility. The
Well-being
Health
results show that window views, which that are rated as being more attractive, are beneficial to building
Social density occupants by reducing discomfort. However, being close to a window and rating the lighting as being of
Comfort lower quality can result in thermal and glare problems (environmental utility). Reduced discomfort at
Discomfort work can improve sleep quality, indicating that physical conditions at work influence home life.
Offices Ó 2010 Elsevier Ltd. All rights reserved.

1. Introduction 2005). Here, we examined the effects of window view, office

Office employees spend most of their waking time inside the
buildings in which they work (Leech, Nelson, Burnett, Aaron, &
Raizenne, 2002; Schweizer et al., 2007). To the extent that envi-
ronmental conditions influence health and well-being, workplace
conditions are important contributors. The effects on individuals
can carry through to influence organizational performance, making
the work environment an indirect influence on measures of orga-
nizational productivity (Veitch, Charles, & Newsham, 2004).
Among the architectural or interior design characteristics that
define the office experience are the proximity of a window to the
individual’s desk, the characteristics of the view outside the
window, and the number of people who share the office (its social
density). Although preferences for windows are well established
(as are discussed below), and the restorative value of natural
views are well-known, we know of no investigations that have
tested an integrated model of the effects of these variables on
both immediate comfort and discomfort in the office with after-
work effects on any measure of well-being. Accordingly, this study
reports a re-analysis of a survey conducted as part of a field study
concerning architectural influences on office lighting (Aries,
* Corresponding author. Department of Architecture, Building and Planning,
Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven, The
Netherlands. Tel.: þ31 40 247 4743.
E-mail address: m.b.c.aries@tue.nl (M.B.C. Aries).
social density, and individual differences on employees’ discom-
fort at work and sleep quality at home. The analysis used Hedge’s
conceptual framework for Sick Building Syndrome [SBS] (Hedge,
Burge, Robertson, Wilson, & Harris-Bass, 1989), which includes
individual differences and architectural characteristics to deter-
mine SBS symptoms, as starting point for our model. Our model
hypothesized that certain individual and architectural factors may
directly and indirectly – through perceived environmental
conditions – lead to physical and psychological discomfort in an
office space (see Fig. 1).
The individual factors in this study were gender, age, and sea-
sonality of mood. Architectural factors included were the distance
from a window, the view type and view quality, and the social
density of the office. The perceived environmental conditions were
office impressions and lighting quality. The outcome measures
were self-reported physical and psychological discomfort, sleep
quality, and environmental utility (thermal and glare problems).
The literature review will focus on these variables.
Many investigators have examined the effects of lighting quality
in an indoor space and its effects on the work performance,
comfort, and satisfaction of occupants (e.g., Chung & Burnett, 2000;
Veitch, 2001b; Veitch & Newsham, 1998) including outcomes that
we have labelled environmental utility (see below). In addition to
these visually mediated psychological effects, light has non-visual
effects on physiology (Commission Internationale de l’Eclairage
(CIE), 2004; Veitch, 2001a). Light controls the human biological

0272-4944/$ – see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jenvp.2009.12.004
534 M.B.C. Aries et al. / Journal of Environmental Psychology 30 (2010) 533–541
Fig. 1. Basic model.
clock and is, therefore, an important regulator of the human
physiology and performance. Regular patterns of light and dark
exposure each day are necessary to regulate circadian rhythms,
including sleep-wake cycles (CIE, 2004; Veitch, 2001a). The co-
ordination of these cycles contributes to sleep quality. Sleep has
powerful restorative effects on the body and poor sleep quality has
adverse effects on physical and mental health (e.g., Haack & Mul-
lington, 2005; Kuppermann et al., 1995; Meerlo, Sgoifo, & Suchecki,
2008). We are not aware of other studies that look at spill-over
effects of day-time office conditions on night-time sleep, but we
hypothesized that such effects should occur. Office conditions will
influence daily patterns of light and dark exposure, which will
influence circadian regulation (CIE, 2004). In addition, Akerstedt
et al. (2002) showed that ‘‘high work demands and physical effort
at work are risk indicators for disturbed sleep, while high social
support is associated with reduced risk’’. We predicted that
conditions that reduce exposure to windows, such as increased
distance to the window, could also reduce sleep quality.
Individual differences also influence sleep quality. Increased
need for sleep is one of the symptoms of seasonal affective disorder,
along with depressed mood, carbohydrate craving, and lethargy
(CIE, 2004; Veitch, 2001a). Murray et al. (2006) suggested that
a depressed person could have a more negative impression of his/
her surroundings, as a consequence of dysregulation between
internal (circadian or annual) rhythms. We further predicted that
sleep quality would vary in relation to whether individuals repor-
ted experiencing seasonal mood changes.
The desire for natural light, rather than electric light, is one of
the reasons why windows are so important to building occupants.
Windows are openings for flows – inward and outward – of air,
light, and sound, and are, therefore, often elements that influence
the indoor environment most (Tregenza & Loe, 1998). Preferences
for windows are well established (Farley & Veitch, 2001; Finnegan
& Solomon, 1981). A window view provides information about time
and weather, decreases the feeling of claustrophobia, and can have
a positive contribution to eye health by providing a distant horizon
at which to gaze. We predicted that proximity to a window would
lead to improved office impressions.
Windows can bring both positive and negative experiences:
access to view and daylight, but also glare and thermal discom-
fort. A good view should normally include the foreground and the
skyline (Littlefair, 1996), but care needs to be taken to control the
glaring effects of the sky. Several authors (e.g., Fisk et al., 1993;
Küller & Wetterberg, 1996; Veitch, Geerts, Charles, Newsham, &
Marquardt, 2005; Yildirim, Akalin-Baskaya, & Celebi, 2007)
showed more health problems and complaints among occupants
farther from the window. However, on closer analysis the rela-
tionship was revealed to be more complex. Veitch et al. (2005)
found that satisfaction with lighting was lower the further one
was from a window; however, overall environmental satisfaction
showed a quadratic relationship, being highest when one could
see a window but did not sit beside it. Satisfaction with venti-
lation (primarily thermal comfort) was lowest for people beside
the window (Charles, Veitch, Newsham, Marquardt, & Geerts,
2006).
Moving further from the window also means decreasing the
size of the largest potential glare source. On the other hand,
Chauval, Collins, Dogniaux, and Longemore (1982) found that
discomfort glare (from daylight) appears to be tolerated to a much
higher degree if there is a pleasant view from the window causing
the glare. Osterhaus (2001) wondered whether glare problems
become less of an issue or could be ignored when office rooms are
provided with a pleasant window view. In contradiction, the
desire for daylight – or at least direct sunlight – is sometimes
limited abruptly when it causes visual discomfort. People will
close their blinds and leave them closed, even when the source of
discomfort is long gone (Drucker-Colin, 1995). We predicted that
problems with glare and heat gain would be lower for people
farther from the window (As described below, we conceptualized
this relationship in the positive direction in a concept we called
environmental utility).
Of all characteristics of windows perhaps the most interesting is
the view. Markus (1967) found that the further occupants were
from a window the less satisfied they were with the view, and the
more they desired to sit nearer a window. The preference of people
for natural over built or urban views is shown in many window
studies (e.g., Chang & Chen, 2005; Hartig, Evans, Jamner, Davis, &
Garling, 2003; Tennessen & Cimprich, 1995; Ulrich, 1984). Natural
scenes are advantageous to human health because they provide an
opportunity for recovery from mental fatigue (Berman, Jonides, &
Kaplan, 2008; Kaplan, 1995). We predicted that urban views would
result in more discomfort and poorer office impressions; suburban
and rural views, containing greenery, should reduce discomfort and
improve office impressions. However, another way to characterize
view is in terms of its aesthetic quality; not all urban views are
alike, nor are all natural views equally attractive. We obtained
quality ratings of the views, and predicted that higher-quality
views would result in fewer discomfort complaints and better office
impressions.
The physical arrangement of the office environment influences
the level and type of social interaction between employees
(Danielsson & Bodin, 2008; Duval, Charles, & Veitch, 2002).
Density is an objective measure that is sub-divided into a measure
of social density and a measure of spatial density (Hayduk, 1983).
In an office setting, social density refers to the number of occu-
pants in an office, regardless of its floor area. Spatial density refers
to the floor area per person, for example, m2 per occupant in an
office (Duval et al., 2002). Chao, Schwartz, Milton, and Burge
(2003) reported that higher social density was associated with
more health complaints, and Brasche, Bullinger, Morfeld, Geb-
hardt, and Bischof (2001) found that this was especially true for
women. Danielsson and Bodin (2008) reported that the lowest
health status was found in medium-sized and small open-plan
offices; the best health was among employees in cell offices and
flex offices. Social density may also influence environmental
satisfaction in a room. According to Duval et al. (2002) occupants
of socially dense offices may perceive themselves as having less
privacy because they have fewer behavioural choices and more
interferences, which in turn leads to environmental dissatisfac-
tion. We predicted that people in offices with fewer occupants
would report fewer problems.
The predicted relationships are shown together in Fig. 2.
2. Method
2.1. Research design
This study was a cross-sectional survey of occupants in office
buildings in the Netherlands. Participants completed a question-
naire concerning their experience in the office, with a focus on
 M.B.C. Aries et al. / Journal of Environmental Psychology 30 (2010) 533–541 535

Fig. 2. Model 1 (hypothesized model).

indoor environment conditions. Physical measurements were
taken at a subset of workstations during a site visit. The analyses
reported here focus only on the survey responses; the physical
measurements are reported elsewhere (Aries, 2005). The survey
took place in the spring (April 8 to May 6) to minimize the possible
effects of seasonal problems such as runny noses in winter, seasonal
mood shifts (e.g., Seasonal Affective Disorder (SAD)), and the peak
of the pollen season (May to July).
individually distributed and collected on the same day. This
resulted in a high return rate (95%). The final sample numbered
333, of which 193 (58%) were men and 140 (42%) were women.
Age was ranked on five categories: under 30 (23%), 30–39 years
(31%), 40–49 years (25%), 50–59 years (19%) and older than 60
(2%).
2.4. Measures

2.2. Sites 2.4.1. Questionnaire

Ten office buildings in the Netherlands, containing private, semi-
private, and open-plan office spaces, were visited. Table 1 summa-
rizes information about the sample buildings. A building marked
with an industrial setting has a suburban location meaning that
there is a lower building density, that other buildings are more
distant, and that there is more green space nearby.
2.3. Survey participants
Each site was visited personally by the first author. Occupants
of the sample floors were individually approached and invited to
complete a questionnaire on paper. Approximately 90% of people
who were approached agreed to participate, and participants
were approximately equally distributed over the ten office
buildings; thus, respondents were a convenience sample based on
availability and willingness to participate. All questionnaires were
Participants received a 43-item questionnaire in Dutch. The
employees were told that the intention of the questionnaire was to
understand how office users experience present-day office envi-
ronments and what they think is important. In addition to the usual
demographic questions (age, sex, use of corrective lenses), the
questions concerned their experiences of the office environment,
their propensity for seasonal changes in mood, and the quality of
their sleep. Composite variables representing scores on the concepts
of interest were created by averaging or summing (as appropriate)
the related individual item scores, as follows.
 Physical and psychological discomfort – the ‘physical and
psychological discomfort’ variable was the mean of the survey
items: ‘concentration problems’, ‘easily tired’, ‘dullness’, ‘irri-
tability’, ‘headache’, ‘bad vision’, ‘dry throat’, ‘dry eyes’, ‘irritated
skin’, ‘sniffles’, each measured on a 5-point scale (0–4), with
higher values indicating more discomfort. The questionnaire

Table 1
General building characteristics.

Building code Setting Sample floors Business sector Questionnaire respondents Respondents/office type

Cell Group Open-plan

1 Urban 2nd; 4th Private financial 46 4 6 36
2 Urban 2nd Private financial 21 1 15 5
3 Urban 6th Private financial 29 10 19 –
4 Urban Ground Private manufacturer 29 2 1 26
5 Urban 4th; 5th Private manufacturer 51 3 9 39
6 Industrial Ground Private manufacturer 20 11 9 –
7 Industrial Ground Private manufacturer 25 4 12 9
8 Industrial 5th; 6th Government laboratory 39 20 12 7
9 Urban 5th Government social services 44 5 8 31
10 Industrial 7th; 8th Private legal services 29 13 15 1

Note: in the Netherlands, floor numbering starts with 0 for the ground floor and continues with 1 for the next floor, etc.
536 M.B.C. Aries et al. / Journal of Environmental Psychology 30 (2010) 533–541
items were selected from the TNO/RUL Chronic Fatigue and
Work Questionnaire (TNO/RUL, 1996).
 Sleep quality – a short form of the Groningen Sleep Quality
Scale (Leppämäki, Meesters, Haukka, Lönnqvist, & Partonen,
2003) was used to assess sleep quality (generalised, 7-items,
yes/no scale). Responses from five of the items were used here:
‘feel refreshed, ‘use energy pills’, ‘lack of sleep’, ‘easily fall
asleep’, and ‘difficulty falling asleep’, resulting in a final scale
from 0 to 5, with higher scores indicating better sleep quality.
The items ‘use energy pills’, ‘lack of sleep’, and ‘difficulty falling
asleep’ were reverse-coded.
 Environmental utility – for the ‘environmental utility’ variable,
five survey items related to visual and thermal disturbance
were used: ‘satisfaction with lighting at the computer screen’,
‘annoying reflections at the desk’, ‘annoying reflections at the
computer screen’, ‘unwanted heat of the sun’, and ‘perceived
warmth’, each measured on a 5-point scale (0–4) and reverse-
coded where necessary. The questions were taken from Hygge
and Löfberg (1999), translated into Dutch and with 5-point
scale instead of the original 3-point scales. The overall result
was a scale indicating the degree to which the visual and
thermal environment supports work, with higher values indi-
cating more environmental utility.
 Light quality – the ‘light quality’ variable was the mean of the
survey items: ‘satisfaction with lighting in office space’, ‘satis-
faction with lighting at the desk’, and ‘light quality’, each
measured on a 5-point scale (0–4), with higher values indicating
better conditions. The questions were from Hygge and Löfberg
(1999).
 Impression – the ‘impression’ was the mean of the survey
impression items ‘pleasant’, ‘interesting’, ‘spacious’, ‘quiet’,
‘enjoyable’, ‘tidy’, ‘clean’, ‘varied’, ‘fine’, ‘attractive’, ‘convenient’,
each measured on a 5-point scale (0–4), with higher values
indicating better (office) conditions. The items were taken from
De Laet et al. (2002) and translated into Dutch.
 Seasonality – the propensity to experience seasonal mood
changes was determined based on whether or not participants
self-reported having experienced differences in sleep
demands, social activities, mood, weight, and energy level in
summer and winter. Each item was scored on a 5-point scale
(1–5), and the total sum for the five items was calculated,
resulting in a final scale from 5 to 25. In a second question, the
employee was asked to what extent seasonal changes limit
activities, measured on a 5-point scale, with a higher score
indicating more limitation. The total experienced seasonal
mood changes in combination with the extent of limitation of
those activities formed three categories (Aries, 2005, table 3.8,
p. 63). People in category 1 declared to feel no influences due to
seasonal changes, category 2 people indicated a moderate
influence and people who experienced great differences
between the dark (winter) and light season (rest of the year)
were in category 3. The two questions were derived from the
Seasonal Pattern Assessment Questionnaire (SPAQ) by Rosen-
thal, Bradt, and Wehr (1987).
2.4.2. Office characteristics
A procedure similar to the International Energy Agency (IEA)
Monitoring Procedures for the Assessment of Daylighting (Velds &
Christoffersen, 2001) was used to record building and workstation
variables. A specially developed, mobile experimental set-up was
used to characterize lighting conditions (Aries, 2005). For each
workstation measured, a list of workstation characteristics (e.g.,
dimensions, window distance, type of VDU, office type, etc.) was
filled in and photos were taken simultaneously. Four of the
variables were used as independent variables in these analyses:
‘view quality’, ‘view type’, ‘window distance’, and ‘social density’.
Variables representing scores on the concepts of interest were
created by summing the related individual item scores, as follows.
 View quality – at a later time, photographs of representative
views for each of the occupants (see Fig. 3 for two examples)
were shown to a small panel of independent raters (N ¼ 4),
who rated each view on a 5-point scale. Their score was aver-
aged and divided into three groups of approximately the same
size: 1.0–2.2: bad, 2.3–3.2: neutral, and 3.3–5.0: good. N ¼ 105
views were rated as bad, N ¼ 100 as neutral, and N ¼ 129 as
good. The average intraclass correlation coefficient was of 0.79.
This is acceptable, although it does indicate differences of
opinion between the raters.
 View type – for the ‘view type’ variable the same set of photos as
for the view quality variable was used. For view type a distinc-
tion between a view of nature and an urban view was made. Out
of all view possibilities available, N ¼ 207 were categorized as
‘urban’ (coded ‘0’) and N ¼ 126 as ‘nature’ (coded ‘1’).
 Window distance – in our analysis, window distance from each
desk was divided in three categories based on the site visits:
0–2 m, 2–4 m, and over 4 m.
 Social density – the typical office design in these buildings was
a set of desks without separating partitions in a windowed
room with a door to a corridor. Social density refers to the
number of people in the room. We used three levels of social
density in this analysis: 1–2 persons, 3–5 persons, and more
than 5 persons.
3. Results
3.1. Descriptive statistics
Descriptive statistics for both questionnaire variables and office
characteristics in this analysis are shown in Table 2. The table also
includes the Cronbach’s alpha values for the composite questionnaire
variables. Only the ‘sleep quality’ variable has a low alpha (a ¼ 0.57);
most likely due to the composition out of dichotomous variables.
3.2. Within and between-group analysis (WABA)
The buildings were selected based on convenience and on the
willingness of organizations to host the study. The possibility of
biases associated with the selection of certain organizations or
certain buildings and by the confound of buildings and workstation
characteristics cannot be excluded. This means that observations
from all the people in one building might be highly correlated by
virtue of coming from one organization or because of commonly
experienced conditions. If so, this would violate a fundamental
statistical assumption, that observations are independent of one
another. Similarly, observations from offices of a given orientation
would be likely to be correlated because of commonalities in the
available daylight.
In this study, neither façade orientation nor building was
a variable of interest; we wished to combine the data into one large
sample. Using a similar approach as Veitch, Charles, Farley, and
Newsham (2007), a series of statistical analyses was conducted to
determine the legitimacy of combining individual data from the
buildings and/or orientations into one large sample in which we
ignored the building and/or orientation as a variable.
The first Within And Between-group Analysis (WABA) was per-
formed for the variable ‘orientation’. The results show that none of
the six dependent variables met all four criteria for group effects (see
Appendix, Table A1). Overall, we concluded that the assumption that
 M.B.C. Aries et al. / Journal of Environmental Psychology 30 (2010) 533–541 537

Fig. 3. A good view and a bad view, according to independent raters.

observations were independent of orientation was met. Further the five dependent variables met all four criteria for group effects.
analyses proceeded by combining all cases in one group, ignoring the Therefore, we concluded that the assumption that observations
office (façade) orientation. were independent of building was met. Further analyses proceeded
The analysis was repeated for the variable ‘building’ (see by combining all cases in one group, ignoring building effects, after
Appendix, Table A2). All but one of the six variable scales showed excluding building 9.
within-groups agreement. In this case, we cannot proceed by
combining all cases in one group. To determine whether it was 3.3. Path analysis
a specific building that caused the within-groups effect, we con-
ducted a univariate analysis of variance (ANOVA) for all question- EQS 6.1 for Windows was used to test the model using path
naire variables (physical and psychological discomfort, analysis (Bentler & Wu, 2003). EQS is a structural equations
environmental utility, sleep quality, light quality, impression, and modeling package that can also be used to perform path analyses.
seasonality). A critical look at all ANOVA results between the ten In path analysis, more than one regression analysis may be called
buildings investigated showed that building 9 jumped out as for. A path analysis model is an ordinary multivariate regression
(significantly) different from the other buildings. Closer inspection of model. At each stage, a variable is regressed on the variables that
the building characteristics (see Table 1) showed that employees in are hypothesized to affect it. We followed rules for path analysis as
building 9 provide Governmental Social Services, including dealing set by Asher (1983). The data input file used is a variance/covari-
with the public. Generally, these kinds of services could cause more ance matrix from the variables in the model.
stress in the job activities. As ‘job stress’ is directly and indirectly Data preparation for the path analyses was as follows: Cases
influenced by physical symptoms and satisfaction (Hedge et al., with missing data on multiple items were excluded from analyses.
1989; Hogdson, Muldoon, Collopy, & Olesen, 1992), we believed that We tested each dependent and independent variable for normality.
this might be a reason for the differences between building 9 and the Following recommendations by Kline (1997), we looked for skew-
other buildings. However, as there were no questions related to job ness values between þ3 and 3, and kurtosis values between þ8
stress in the questionnaire or any other indication in this field study and 8. All the variables met these criteria. We further examined
to confirm the assumption, there was no way to test this notion. the data for univariate and multivariate outliers. Univariate outliers
Instead, we repeated the WABA after excluding the cases from were defined as cases on which the absolute value of the stan-
building 9 (see Appendix, Table A3). Without building 9, none of dardized score for that variable was greater than 3. These 13 cases
were omitted from analysis. We examined the Mahalanobis
Table 2
distance statistic for each case. We found one multivariate outlier
Descriptive statistics. and excluded this case from further analyses. We did not look for
further multivariate outliers after the first exclusion. To check for
Questionnaire variables N Mean Std. dev. Min.a Max. Cronbach’s a
multicollinearity, the correlations between the questionnaire
Phys. and psych. 321 9.07 5.83 0 32 0.82
variables and the office characteristics were calculated. No high
discomfort
Sleep quality 329 4.02 1.17 0 5 0.57 correlations (above r ¼ 0.500) were found.
Environmental utility 302 13.64 4.27 1 20 0.79 Model fit for the path analyses were assessed using multiple
Light quality 325 9.83 2.09 1 12 0.76 statistical and fit indices including c2, Joreskog–Sorbom’s Goodness
Impression 318 24.48 5.51 0 43 0.84 of Fit Index (GFI), Joreskog–Sorbom’s Adjusted Goodness of Fit
Seasonality 316 1.64 0.70 1 3 0.82b
Index (AGFI), Comparative Fit Index (CFI), Bentler-Bonnet Normed
Valid N (listwise) 272 Fit Index (NFI), Bentler–Bonnet Non-Normed Fit Index (NNFI), Root
Office characteristics N Mean Std. dev. Min.a Max. Mean-Square Residual (RMR), Standardized RMR (SRMR), and Root
View quality 333 1.38 0.49 1 2 Mean-Square Error of Approximation (RMSEA). In addition, we also
View type 333 0.38 0.49 0 1 looked for statistical significance of the paths and the proportion of
Window distance 332 1.29 0.45 1 2 standardized residuals between 0.1 and 0.1. The Legrange Multi-
Social density 333 1.46 0.50 1 2
plier (LM) test and the Wald W statistic were examined to deter-
Valid N (listwise) 332 mine possible misfits. The LM test provides an estimate of how
a
Answer range starts at 0.00. much the overall c2 statistic would decrease if a particular variable
b
a for first 5 items only. were added. The Wald W-test estimates the amount the overall c2
538 M.B.C. Aries et al. / Journal of Environmental Psychology 30 (2010) 533–541
would increase if a particular free variable were dropped from the
model (Kline, 1997).
3.3.1. Model 1
The first model (Fig. 2, above) used the variables ‘gender’, ‘age’,
and ‘seasonality’ as individual factors. The variables ‘Social density’,
‘Window distance’, ‘View type’, and ‘View quality’ were included as
architectural factors. Within this group, we hypothesized that all
variables were correlated to correct for possible influences of specific
buildings; despite the results of the WABA for the variable ‘building’
(see paragraph 3.2). The two variables ‘Office impression’, and ‘Light
quality’ represented perceived environmental conditions. We
assumed that all individual and architectural factors would affect the
outcome variables directly and indirectly (by the perceived envi-
ronmental conditions), and that they did not affect each other. The
variables ‘physical and psychological discomfort’, ‘sleep quality’, and
‘environmental utility’ were considered outcome variables.
After running the model with EQS, the model showed a c2
(18) ¼ 97.16 (p ¼ 0.000), with the statistical significance of the c2
indicating a poor fit (see Table 3). This result was confirmed by all of
the other fit indices (Table 3).
3.3.2. Model 2
We modified Model 1 based on the results of the Wald W-test
and on theoretical considerations. Paths with insignificant path
coefficients (ß-values) were deleted. As our first model contained
almost all paths possible (except between individual and archi-
tectural factors), the suggestions according to the LM test were
ignored. The test suggested mainly paths between the individual
and architectural factors, which in this context should be consid-
ered artefacts or sources of bias.
The only addition to the model was the path between the
outcome variables ‘sleep score’ and ‘physical and psychological
discomfort’. This theoretical relationship was based on evidence
that a poor psychosocial work environment doubled the risk of
developing a sleep problem and that there are relationships
between sleep quality and physical symptoms such as headaches
(e.g., Bastien, Vallieres, & Morin, 2004; Bruni et al., 2007; Linton,
2004). Although not suggested by the Wald W-test, we removed
the variable ‘age’ from the model to increase the degrees of freedom
of the model. There was a sizeable correlation (r ¼ 0.262, p < 0.001)
between ‘age’ and ‘gender’ that we viewed as an artefact. The
employees’ ages had not been requested in years (continuous), but
by means of age categories. Moreover, the distribution of the age
groups was arbitrary and without theoretical explanation in rela-
tion to specific variables.
The results of the path analysis for the final model are shown in
Fig. 4 and Table 3. All path coefficients were significant (p ¼ 0.05).
The final model had a Chi-square of 49.73 on 38 degrees of freedom
(p ¼ 0.10). The frequency distribution of the standardized residuals
revealed that most residuals (86.4%) fell between 0.10 and 0.10,
which is desirable. The reliability of the model was tested by means
of different fit indices. The minimum discrepancy in our final model
is 1.31, a good fit as it is below 3.0. The Root Mean-Square Error of
Table 3
Goodness of fit for path analysis.
N c2 c2/df GFI AGFI
Targets <3 >0.90 >0.90
Model 1 211 97.16 5.40 0.92 0.71
Model 2 (final) 211 49.73 1.31 0.96 0.93
Note: NFI ¼ Bentler–Bonnet Normed Fit Index; NNFI ¼ Bentler–Bonnet Non-Normed Fit Index; CFI ¼ Comparative Fit Index; GFI ¼ Joreskog–Sorbom’s GFI Fit Index;
AGFI ¼ Joreskog–Sorbom’s AGFI Fit Index; RMR ¼ Root Mean-Square Residual; RMR ¼ Standardized RMR; RMSEA ¼ Root Mean-Square Error of Approximation;
St. Res ¼ Standardized Residuals. Optimal values are based on Byrne (1994) and Kline (1997).
Approximation (RMSEA) of 0.04 indicates a reasonable error of
approximation. The model was tested according to other fit indices
(see Table 3). As all but one of the indices show a value over 0.90, we
concluded that the modified model fits the data extremely well.
Additional statistics were examined to assess the possibilities for
improving model fit by adding or dropping parameters. The Wald W-
test indicated the model would not be improved by dropping any
parameters. The LM test indicated only unrealistic improvements
(e.g., ‘social density’ on ‘gender’) and, therefore, it was decided not to
add these links. All estimated factor loadings were significant.
One of the outcomes of the EQS output is a standardized solu-
tion. All model variables are rescaled to have unit variance. The
magnitude of these standardized coefficients may be easier to
interpret than coefficients from covariance or raw data. The stan-
dardized residual variance (R2) in an equation can be obtained as
the square of the coefficient associated with the residual variable. In
the behavioural sciences, it is unusual to find extremely high R2
values, and generally an R2 greater than 0.25 is considered a large
effect size (Cohen, 1988), 0.09 or greater is medium-sized; 0.04 is
small (Table 4).
The main focus of the analysis was to investigate predictors for
physical and psychological discomfort. The standardized solution for
the final model shows that for ‘Physical and psychological discom-
fort’, the variables ‘gender’, ‘seasonality’, ‘impression’, ‘view quality’,
‘view type’, and ‘social density’ explain almost 22% of the variance.
‘Seasonality’ and ‘physical and psychological discomfort’ explain 27%
of the variance for ‘sleep quality’. Variance for environmental utility
is for 13% explained by light quality and office impression.
4. Discussion
The final model showed a good fit to the data. Moreover, the
relationships were largely in the predicted directions. The results
open new avenues for office environment research. The first of
these is to extend the understanding of the aesthetic quality of
window views beyond the simple natural/built environment
dichotomy. The second is the consideration of the possible spill-
over effects of physical work conditions and their evaluations on
well-being after work.
Three out of four architectural factors had a significant influence
on physical and psychological discomfort: view type (ß ¼ 0.23),
view quality (ß ¼ 0.20), and social density (ß ¼ 0.15). There is
comparatively little data on social density effects (Duval et al.,
2002), but the finding here is broadly consistent with the literature
(Brasche et al., 2001; Chao et al., 2003; Danielsson & Bodin, 2008).
The more people present in the office room, the more discomfort
the occupants reported. A denser population and its activities could
make it harder to concentrate, could reduce the air quality, and
could mean more effluents.
Indirectly, nature views reduced discomfort through office
impressions; unexpectedly, nature views increased discomfort
directly. This seems not consistent with data of others (e.g., Chang &
Chen, 2005; Kaplan, 1993), which indicated that a window with
a view of nature has a more positive effect on the human psycho-
CFI NFI NNFI SRMR RMSEA (90% CI) St. Res
>0.90 >0.90 >0.90 <0.10 <0.10 %
0.75 0.74 0.25 0.07 0.15 89.4
0.96 0.87 0.95 0.06 0.04 86.4
 M.B.C. Aries et al. / Journal of Environmental Psychology 30 (2010) 533–541
Fig. 4. Model 2 (final model).
physiological response than one of an urban scene. However, view
quality negatively predicted discomfort directly (i.e., a better-
quality view was associated with lower discomfort, ß ¼ 0.20). The
finding that a good view can be associated with fewer self-reported
discomfort problems is consistent with other studies
(e.g., Heschong, 2003).
Surprisingly, the architectural factors had few direct effects on
office impressions and light quality ratings. We did observe that
nature views predicted higher office impressions (ß ¼ 0.29), but
neither view quality nor window distance had the predicted
effects; nor did any of the architectural variables influence judge-
ments of light quality.
Also unexpectedly, window distance did not contribute to
discomfort directly, although it did predict environmental utility as
we had predicted (ß ¼ 0.23). The lack of relationship between
window distance and discomfort might arise from the fact that
window access is mandatory in the Netherlands, and there were no
partitions to block view access. Environmental utility on the other
hand does have a significant relationship with window distance; the
farther the participant was from the window, the higher the envi-
ronmental utility (in the form of fewer problems with heat or glare).
This is consistent with the literature (Veitch et al., 2005) and with
building physics: Deeper in the room there is less chance for direct
sunlight to result in glare and thermal problems. Environmental
utility had also a positive link to light quality (ß ¼ 0.28): People who
rated the lighting quality as higher had higher environmental utility,
Table 4
Standardized solution.
Physical & ¼ 0.34  Seasonality þ 0.19  Gender  0.15
psychological  Impression þ 0.23  View type  0.20 
discomfort View quality þ 0.15  Social density
þ 0.89  E
Environmental ¼ 0.28  Light quality þ 0.23 
utility Window distance þ 0.93  E
Sleep quality ¼ 0.12  Seasonality  0.47 
Discomfort þ 0.85  E
Light quality ¼ 0.14  Gender þ 0.99  E
Impression ¼ 0.29  View type þ 0.96  E
539
which is also consistent with other research (Veitch, Newsham,
Boyce, & Jones, 2008).
Future research should further examine the relationship
between view characteristics beyond view content and view
quality. We did not find a significant relationship between view
quality and/or type and environmental utility. Tuaycharoen and
Tregenza (2007) found that glare discomfort (which is one of the
aspects of our ‘environmental utility’ variable) decreased as the
interest of view increased. Similar to our approach, independent
groups of subjects rated the interest of all views. View type and
view quality were strongly linked to each other (ß ¼ 0.44), but they
both had different relationships with other variables. Next to view
quality and view type, the interest evoked by a view might be
a third element that influences employee’s comfort ratings (visual
and/or physical and psychological). Such research could have
practical implications for urban design and architecture; not all
views can be natural, but many possibilities exist for making views
interesting or attractive.
The direct relation between seasonality and sleep (ß ¼ 0.12) is
not surprising; sleep quality problems are often an indicator for
seasonal mood disorders. Seasonality also indirectly affected sleep
quality through its effect on physical and psychological discomfort
at work. People with a higher seasonality score reported more
discomfort at work. We could not study whether people who are
more or less affected by seasonality have differing reactions to
window and view, because the sample size was too small to split.
The influence of seasonality on work experience, including possible
environmental moderators, is a topic for future research.
A building or interior designer cannot change employees’
gender (women reported more discomfort than men; ß ¼ 0.19) or
depression sensitivity, but can influence the office environment.
R2 The negative relationship between impression and discomfort in
0.22 the model (ß ¼ 0.15) showed that the more positive an employee
is about his/her office environment, the less physical and psycho-
logical discomfort is reported. This is logical: A spacious, quiet,
convenient, and clean office environment is likely to provide fewer
0.13
irritations, concentration problems, stress, and sources of bacteria
0.27 that can cause illness. Given the further link from discomfort to
sleep quality (ß ¼ 0.47), the potential for work environments to
0.02 influence workers’ health and well-being is broader than their
0.09
experiences during the working day. The direct effect of physical
540 M.B.C. Aries et al. / Journal of Environmental Psychology 30 (2010) 533–541

and psychological discomfort at work on sleep quality at home is Table A3. Within and between-group analysis – building (N ¼ 9;
well established for psychosocial stressors (e.g., Bastien et al., 2004; building 9 out).
Kuppermann et al., 1995; Linton, 2004) but to our knowledge has
not previously been studied in relation to the physical work envi- ICC (1) ICC (2) hbn h2bn hwn h2wn E testa F Corr. F

ronment. This and other possible instances of spillover from work Phys. and psych. 0.11 0.51 0.24 0.06 0.97 0.94 0.25 2.06* 0.49
discomfort
to home are worthy of further study. b
Environmental 0.24 0.74 0.33 0.11 0.95 0.89 0.35 3.80* 0.26*
The results of this investigation add weight to the argument utility
that even within the relatively benign environment of present-day Sleep quality 0.07 1.39 0.11 0.01 0.99 0.99 0.11 0.42 2.39
offices, workplace conditions influence employees’ health and Light quality 0.15b 0.61 0.27 0.07 0.96 0.93 0.28 2.56* 0.39
well-being. Architects, interior designers, and lighting designers Impression 0.264 0.76 0.34 0.11 0.94 0.89 0.36 4.22* 0.24*
Seasonality 0.06b 0.34c 0.21 0.04 0.98 0.96 0.22 1.53* 0.66*
all together can contribute to the reduction of the employee’s
physical and psychological discomfort at work, and these Note. *p  0.05.
a
E  0.58 indicates independence.
improvements appear to carry over to life at home by influencing b
ICC (1)  0.12.
sleep quality. c
ICC (2)  0.85.

Acknowledgements
References
This paper is an extended analysis of data collected as part of the
first author’s Ph.D. dissertation research at the Eindhoven Univer- Akerstedt, T., Knutsson, A., Westerholm, P., Theorell, T., Alfredsson, L., & Kecklund, G.
(2002). Sleep disturbances, work stress and work hours: a cross-sectional study.
sity of Technology (TU/e), the Netherlands (Aries, 2005). M. B.C.
Journal of Psychosomatic Research, 53(3), 741–748.
Aries acknowledges the advice of the promoter Dr. S. H. A. Bege- Aries, M. B. C. (2005). Human lighting demands – healthy lighting in an office envi-
mann, the co-promotor Dr. A. D. Tenner, and L. Zonneveldt, M.Sc., ronment. Eindhoven, Eindhoven University of Technology; Ph.D. thesis, 152 pages.
and the financial support of Netherlands Organisation for Applied Aries, M. B. C., Veitch, J. A., & Newsham, G. R. (2007). Physical and psychological
discomfort in the office environment. In Proceedings of the light, performance
Scientific Research (TNO), during her studies at TU/e. The authors and quality of life symposium. Eindhoven, The Netherlands: Light & Health
would like to thank all the participants and facility managers in the Research Foundation SOLG. pp. 45–49.
Dutch buildings visited for their cooperation. They also thank Dr. M. Asher, H. B. (1983). Causal modeling. Beverly Hills, CA: Sage University. Paper Series
on Quantitative Applications in the Social Sciences, 07-003.
R. Atif at NRC Institute for Research in Construction for supporting Bastien, C. H., Vallieres, A., & Morin, C. M. (2004). Precipitating factors of insomnia.
this extended analysis. A version of this paper was presented at the Behavioral Sleep Medicine, 2(1), 50–62.
symposium of the Dutch Light and Health Research Foundation Bentler, P. M., & Wu, E. J. (2003). EQS structural equations program. Version 6. 1 (Beta
build 62). Encino, CA: Multivariate Software, Inc.
(November 8th, 2007) in Eindhoven, the Netherlands (Aries, Veitch, Berman, M. G., Jonides, J., & Kaplan, S. (2008). The cognitive benefits of interacting
& Newsham, 2007). with nature. Psychological Science, 19(12), 1207–1212.
Brasche, S., Bullinger, M., Morfeld, M., Gebhardt, H. J., & Bischof, W. (2001). Why do
women suffer from sick building syndrome more often than men? Subjective
Appendix A. Building comparison higher sensitivity versus objective causes. Indoor Air, 11(4), 217–222.
Bruni, O., Russo, P. M., Ferri, R., Novelli, L., Galli, F., & Guidetti, V. (2007). Relation-
ships between headache and sleep in a non-clinical population of children and
Table A1. Within and between-group analysis – orientation. adolescents. Sleep Medicine, 9(5), 542–548.
Byrne, B. M. (1994). Structural equation modeling with EQS and EQS/Windows: Basic
ICC (1) ICC (2) hbn h2bn hwn h2wn E testa F Corr. F concepts, applications, and programming. Thousand Oaks, CA: Sage Publications.
Chang, C.-Y., & Chen, P.-K. (2005). Human response to window views and indoor
Phys. and psych. 0.14 0.93 0.07 0.01 1.00 1.00 0.07 0.52 1.93
plants in the workplace. HortScience, 40(5), 1354–1359.
discomfort
Chao, H. J., Schwartz, J., Milton, D. K., & Burge, H. A. (2003). The work environment
Environmental 0.49b 0.80 0.22 0.05 0.98 0.95 0.22 4.87* 0.21
and workers’ health in four large office buildings. Environmental Health
utility Perspectives, 11(9), 1242–1248.
Sleep quality 0.06 0.21 0.11 0.01 0.99 0.99 0.11 1.27 0.79 Charles, K. E., Veitch, J. A., Newsham, G. R., Marquardt, C. J. G., & Geerts, J. (2006).
Light quality 0.14b 0.38 0.12 0.02 0.99 0.99 0.12 1.62 0.62 Satisfaction with ventilation in open-plan offices: COPE field findings. In
b
Impression 0.60 0.86c 0.25 0.06 0.97 0.94 0.26 6.89* 0.15 Proceedings of healthy buildings 2006: Creating a healthy indoor environment for
Seasonality 0.03 0.13 0.09 0.01 1.00 0.99 0.09 0.89 1.13 people. Lisboa, Portugal: International Society of Indoor Air Quality and Climate.
pp. 93–98.
Note. * p  0.05. Chauval, P., Collins, J. B., Dogniaux, R., & Longemore, J. (1982). Glare from windows:
a
E  0.58 indicates independence. current views of the problem. Lighting Research and Technology, 14(1), 31–46.
b
ICC (1)  0.12. Chung, T. M., & Burnett, J. (2000). Lighting quality surveys in office premises. Indoor
c
ICC (2)  0.85. and Built Environment, 9(6), 335–341.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.).
Hillsdale, NJ: Erlbaum, Lawrence Erlbaum Associates. 572 pages.
Commission Internationale de l’Eclairage (CIE). (2004). Ocular lighting effects on
human physiology and behaviour. No. CIE 158:2004. Vienna, Austria: CIE.
Danielsson, C. B., & Bodin, L. (2008). Office type in relation to health, well-being, and
Table A2. Within and between-group analysis – building (N ¼ 10).
job satisfaction among employees. Environment and Behavior, 40(5), 636–668.
De Laet, I., Meynen, G., Van Tichelen, P., Vanuytsel, G., Taeymans, F., Zonneveldt, L.
ICC (1) ICC (2) hbn h2bn hwn h2wn E testa F Corr. F User acceptance study of the EE-SYLK daylighting system, Proceedings Right
Light conference, 2002, Nice, France, pp. 93–97.
Phys. and psych. 0.13b 0.60 0.26 0.07 0.97 0.93 0.27 2.50* 0.40
Drucker-Colin, R. (1995). The function of sleep is to regulate brain excitability in
discomfort order to satisfy the requirements imposed by waking: the function of sleep.
Environmental 0.40b 0.87c 0.44 0.19 0.90 0.81 0.49 7.75* 0.13* Behavioural Brain Research, 69(1–2), 117–124.
utility Duval, C. L., Charles, K. E., & Veitch, J. A. (2002). Open-plan office density and envi-
Sleep quality 0.07 1.55 0.11 0.01 1.00 0.99 0.11 0.39 2.55 ronmental satisfaction. NRC-IRC. NRC-IRC Research report IRC-RR-150, 25 pages.
Light quality 0.18b 0.69 0.29 0.09 0.96 0.92 0.30 3.24* 0.31* Farley, K. M. J., & Veitch, J. A. (2001). A room with a view: A review of the effects of
Impression 0.34b 0.84 0.39 0.15 0.92 0.85 0.42 6.12* 0.16* windows on work and well-being. NRC-IRC. NRC-IRC Research report IRC-RR-136,
Seasonality 0.03b 0.26 0.20 0.04 0.98 0.96 0.20 1.35 0.74 33 pages.
Finnegan, M. C., & Solomon, L. Z. (1981). Work attitudes in windowed vs.
Note. *p  0.05. windowless environments. Journal of Social Psychology, 115, 291–292.
a
E  0.58 indicates independence. Fisk, W. J., Mendell, M. J., Daisey, J. M., Faulkner, D., Hodgson, A. T., Nematollahi, M.,
b
ICC (1)  0.12. et al. (1993). Phase 1 of the california healthy building study: a summary. Indoor
c
ICC (2)  0.85. Air, 3(4), 246–254.
 M.B.C. Aries et al. / Journal of Environmental Psychology 30 (2010) 533–541
Haack, M., & Mullington, J. M. (2005). Sustained sleep restriction reduces emotional
and physical well-being. Pain, 119(1–3), 56–64.
Hartig, T., Evans, G. W., Jamner, L. D., Davis, D. S., & Garling, T. (2003). Tracking
restoration in natural and urban field settings: restorative environments.
Journal of Environmental Psychology, 23(2), 109–123.
Hayduk, L. A. (1983). Personal space: where we now stand? Psychological Bulletin,
94(2), 293–335.
Hedge, A., Burge, P. S., Robertson, A. S., Wilson, S., & Harris-Bass, J. (1989). Work-
related illness in offices: a proposed model of the ‘‘sick building syndrome’’.
Environment International, 15(1–6), 143–158.
Heschong, L. (2003). Windows and offices: A study of office worker performance and
the indoor environment. Fair Oaks, California: Heschong Mahone Group. Tech-
nical Report.
Hogdson, M. J., Muldoon, S., Collopy, P., & Olesen, B. (1992). Sick building syndrome
symptoms, work stress, and environment measures. In Proceedings IAQ ’92
Environment for people. Atlanta, GA, USA: American Society for Heating,
Refrigerating and Air-Conditioning Engineers. pp. 33–42.
Hygge, S., & Löfberg, H. A. (1999). Post occupancy evaluation of daylight in build-
ings, A report of IEA SHC Task 21/ECBCS Annex 20, 56 pages.
Kaplan, R. (1993). The role of nature in the context of the workplace. Landscape and
Urban Planning, 26, 193–201.
Kaplan, S. (1995). The restorative benefits of nature: toward an integrative frame-
work. Psychological Science, 15(3), 169–182.
Kline, R. B. (1997). Principles and practice of structural equation modeling. New York:
Guilford Press.
Küller, R., & Wetterberg, L. (1996). The subterranean work environment: impact on
well-being and health: the psychology of reactions to environmental agents.
Environment International, 22(1), 33–52.
Kuppermann, M., Lubeck, D. P., Mazonson, P. D., Patrick, D. L., Stewart, A. L.,
Buesching, D. P., et al. (1995). Sleep problems and their correlates in a working
population. Journal of General Internal Medicine, 10(1), 25–32.
Leech, J. A., Nelson, W. C., Burnett, R. T., Aaron, S., & Raizenne, M. E. (2002). It’s about
time: a comparison of Canadian and American time-activity patterns. Journal of
Exposure Analysis and Environmental Epidemiology, 12(6), 427–432.
Leppämäki, S., Meesters, Y., Haukka, J., Lönnqvist, J., & Partonen, T. (2003). Effect of
simulated dawn on quality of sleep – a community-based trial. BMC Psychiatry,
3(14).
Linton, S. J. (2004). Does work stress predict insomnia? A prospective study. British
Journal of Health Psychology, 9(2), 127–136.
Littlefair, P. J. (1996). Designing with innovative daylighting. London: BRE Publication
Crown Research Communications.
Markus, T. A. (1967). The function of windows – a reappraisal. Building Science, 2(2),
97–121.
Meerlo, P., Sgoifo, A., & Suchecki, D. (2008). Restricted and disrupted sleep: effects
on autonomic function, neuroendocrine stress systems and stress responsivity.
Sleep Medicine Reviews, 12(3), 197–210.
Murray, G., Michalak, E. E., Levitt, A. J., Levitan, R. D., Enns, M. W., Morehouse, R., et
al. (2006). O sweet spot where art thou? Light treatment of seasonal affective
541
disorder and the circadian time of sleep. Journal of Affective Disorders, 90(2–3),
227–231.
Osterhaus, W. K. E. (2001). Discomfort glare from daylight in computer offices: how
much do we really know? Proceedings of LUX Europa 2001, 9th European
Lighting Conference, Reykjavik, Iceland, pp. 448–456.
Rosenthal, N. E., Bradt, G. H., & Wehr, T. A. (1987). Seasonal pattern assessment
questionnaire. Washington, DC, USA: National Institute of Mental Health.
Schweizer, C., Edwards, R. D., Bayer-Oglesby, L., Gauderman, W. J., Ilacqua, V.,
Jantunen, M., et al. (2007). Indoor time-microenvironment-activity patterns in
seven regions of Europe. Journal of Exposure Analysis and Environmental Epide-
miology, 17(2), 170–181.
Tennessen, C. M., & Cimprich, B. (1995). Views to nature: effects on attention.
Journal of Environmental Psychology, 15(1), 77–85.
TNO/RUL. (1996). TNO Preventie en Gezondheid/Afdeling interne geneeskunde
Rijksuniversiteit Leiden. In Vragenlijst chronische vermoeidheid en werk
(Questionnaire chronic fatigue and work) (in Dutch).
Tregenza, P., & Loe, D. (1998). The design of lighting (1st ed.). London and New York:
E & FN Spon Press. 176 pages.
Tuaycharoen, N., & Tregenza, P. R. (2007). View and discomfort glare from windows.
Lighting Research and Technology, 39(2), 185–200.
Ulrich, R. S. (1984). View through a window may influence recovery from surgery.
Science, 224, 420–421.
Veitch, J. A. (2001a). Lighting quality contributions from biopsychological processes.
Journal of the Illuminating Engineering Society, 30(1), 3–16.
Veitch, J. A. (2001b). Psychological processes influencing lighting quality. Journal of
the Illuminating Engineering Society, 30(1), 124–140.
Veitch, J. A., Charles, K. E., Farley, K. M. J., & Newsham, G. R. (2007). A model of
satisfaction with open-plan office conditions: COPE field findings. Journal of
Environmental Psychology, 27(3), 177–189.
Veitch, J. A., Charles, K. E., & Newsham, G. R. (2004). Workstation design for the open-
plan office. Ottawa, ON: NRC Institute for Research in Construction, Construction
Technology. Update No. 61.
Veitch, J. A., Geerts, J., Charles, K. E., Newsham, G. R., & Marquardt, C. J. G. Satis-
faction with lighting in open-plan offices: COPE field findings, Proceedings of
Lux Europa 2005, 10th European Lighting Conference, 2005, Berlin, Germany,
pp. 414–417.
Veitch, J. A., & Newsham, G. R. (1998). Lighting quality and energy-efficient effects
on task performance, mood, health, satisfaction and comfort. Journal of the
Illuminating Engineering Society, 27, 107–130.
Veitch, J. A., Newsham, G. R., Boyce, P. R., & Jones, C. C. (2008). Office lighting
appraisal, performance, and well-being, a linked mechanisms approach.
Lighting Research & Technology, 40(2), 133–151.
Velds, M., & Christoffersen, J. (2001). Monitoring procedures for the assessment of
daylighting performance of buildings. Bruxelles: International Energy Agency
(IEA). A report of IEA SHC task 21, ECBCS annex 29.
Yildirim, K., Akalin-Baskaya, A., & Celebi, M. (2007). The effects of window prox-
imity, partition height, and gender on perceptions of open-plan offices. Journal
of Environmental Psychology, 27(2), 154–165.