ArticlePDF Available

Environmental Satisfaction in Open-Plan Environments: 5. Workstation and Physical Condition Effects

Authors:

Abstract and Figures

Open-plan offices are notorious for their unpopularity with occupants. Among the most common anecdotal complaints are problems with distraction and inadequate privacy. As part of the Cost-effective Open-Plan Environments project, a field study was conducted to examine the relationships between measured physical conditions and occupant satisfaction with those conditions. A total of 779 workstations in nine buildings were visited. Lighting, acoustic, thermal and air movement conditions were recorded along with descriptive data about workstation size, partition height, and other characteristics. Occupants completed a 27-item questionnaire simultaneously with the measurements in their own workstations. The questionnaire covered satisfaction with individual features of the workstation, the environment overall, and the job, the rank ordered importance of seven physical features, and basic demographic characteristics. A mail-back questionnaire was provided to allow for longer comments about likes and dislikes.This report concerns the effects of workstation physical conditions on five aspects of satisfaction: satisfaction with privacy and acoustics; satisfaction with lighting; satisfaction with ventilation; overall environmental satisfaction, and job satisfaction. Hierarchical multiple regression analyses controlled for age, job type, and gender first; then examined the effects of workstation characteristics and additional physical variables. Separate nonparametric analyses were conducted for the rank order data, and the text comments were transcribed and characterized.
Content may be subject to copyright.
NRC Publications Archive (NPArC)
Archives des publications du CNRC (NPArC)
Web page / page Web
Environmental Satisfaction in Open-Plan Environments: 5.
Workstation and Physical Condition Effects
Veitch, J. A.; Charles, K. E.; Newsham, G. R.; Marquardt, C. J. G.;
Geerts, J.
Access and use of this website and the material on it are subject to the Terms and Conditions set forth at
http://nparc.cisti-icist.nrc-cnrc.gc.ca/npsi/jsp/nparc_cp.jsp?lang=en
L’accès à ce site Web et l’utilisation de son contenu sont assujettis aux conditions présentées dans le site
http://nparc.cisti-icist.nrc-cnrc.gc.ca/npsi/ctrl?lang=en
http://nparc.cisti-icist.nrc-cnrc.gc.ca/npsi/ctrl?lang=fr
LISEZ CES CONDITIONS ATTENTIVEMENT AVANT D’UTILISER CE SITE WEB.
READ THESE TERMS AND CONDITIONS CAREFULLY BEFORE USING THIS WEBSITE.
Contact us / Contactez nous: nparc.cisti@nrc-cnrc.gc.ca.
http://nparc.cisti-icist.nrc-cnrc.gc.ca/npsi/jsp/nparc_cp.jsp?lang=fr
Environmental Satisfaction in Open-Plan Environments: 5.
Workstation and Physical Condition Effects
Veitch, J.A.; Charles, K.E.; Newsham, G.R.;
Marquardt, C.J.G.; Geerts, J.
IRC-RR-154
October 20, 2003
http://irc.nrc-cnrc.gc.ca/ircpubs
Workstation and Physical Condition Effects on Environmental Satisfaction
Environmental Satisfaction in Open-Plan Environments:
5. Workstation and Physical Condition Effects
Jennifer A. Veitch, Kate E. Charles, Guy R. Newsham,
Clinton J. G. Marquardt, and Jan Geerts
Institute for Research in Construction
National Research Council Canada, Ottawa, ONT, K1A 0R6, Canada
IRC Research Report RR-154
October 20, 2003
IRC RR-154 1
Workstation and Physical Condition Effects on Environmental Satisfaction
Environmental Satisfaction in Open-Plan Environments:
5. Workstation and Physical Condition Effects
Jennifer A. Veitch, Kate E. Charles, Guy R. Newsham,
Clinton J. G. Marquardt, and Jan Geerts
Executive Summary
Open-plan offices are notorious for their unpopularity with occupants. Among the most common
anecdotal complaints are problems with distraction and inadequate privacy. As part of the Cost-effective
Open-Plan Environments project, a field study was conducted to examine the relationships between
measured physical conditions and occupant satisfaction with those conditions.
A total of 779 workstations in nine buildings were visited. Lighting, acoustic, thermal and air
movement conditions were recorded along with descriptive data about workstation size, partition height,
and other characteristics. Occupants completed a 27-item questionnaire simultaneously with the
measurements in their own workstations. The questionnaire covered satisfaction with individual features
of the workstation, the environment overall, and the job, the rank ordered importance of seven physical
features, and basic demographic characteristics. A mail-back questionnaire was provided to allow for
longer comments about likes and dislikes.
This report concerns the effects of workstation physical conditions on five aspects of satisfaction:
satisfaction with privacy and acoustics; satisfaction with lighting; satisfaction with ventilation; overall
environmental satisfaction, and job satisfaction. Hierarchical multiple regression analyses controlled for
age, job type, and gender first; then examined the effects of workstation characteristics and additional
physical variables. Separate nonparametric analyses were conducted for the rank order data, and the text
comments were transcribed and characterized.
Key findings are:
Environmental conditions in the offices generally met accepted standards. This sample of
workplaces was not random, but was not chosen to exemplify good or bad workplaces. Overall, there
were relatively few instances of conditions that did not meet applicable guidelines or standards.
Having access to a window or to daylight strongly improves satisfaction with lighting. Having a
window, or daylight within 15 ft (5 m), strongly improves satisfaction with lighting. The desire for a
window was a frequently mentioned comment among “things I would change” in the open-ended
remarks.
Having a window in the workstation has a detrimental effect on satisfaction with ventilation
and overall environmental satisfaction. We believe this reflects the problems of heat gain and
radiant cooling. Having a window is desirable, but poor thermal conditions are not.
Larger workstations are more satisfactory. Increasing workstation size improves satisfaction with
privacy.
Lower partition heights appear to improve satisfaction. This finding is paradoxical, as it is
contrary to previous research and common sense, particularly with respect to privacy. We suspect that
it might reflect the desire for better daylight penetration, which lower partitions afford, and to the
perception that lower partitions improve ventilation.
Concentrations of pollutants influence satisfaction with ventilation. Even at concentrations within
accepted limits, higher concentrations of carbon dioxide and other contaminants reduce satisfaction
with ventilation.
The next steps for research in this area should include a wider range of variables relating to
occupants, their work, and their organizations, to enable a finer-grained analysis and prescriptions for
workplace design that are tailored to individuals and their specific requirements.
IRC RR-154 2
Workstation and Physical Condition Effects on Environmental Satisfaction
Table of Contents
1.0 Introduction........................................................................................................................................... 5
2.0 Method .................................................................................................................................................. 6
2.1 Participants ........................................................................................................................................ 6
2.1.1 Buildings..................................................................................................................................... 6
2.1.2 Occupants. .................................................................................................................................. 6
2.2 Independent Variables....................................................................................................................... 9
2.3 Dependent Variables ....................................................................................................................... 10
2.4 Procedure......................................................................................................................................... 10
3.0 Results and Discussion ....................................................................................................................... 10
3.1 Descriptive Statistics ....................................................................................................................... 10
3.1.1 Dependent variables: Satisfaction............................................................................................. 10
3.1.2 Independent variables: Physical conditions.............................................................................. 11
3.1.3 Intercorrelations........................................................................................................................ 15
3.2 Analytic Strategy............................................................................................................................. 17
3.2.1 Data cleaning. ........................................................................................................................... 17
3.2.2 Independence of observations................................................................................................... 18
3.2.3 Hierarchical regression models................................................................................................. 19
3.3 Predicting Satisfaction with Privacy ............................................................................................... 20
3.3.1 Workstation characteristics....................................................................................................... 20
3.3.2 Acoustic conditions. ................................................................................................................. 20
3.3.3 Discussion: Satisfaction with privacy....................................................................................... 21
3.4 Predicting Satisfaction with Lighting.............................................................................................. 22
3.4.1 Workstation characteristics....................................................................................................... 22
3.4.2 Lighting conditions................................................................................................................... 23
3.4.3 Discussion: Satisfaction with lighting. ..................................................................................... 26
3.5 Predicting Satisfaction with Ventilation.......................................................................................... 27
3.5.1 Workstation characteristics....................................................................................................... 27
3.5.2 Ventilation/IAQ conditions. ..................................................................................................... 27
3.5.3 Discussion: Satisfaction with ventilation.................................................................................. 28
3.6 Predicting Overall Environmental Satisfaction ...............................................................................29
3.6.1 Workstation characteristics....................................................................................................... 29
3.6.2 Acoustic conditions. ................................................................................................................. 29
3.6.3 Lighting conditions................................................................................................................... 30
3.6.4 Ventilation/IAQ conditions. ..................................................................................................... 32
3.6.5 Discussion: Overall environmental satisfaction........................................................................ 33
3.7 Predicting Job Satisfaction .............................................................................................................. 34
3.7.1 Workstation characteristics....................................................................................................... 34
3.7.2 Acoustic conditions. ................................................................................................................. 34
3.7.3 Lighting conditions................................................................................................................... 35
3.7.4 Ventilation/IAQ conditions. ..................................................................................................... 37
3.7.5 Discussion: Job satisfaction...................................................................................................... 38
3.8 Cumulative Risk Factors ................................................................................................................. 39
3.8.1 Cumulative risk variables. ........................................................................................................ 39
3.8.2 Predicting overall environmental satisfaction........................................................................... 40
3.8.3 Predicting job satisfaction......................................................................................................... 41
3.8.4 Discussion: Cumulative risk. .................................................................................................... 41
3.9 Ranked Order of Importance of Workstation Features ................................................................... 42
3.9.1. Frequencies.............................................................................................................................. 42
3.9.2. Nonparametric analyses........................................................................................................... 42
IRC RR-154 3
Workstation and Physical Condition Effects on Environmental Satisfaction
3.9.3 Crosstabulations by workstation area. ...................................................................................... 42
3.9.4 Crosstabulations by partition height ......................................................................................... 43
3.9.5 Crosstabulations by windows and daylight .............................................................................. 43
3.9.6 Crosstabulations by age ............................................................................................................ 43
3.9.7 Crosstabulations by job category.............................................................................................. 44
3.9.8 Discussion: Ranked importance of features............................................................................. 44
3.10 Open-Ended Comments ................................................................................................................ 45
4.0 Conclusions......................................................................................................................................... 46
5.0 References........................................................................................................................................... 48
Acknowledgements................................................................................................................................... 52
Appendix A: Derived Thermal Indices Regressions.................................................................................53
IRC RR-154 4
Workstation and Physical Condition Effects on Environmental Satisfaction
1.0 Introduction
Open-plan offices have become the dominant interior design strategy for North American
organizations, driven both by the opportunity for lower real-estate costs and by the appeal of the notion
that reducing physical barriers between individuals might also remove social barriers (Brill, Margulis,
Konar, & BOSTI, 1984; Sundstrom, 1987). However, persistent problems in open-plan offices have made
them fodder for cartoonists such as Scott Adams (Dilbert™) and Francesco Marciuliano and Craig
Macintosh (“Sally Forth”). Among the most common complaints are lack of privacy and distractions that
prevent concentration (Brill, Weidemann, & BOSTI Associates, 2001; Brookes & Kaplan, 1972; Marans
& Spreckelmeyer, 1982; Mercer, 1979; Sundstrom, 1982; Zalesny & Farace, 1987). Problems with other
ambient conditions have also been reported, for instance poor indoor air quality and poor thermal comfort
(Hedge, 1982; Woods, Drewry, & Morey, 1987).
Two factors have consistently emerged as important influences on environmental satisfaction: the
area available to each employee and the degree of enclosure. Marans and Spreckelmeyer (1982) found
that the amount of space available to the employee was the strongest predictor of satisfaction with the
workstation, with larger sizes being more satisfactory. Other investigators have found that workstation
size predicts assessments of privacy, with larger workstations being perceived as more private (Oldham,
1988; O'Neill & Carayon, 1993). Increasing enclosure is associated with higher ratings of privacy
(Oldham, 1988; Sundstrom, Burt, & Kamp, 1980) and environmental satisfaction (e.g., Brennan, Chugh,
& Kline, 2002; Brill et al., 1984; Marans & Yan, 1989; Oldham, 1988).
Anecdotal reports from interior designers and facilities managers indicate that cubicles in open-
plan offices are smaller than ever, and often feature lower partitions than would have been typical in the
1980s ("Space planning", 2003). Concern that these changes would result in the creation of physical
conditions that would reduce environmental satisfaction was among the reasons for the creation of the
Cost-effective Open-Plan Environments project in 1999. It seemed likely that reducing cubicle size would
increase noise and distraction along with occupancy, and that more cubicles might mean more barriers to
light and air circulation. However, we could find few investigations that reported the physical conditions
in sufficient detail to predict precisely how the physical conditions might relate to environmental
satisfaction. Most of the investigations compared open-plan versus enclosed offices (e.g., Brennan et al.,
2002; Mercer, 1979; Oldham & Brass, 1979; Spreckelmeyer, 1993) and in general lacked detailed
measurements of physical conditions, or reported them in a manner that did not lend itself to open-plan
design recommendations (e.g., Brennan et al., 2002; Marans & Spreckelmeyer, 1982; Oldham, Kulik, &
Stepina, 1991; Oldham & Rotchford, 1983; Oldham & Fried, 1987; Sutton & Rafaeli, 1987). The few
studies that did measure a wide range of physical conditions did not measure at individual workstations,
but took averages across wide areas over periods of time (e.g., Hedge, Erickson, & Rubin, 1992).
Moreover, relatively few investigations appear to have taken place since the late 1980s, which means that
most predate the change to ubiquitous personal computing.
This field investigation was designed to fill a gap in the literature, by taking detailed
measurements of both the physical conditions and the opinions of the occupants. Although a truly random
sample of North American offices was not possible, the participants were from a variety of organizations,
both public and private-sector, in several cities. Basic demographic variables were controlled, with the
aim of providing information that could guide designers to providing open-plan office designs that will be
satisfactory to the wide range of potential occupants. In addition to enlarging our understanding of how
physical conditions influence environmental satisfaction, this cross-sectional field investigation is (to our
knowledge) the only source of descriptive statistics on the physical conditions experienced in North
American open-plan offices in the early 21st century.
IRC RR-154 5
Workstation and Physical Condition Effects on Environmental Satisfaction
2.0 Method
Detailed presentations of the method and participants in this cross-sectional field study have been
presented elsewhere (Charles, Veitch, Farley, & Newsham, 2003; Veitch, Farley, & Newsham, 2002). A
brief outline is provided here.
2.1 Participants
Data were collected in nine buildings between spring 2000 and spring 2002.
Five of the buildings were occupied by public sector Canadian organizations.
Four were occupied by private sector organizations in either Canada or the United States. The buildings
were located in Ottawa and Toronto (Ontario), Montreal and Quebec City (Quebec), and in the San
Francisco Bay area (California). All buildings, and the specific locations within them, were selected
because they contained open-plan offices occupied by white-collar workers, and because their
management was willing to host the visit. A summary of the building characteristics at each site is shown
in Table 1.
2.1.1 Buildings.
A total of 779 occupants of the nine buildings participated in the
investigation. They responded to a questionnaire about their satisfaction with
the physical environment while the NRC team collected physical data pertaining to their workstations (see
below). Demographic characteristics of these participants are shown in Table 2.
2.1.2 Occupants.
As may be seen in Table 2, several characteristics varied between buildings. One of the most
striking differences occurred in the frequency with which respondents chose to respond to the
questionnaire in English or French. We merged all the data, regardless of the language in which the
questionnaire had been completed. The study had not been designed to provide data for a comparison
between the two translations; moreover we were fairly confident that our translation and back-translation
procedure had provided equivalent forms, given that the responses did not in general involve subtle
emotional concepts that might differ from one language to another.
IRC RR-154 6
Workstation and Physical Condition Effects on Environmental Satisfaction
Table 1. Summary of site characteristics.
Building Year
Built
City Sector Visited # Floors Floor plate
(sf)
Lighting HVAC Windows Sound
1
1977 Ottawa public spring
2000
11
(4 visited)
39,000 (x 2
towers)
4' coffered
prismatic
fluorescent
ducted air VAV cooling
/ perimeter hot-water
heating
non-operable no sound
masking
2
1975 Toronto public summer
2000
12
(3 visited)
40,000 4' recessed
parabolic cube
ducted air VAV cooling
/ perimeter convention
heating
non-operable no sound
masking
3
1975 Ottawa public spring
2000 &
winter
2000
22
(4 visited)
18,000 4' recessed
prismatic (some
parabolic)
ducted air VAV cooling
/ perimeter hot and
chilled water heating &
cooling
non-operable sound
masking in
use
4
1976 Ottawa private winter
2002
15
(1 visited)
16,000 2’ x 4’ prismatic ducted air VAV cooling
/ perimeter hot-water
heating
non-operable no sound
masking
5 1994 San Rafael private spring
2002
3
(3 visited)
40,000 2’ x 4’ recessed
parabolic
ducted air VAV cooling
/ hot-water reheat
non-operable sound
masking in
use
6
1984 San Rafael private spring
2002
5
(1 visited)
35,000 2’ x 4’ recessed
parabolic
ducted air VAV cooling
perimeter hot-water
heating
non-operable no sound
masking
7
1916
(renovated
2000)
San
Francisco
private spring
2002
8
(1 visited)
41,000 8’ direct/ indirect ducted air VAV operable
windows
sound
masking in
specific
locations
8
1954 Montreal public spring
2002
4
(2 visited)
6,700 50% indirect /
50% 2’x 4’
parabolic
ducted air VAV /
perimeter heating
non-operable no sound
masking
9
1989/90 Quebec
City
public spring
2002
3
(3 visited)
15,300 1’ x 4’ parabolic Fan-coil with occupant-
controlled ceiling vents,
perimeter electric
heating
non-operable no sound
masking
IRC RR-154 7
Workstation and Physical Condition Effects on Environmental Satisfaction
IRC RR-154 8
Workstation and Physical Condition Effects on Environmental Satisfaction
Table 2. Demographic characteristics of participating occupants.
Site N % English % female /% male Mean age (SD)
Full sample 779 79.5 47.6 / 51.5 36.2 (10.6)
Building 1 132 85.6 47.7 / 51.5 38.2 (12.7)
Building 2 160 98.8 48.8 / 50.6 37.8 (9.4)
Building 3 127 75.6 49.6 / 48.8 39.5 (10.1)
Building 4 52 94.2 23.1 / 75.0 32.1 (8.0)
Building 5 85 97.6 67.1 / 31.8 33.1 (9.6)
Building 6 48 100.0 62.5 / 37.5 29.8 (9.4)
Building 7 72 100.0 31.9 / 68.1 30.7 (7.3)
Building 8 47 0.0 53.2 / 44.7 38.8 (9.9)
Building 9 56 0.0 35.7 / 64.3 37.3 (10.1)
Job Category (%)
Administration Technical Professional Management
Full sample 27.1 24.9 38.4 8.6
Building 1 18.9 11.4 68.2 0.0
Building 2 47.5 11.3 32.5 8.1
Building 3 39.4 22.8 24.4 11.8
Building 4 1.9 57.7 30.8 7.7
Building 5 20.0 20.0 41.2 17.6
Building 6 33.3 14.6 35.4 16.7
Building 7 6.9 52.8 25.0 15.3
Building 8 31.9 34.0 29.8 2.1
Building 9 10.7 42.9 46.4 0.0
Education (%)
High School Community
College
University
courses
Undergraduate
Degree
Graduate
Degree
Full sample 11.6 15.1 14.6 34.0 22.7
Building 1 9.1 8.3 13.6 30.3 37.1
Building 2 13.1 21.3 16.9 26.3 20.0
Building 3 26.8 22.8 12.6 21.3 12.6
Building 4 0.0 5.8 13.5 36.5 42.3
Building 5 4.7 3.5 12.9 58.8 17.6
Building 6 6.3 8.3 18.8 41.7 25.0
Building 7 2.8 5.6 19.4 48.6 23.6
Building 8 12.8 27.7 14.9 25.5 17.0
Building 9 14.3 30.4 8.9 35.7 10.7
Note. Percentages that do not sum to 100 are the result of rounding error and missing data.
2.2 Independent Variables
During the data collection visit, the NRC team used a specially designed and constructed cart
attached to a modified office chair to take measurements of the physical conditions at the workstation.
These measurements included illuminance at various points on the work surface, sound level at the
approximate location of a seated occupant’s ear, temperature and air movement at head, knee, and ankle
height of a seated occupant, relative humidity at torso height, and concentrations of carbon monoxide,
carbon dioxide, total hydrocarbons and methane, as well as the size of the workstation, height of
partitions surrounding the workstation, and number of enclosed sides of the workstation. Additional
IRC RR-154 9
Workstation and Physical Condition Effects on Environmental Satisfaction
acoustic and illuminance measurements were taken at night, with no occupants and no daylight. This
equipment was described in detail by Veitch et al. (2002).
2.3 Dependent Variables
Participating occupants completed a 27-item questionnaire. It consisted of 18 individual ratings of
their satisfaction with specific environmental conditions, two overall ratings of environmental
satisfaction, two items assessing job satisfaction, one set of rankings of the relative importance to that
individual of 7 environmental features, and four demographic characteristics: age, sex, job type, and
education level (Veitch et al., 2002).
Exploratory and confirmatory factor analyses were used to create three subscales of satisfaction
from the 18 individual items (Charles et al., 2003; Veitch et al., 2002). Thus, the final set of dependent
variables for this field study comprised Satisfaction with Privacy (Sat_Priv), Satisfaction with Lighting
(Sat_Light), Satisfaction with Ventilation (Sat_Vent), Overall Environmental Satisfaction (OES), and Job
Satisfaction (JobSatis). Each was calculated as the average of the contributing items, on scales from 1 to
7. The demographic characteristics were used as control variables in the regression analyses. Ranked
importance was analysed separately and is reported below.
2.4 Procedure
Building occupants were contacted by memo or e-mail by their management prior to the visit by
the NRC research team, to inform them about the investigation and to invite their participation. During
the visit, a research team of two NRC staff visited individual workstations in the designated areas of the
building. The team approached the occupants individually to invite their participation; over 95% agreed to
participate. Having accepted the invitation, the participant was conducted to an adjacent workstation to
complete the questionnaire on a handheld computer. At the same time, the NRC team replaced the
participant’s usual chair with the instrumented chair and took the physical measurements of the
workstation. Each workstation visit took approximately 13 minutes. At the end of the visit the team
moved on to the next occupied workstation. Two teams returned to the building to take night-time
measurements of acoustic conditions and illuminances. Further details of the procedure are in Veitch et al.
(2002).
Participants received no reward for participation, but both employees and management of the
building received a report summarizing the physical conditions and aggregate satisfaction responses in
that building.
3.0 Results and Discussion
3.1 Descriptive Statistics
As previously stated, the dependent variables were
scale scores for five aspects of satisfaction, each
measured on a scale from 1 to 7, with rising numbers reflecting greater satisfaction. The overall
descriptive statistics are shown in Table 3. These statistics are for all cases with valid data (including
cases excluded from regression analyses as univariate outliers). In general, among the aspects of the
physical environment, Sat_Priv scores were lowest, and Sat_Light scores highest. JobSatis was very high,
with a mean of 5.07 and median of 5.00. Over half of the sample were “satisfied” or “very satisfied” with
their jobs.
3.1.1 Dependent variables: Satisfaction.
IRC RR-154 10
Workstation and Physical Condition Effects on Environmental Satisfaction
Table 3. Full sample descriptive statistics for dependent variables.
Variable N Mean SD Median Minimum Maximum
SAT_PRIV 775 3.88 1.12 3.90 1.00 6.70
SAT_VENT 775 4.25 1.41 4.33 1.00 7.00
SAT_LIGHT 776 4.75 1.20 5.00 1.40 7.00
OES 745 4.05 1.31 4.00 1.00 7.00
JOBSATIS 767 5.07 1.08 5.00 1.00 7.00
From the many physical measurements we
selected a subset for the regression analyses.
These were selected because of their relevance to the project (workstation size and partition height), their
theoretical relevance (degree of enclosure), or to cover the most important elements of the acoustic,
ventilation/temperature, and lighting conditions, as identified in previous COPE research or in the
scientific literature. The descriptive statistics for each variable and their definitions are provided in tables
4, 5, 6, and 7. Sample sizes differ slightly from one variable to another because of equipment failures and
operator errors; these were random losses. Acoustic variables have the greatest losses because in two
buildings some workstations were not available for the necessary night-time measurements.
3.1.2 Independent variables: Physical conditions.
Table 4 shows the general workstation characteristics. The indicator for workstation size was the
square root of the workstation area. Areas were calculated from the measured data for length and width,
and corrected when the shape was known not to be square or rectangular (a few were triangular). We
chose to convert to the square root for easier comparisons to other COPE project results, where square
workstations were studied and results reported according to their length. For partition height, we took the
height of the lowest side on which there was a partition on the basis that this was the most conservative
estimate of enclosure that might affect privacy. We excluded open sides in this determination because all
workstations were open on at least one side to provide an entrance, and because the degree of enclosure
was separately captured. Figure 1 shows the histograms for workstation area and partition height, which
were the principal variables of interest for the COPE project.
The median value of 8.70 ft for the square root of workstation area converts to approximately 76
square feet per workstation, which is within the range reported in a recent IFMA survey . In that survey,
professional staff averaged 79 sf, senior clerical staff 77 sf, and general clerical staff 66 sf (our value here
is across all job types).
IRC RR-154 11
Workstation and Physical Condition Effects on Environmental Satisfaction
Table 4. Full sample descriptive statistics for workstation characteristics.
Variable Definition unit N Mean SD Median Minimum Maximu
m
SQRTAREA workstation area
(L*W)
ft 779 8.90 2.06 8.70 3.51 15.83
MINPH_NOOPE
N
Minimum partition
height, excluding open
sides
in 779 60.84 9.82 64.00 30.00 109.00
N # = 0 # = 1 # = 2
PANELS_CAT 1 = not fully enclosed
2 = enclosed except for
entrance
779 N/A 203 576
NO_DL_WI 0 = no daylight (more
than 15 ft / 5 m from
window)
1 = daylight available
(within 15 ft / 5 m of
window), but no
window
2 = window in cubicle
779 330 131 318
WINDOW 0 = no window
1 = window in
workstation
779 461 318 NA
Figure 1. Histograms for workstation area (SQRTAREA) and partition height (MINPH_NOOPEN).
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Square Root of WS Area
0
100
200
300
Count
0.0
0.1
0.2
0.3
Proportion per Bar
30 38 46 54 62 70 78 86 94 102110
Minimum Partition Height
0
50
100
150
200
250
Count
0.0
0.1
0.2
0.3
Proportion per Bar
Table 5 shows the acoustic conditions. The acoustic variables used both daytime sound level
measurements (excluding speech) and night-time measurements of sound propagation (cf. Veitch et al.,
2002). For these analyses, we used the Speech Intelligibility Index calculation with the assumption of
“normal” speech levels (American National Standards Institute, 1997). There is some evidence that actual
speech in open-plan workstations is quieter than this (Warnock & Chu, 2002), in which case the true SII
values would be lower. The values reported here may therefore be viewed as the worst-case scenario. The
mean value for SII indicates that speech intelligibility is quite high; however, overall noise levels are in
the range of desired conditions, as determined in a recent literature review (Navai & Veitch, 2003).
We also examined a new indicator of acoustic conditions, the difference between the low-
frequency and high-frequency components of ambient noise. This characteristic was a good predictor of
IRC RR-154 12
Workstation and Physical Condition Effects on Environmental Satisfaction
acoustic satisfaction in a COPE laboratory experiment (Veitch, Bradley, Legault, Norcross, & Svec,
2002) and we wished to examine its distribution and effects in the field.
Table 5. Full sample descriptive statistics for acoustic conditions.
Variable Definition unit N Mean SD Median Minimu
m
Maximu
m
LNOISEA A-weighted ambient
sound level during
working hours
dB(A
)
734 46.43 3.77 46.66 36.24 59.87
SII Speech Intelligibility
Index (American
National Standards
Institute, 1997),
calculated using
‘normal speech’,
measured sound
propagation, and
daytime ambient sound
level
Ratio,
0 - 1
734 0.51 0.15 0.51 0.00 0.91
LOHI_DBA Difference between the
A-weighted level of the
low frequency sounds
Low(A) (16 - 500 Hz)
and the A-weighted
level of the higher
frequency sounds
High(A) (1000 - 8000
Hz) (Veitch et al.,
2002)
dB(A
)
779 1.96 3.29 1.91 -12.54 13.29
Table 6 shows the lighting conditions observed in this sample. For the lighting conditions there
was one subjectively scored variable, VDT_CAT. Three independent raters viewed photos of the VDT
screen in each cubicle and judged the degree to which the photo showed reflected images of luminaires.
The standards for “low”, “medium”, and “high” had been produced using computer simulations of
lighting installations in open-plan offices in a separate COPE task (Newsham & Sander, 2003). Interrater
agreement was not as good as hoped, with 3-way agreement on only 49% of cases , an average correlation
between raters of r=.72, and kappa values between any two raters in the range 0.42 - 0.50. However,
Cronbach’s alpha (using each rater’s score as an item, and each workstation as a subject) was very good,
being equal to 0.88. Therefore, VDT_CAT scores were created by averaging the three ratings for each
workstation and binning into three categories representing the low, middle, and highest thirds.
We selected three lighting characteristics for inclusion in the regression analyses. The average
illuminance reaching the eye from all directions (called CUBEDAYT here) was selected as the
illuminance value because it was the most consistent measurement, being affixed to the data-collection
chair; desktop measurements proved to be less reliable because physical constraints or operator error led
to variation in where the sensors were placed. There are no standards for desirable illuminance at the eye,
but examination of the entire data set showed that the mean desktop illuminance was 362 lx (SD - 159)
for workstations with no window, which is within recommendations for VDT offices (Illuminating
Engineering Society of North America (IESNA), 1993). Depending on daylight and blind conditions,
windowed workstations had desktop illuminances as high as 6700 lx. There are also no exact equivalents
for our measure of desktop uniformity (UNIFDAYT) or directionality (EH2V). However, recommended
practice is for fairly high desktop illuminance uniformity (Chartered Institution of Building Services
Engineers (CIBSE), 1994; IESNA, 1993).
IRC RR-154 13
Workstation and Physical Condition Effects on Environmental Satisfaction
Table 6. Full sample descriptive statistics for lighting conditions.
Variable Definition unit N Mean SD Median Minimum Maximu
m
CUBEDAYT Average
illuminance on 6
faces of a cube in
location of head of
seated occupant
lux 779 261.81 251.93 202.60 9.20 3655.90
UNIFDAYT (Maximum desktop
illuminance over 4
locations -
minimum desktop
illuminance) /
Maximum desktop
illuminance
Ratio 0 -
1 (lower
values
more
uniform
)
779 0.44 0.20 0.41 0.01 1.00
EH2V Ratio of
illuminance on top
of cube to average
vertical on 4 sides
Ratio 779 2.34 0.92 2.25 0.37 12.16
N # = 1 # = 2 # = 3
VDT_CAT Categorical rating
of degree of
luminaire
reflections in VDT
screen photo.
773 336 163 274
Table 7 shows the descriptive statistics for ventilation and thermal variables. Although we
measured ventilation and thermal conditions at three heights, we used only the head-height measurements
for further analyses. As expected, the measurements at the three heights were highly intercorrelated. It
seemed likely that draught might be most problematic at the head because the majority of the offices used
ceiling air diffusers, hence the choice of air velocity at that location. We chose the temperature
measurement at that location to be consistent. Relative humidity was only measured at torso height. For
indoor air quality we used carbon dioxide as an indicator of ventilation system activity, and created a new
variable to indicate the total concentration of other pollutants. The new variable is the sum of
standardized scores for three individual measurements; standardized scores were chosen because although
each may be reported in parts per million, they differed widely in their expected distributions and in the
levels at which each might be considered problematic. This new variable showed acceptable internal
consistency reliability (Cronbach’s alpha = 0.64). The values of air movement, relative humidity,
temperature and carbon dioxide concentrations were all within recommended levels (American Society of
Heating Refrigerating and Air Conditioning Engineers (ASHRAE), 1992; ASHRAE, 2001).
IRC RR-154 14
Workstation and Physical Condition Effects on Environmental Satisfaction
Table 7. Full sample descriptive statistics for ventilation conditions.
Variable Definition unit N Mean SD Median Minimum Maximu
m
AIR_V_H Air velocity at head
height of seated
occupant (an indicator
of draught)
m/s 779 0.10 0.05 0.08 0.01 0.43
RTD_H Air temperature at
head height of seated
occupant
°C 779 23.27 0.95 23.27 20.39 28.71
REL_HUMID Relative humidity,
measured at torso of
seated occupant
% 779 29.91 10.78 28.70 13.05 58.82
FDCO2 Carbon dioxide
concentration
ppm 779 648.37 97.28 639.79 469.51 1103.90
POLLUT Summed standardized
score of 3 pollutants:
carbon monoxide,
total hydrocarbons,
and methane
779 0.00 1.78 -0.11 -4.70 8.05
N # = 0 # = 1
DL_OUT Air diffuser location
0 = in workstation
1 = outside
workstation
777 590 187
Table 8 contains intercorrelations between all the independent
variables, using pairwise deletion of missing data. With one exception,
there is no evidence of multicollinearity. The exception is the correlation of .61 between SQRTAREA and
MINPH_NOOPEN, which approaches the level at which statistical problems may arise (Tabachnick &
Fidell, 2001). The correlation indicates that larger workstations tend to have higher partitions. Although it
might be considered an artifact of the way in which we selected buildings; we sought buildings with
smaller workstations and shorter partitions to expand the original 3-building data set reported by Charles
and Veitch (2002), it could also reflect a real condition in workplaces. Indeed, concerns expressed to us
by design professionals about the shift to smaller workstations and lower partitions was an initial impetus
behind the development of the COPE project. It seems likely that the incidence of small workstations with
tall partitions is relatively rare in open-plan offices generally.
3.1.3 Intercorrelations.
IRC RR-154 15
Workstation and Physical Condition Effects on Environmental Satisfaction
Table 8. Intercorrelations between independent variables
AGE_COMBINED
GENDER
ADMIN
MGR
PROF
SQRTAREA
MINPH_NOOPEN
PANELS_CAT
NO_DL_WI
WINDOW
LNOISEA
SII
LOHI_DBA
CUBEDAYT
UNIFDAYT
EH2V
VDT_CAT
AIR_V_H
RTD_H
REL_HUMID
FDCO2
POLLUT
DL_OUT
AGE_COMBINED 1.00
GENDER 0.04 1.00
ADMIN
0.04 -0.37 1.00
MGR 0.03 0.10 -0.19 1.00
PROF 0.09 0.07 -0.49 -0.25 1.00
SQRTAREA 0.27 -0.01 0.13 -0.01 0.14 1.00
MINPH_NOOPEN
0.08 -0.01 0.06 -0.07 0.13 0.61 1.00
PANELS_CAT
0.20 -0.07 0.11 -0.04 0.13 0.49 0.36 1.00
NO_DL_WI 0.18 0.02 0.02 0.10 0.01 0.31 0.06 0.08 1.00
WINDOW 0.20 0.03 0.04 0.08 0.03 0.41 0.15 0.11 0.92 1.00
LNOISEA -0.15 0.00 0.02 0.14 -0.20 -0.36 -0.35 -0.21 0.07 0.03 1.00
SII 0.04 0.05 -0.09 -0.04 0.09 -0.06 -0.11 -0.29 0.00 0.01 -0.58 1.00
LOHI_DBA 0.06 -0.01 -0.06 -0.06 0.19 0.24 0.14 0.20 -0.11 -0.08 -0.38 0.38 1.00
CUBEDAYT 0.09 0.04 0.03 0.02 0.06 0.14 -0.06 0.03 0.37 0.37 0.01 0.06 0.01 1.00
UNIFDAYT
0.06 -0.06 0.13 -0.01 -0.06 0.02 0.02 0.06 0.08 0.11 0.05 -0.03 -0.04 0.06 1.00
EH2V -0.07 0.05 -0.03 -0.11 0.04 0.03 0.21 0.08 -0.37 -0.34 -0.12 -0.03 -0.01 -0.13 -0.28 1.00
VDT_CAT 0.02 0.01 0.00 0.00 0.00 0.09 0.06 0.04 0.00 -0.01 -0.01 -0.04 -0.08 -0.04 -0.14 0.00 1.00
AIR_V_H -0.08 0.14 0.00 0.04 -0.11 -0.16 -0.23 -0.08 -0.01 -0.03 0.17 -0.03 -0.07 0.06 0.07 -0.10 0.00 1.00
RTD_H -0.10
0.12 -0.14 -0.06 0.01 -0.33 -0.30 -0.23 -0.15 -0.18 0.12 0.13 -0.03 0.11 -0.09 0.01 -0.07 0.15 1.00
REL_HUMID
0.06 -0.01 0.15 0.01 -0.13 -0.19 -0.34 -0.03 -0.01 -0.02 0.06 0.11 -0.03 0.00 0.19 -0.13 -0.06 0.20 -0.06 1.00
FDCO2 0.04 0.06 -0.08-0.04 0.07 -0.05 0.08 -0.01 -0.05 -0.06 -0.17 0.11 -0.13 -0.02-0.06 0.07 0.02 -0.04 0.13 0.03 1.00
POLLUT 0.09 -0.03 0.14 -0.08 0.00 0.10 -0.01 0.11 -0.08 -0.03 -0.17 0.13 0.13 -0.01 0.13 -0.03 -0.07 0.02 -0.09 0.58 0.09 1.00
DL_OUT -0.09 -0.03 -0.07 0.11 -0.06 -0.41 -0.34 -0.18 -0.15 -0.19 0.21 0.05 -0.12 -0.05 -0.03 -0.10 -0.06 0.04 0.25 0.11 0.00 0.09 1.00
IRC RR-154 16
Workstation and Physical Condition Effects on Environmental Satisfaction
Another area of high intercorrelation is between LNOISEA and SII. This is not surprising;
LNOISEA is an input to SII. Louder ambient noise can mask speech sounds.
The bivariate correlations between the independent variables and the dependent variables are
shown in Table 9, also using pairwise deletion of missing data. These are low, which suggests that effect
sizes are likely to be small. Intercorrelations between the dependent variables are not shown, as they are
reported and analyzed in more detail elsewhere (Charles et al., 2003).
Table 9. Correlations of independent variables (rows) with dependent variables (columns)
SAT_PRI
V
SAT_LIGH
T
SAT_VEN
TOESJOBSATIS
AGE_COMBINE
D -0.09 0.03 -0.05 -0.04 -0.09
GENDER 0.03 0.03 0.21 0.02 0.00
ADMIN 0.07 0.05 -0.13 0.09 -0.06
MGR -0.02 0.03 0.03 -0.03 0.02
PROF -0.03 -0.03 0.04 -0.10 -0.01
SQRTAREA 0.12 0.07 -0.17 -0.01 -0.12
MINPH_NOOPEN 0.08 -0.05 -0.20 -0.09 -0.14
PANELS_CAT 0.03 -0.01 -0.13 -0.04 -0.12
NO_DL_WI 0.01 0.28 -0.08 0.04 0.03
WINDOW -0.01 0.26 -0.12 -0.01 0.00
LNOISEA 0.00 -0.05 0.06 0.03 0.14
SII -0.08 0.05 0.03 0.01 -0.03
LOHI_DBA -0.03 -0.03 0.02 0.02 -0.04
CUBEDAYT 0.01 0.08 -0.02 0.06 0.00
UNIFDAYT -0.05 -0.07 -0.03 -0.02 0.01
EH2V 0.06 -0.10 -0.02 -0.03 -0.04
VDT_CAT 0.04 -0.08 0.00 -0.01 -0.03
AIR_V_H 0.01 0.00 -0.03 0.00 0.03
RTD_H -0.06 -0.10 -0.05 -0.05 0.07
REL_HUMID 0.02 0.12 0.10 0.09 0.00
FDCO2 -0.07 -0.06 -0.12 -0.06 -0.12
POLLUT 0.01 0.03 -0.02 0.03 -0.11
DL_OUT -0.14 -0.12 0.02 -0.09 0.00
3.2 Analytic Strategy
This investigation was a cross-sectional field survey. We sought in the analyses reported here to
relate the measured physical conditions with the satisfaction of the occupants with those conditions. A
separate report describes analyses involving only the physical conditions (Newsham et al., 2003a), and
another reports a general model of the relationships between the questionnaire variables alone (Charles et
al., 2003).
The general approach taken is hierarchical linear regression and follows generally accepted
practices within the behavioural sciences, as described in standard works such as those by Kerlinger and
Lee (2000), Pedhazur (1997), and Tabachnick and Fidell (2001). This section describes the criteria
applied to all analyses reported here.
We examined all of the data carefully for inconsistencies and errors in
data entry, and corrected these where possible, leaving data as missing if
there were any question.
3.2.1 Data cleaning.
IRC RR-154 17
Workstation and Physical Condition Effects on Environmental Satisfaction
For the dependent variables there was very little missing data, and no evidence of any systematic
missing data. We calculated scale scores as the average of available data on the contributing items, but
required valid data on more than 50% of the contributing items for the scale score to be valid. Otherwise,
the scale score was set to missing. This criterion resulted in four missing cases for Sat_Priv, four for
Sat_Vent, three for Sat_Light, 34 for OES, and 12 for JobSatis.
We tested each dependent and independent variable for normality. Following recommendations
by Kline (1997), we looked for skewness values between +3 and -3, and kurtosis values between +8 and -
8. All the variables met these criteria.
We further examined the data for univariate and multivariate outliers. Univariate outliers were
defined as cases on which the absolute value of the standardized score for that variable was greater than 3.
These cases were omitted from analysis.
Multivariate outliers were examined for each analysis. We first ran the analysis with all cases
except for the univariate outliers. We then examined the Mahalanobis distance statistic for each case.
Very large values of this statistic indicate that the case is an extreme outlier and probably is having an
undue effect on the outcome. Cases for which the Mahalanobis distance exceeded the critical value for
that analysis were identified as multivariate outliers and excluded from analysis. (Mahalanobis distance is
distributed as a chi-square and is tested against the degrees of freedom, which is the number of predictor
variables in the model. We tested against a very conservative alpha of p<=.001.) For most analyses there
were no multivariate outliers, and there were never more than two. We did not look for further
multivariate outliers after the first exclusion.
The results presented below are for the final sample, excluding both univariate and multivariate
outliers. Because outliers were determined separately for each analysis, sample sizes vary somewhat from
one analysis to another. We chose this approach to preserve as large a sample size as possible for each
analysis.
The buildings were not randomly sampled; rather they
were selected based on the willingness of management to
provide access and on the availability of a suitable number of open-plan workstations. Later buildings
were selected deliberately to ensure a broader range of workstation sizes in the overall sample. Moreover,
although workstation size tends to vary within a building it is often the case that organizations use a
limited range of workstation sizes and partition heights, so that the range within any one building is
limited. The sample therefore has the possibility of being biased by the selection of certain organizations
or certain buildings and by the confound of buildings and workstation characteristics. 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.
3.2.2 Independence of observations..
Although the remedies for these problems are few, we did conduct a series of statistical analyses
to determine the legitimacy of combing individual data from the buildings into one large sample in which
we ignored the building as a variable. These tests followed the guidance of Dansereau, Alutto, and
Yammarino (1984)and Yammarino and Markham (1992) regarding independence of observations.
Four statistical criteria were used to examine the agreement among occupants in a building
regarding the five satisfaction measurements. Traditional one-way ANOVAs were conducted in which
building served as the independent variable and the five satisfaction scales served as the dependent
variables. Then, using the information produced in the ANOVA (i.e., sums-of-squares and mean square),
other relevant statistics were calculated (Table 10). All cases were used in these analyses, although there
were 7 cases with missing data on JobSatis, 6 missing on OES, and three each on Sat_Priv, Sat-Light, and
Sat_Vent.
The Intraclass Correlation Coefficient 1, or ICC(1), assesses whether occupants in the same
building reliably agreed in their responses. The ICC(1) has values that range from .00 to .50, with a
IRC RR-154 18
Workstation and Physical Condition Effects on Environmental Satisfaction
median of .12 (James, 1982). An ICC(1) value of .12 or greater indicates reliable agreement. By this
criterion, all of the satisfaction scales showed within-building agreement, as all had ICC(1) values >= .12.
The ICC(2) measures how reliably buildings can be differentiated based on satisfaction scores
(Bartko, 1976). The closer ICC(2) is to 1.00, the better the measurement indicates whether the buildings
can be reliably differentiated in terms of individual responses on the five satisfaction scales. The criterion
of .85 or higher was used as an indicator of between-building differences, following Griffith (1997). Only
one of the five scales passed this criterion, with Sat_Vent having ICC(2) = .861.
The E-tests (a ratio of the between-eta and within-eta) of practical significance provides an index
of the magnitude of the effects (within- and between-group analysis - WABA) (Yammarino & Markham,
1992). The test of significance for the E-Test is not dependent on degrees of freedom but is geometrically
based. Briefly, a 900 angle representing the relationship between two sets of scores indicates that the
scores are orthogonal. The smaller the angle becomes, the stronger the observed relationship. In WABA,
angles are considered for between versus within etas. The difference between the pair of angles in each
case is what is tested. Thus, the larger the angular difference, the more likely that the etas are significantly
different. We used the critical value for the most conservative, 300 test, E <= 0.58 (Dansereau et al.,
1984). On this test all five satisfaction scales showed practical significance, meaning that the within-
groups variance is greater than the between-groups variance. This is an indicator that observations are
independent of building.
Two F tests examine the statistical significance of the between-groups versus within-groups
variance. The traditional F test compares between-group variance/within-group variance, to determine
whether there are meaningful differences between buildings on the variable of interest. All five traditional
F tests are statistically significant, indicating that there are between-buildings differences in all of the
satisfaction scales. However, in cases in which the within-groups eta is the larger of the two, as in the
present results (as shown by the E tests), a corrected F test is the appropriate indicator of the significance
of the within-groups effect (Dansereau et al., 1984). The test is the inverse of the traditional F test. For the
present results, three of the corrected F tests are statistically significant, suggesting that there are
differences between buildings in the amount of within-groups variability on these three variables (OES,
Sat_Priv, and Sat_Light).
Table 10. Summary of independence analyses.
Variable ICC(1) ICC(2) etabn eta2
bn etawn eta2
wn E Test**
Traditional
F Test
Corrected
F Test
SAT_PRIV 0.270† 0.769 0.208 0.043 0.978 0.957 0.212 4.326* 0.231*
SAT_VENT 0.408† 0.861‡ 0.264 0.070 0.964 0.930 0.274 7.195* 0.139
SAT_LIGHT 0.202† 0.695 0.182 0.033 0.983 0.967 0.185 3.280* 0.305*
OES 0.240† 0.740 0.197 0.039 0.980 0.961 0.200 3.840* 0.260*
JOBSATIS 0.308† 0.800 0.223 0.050 0.975 0.950 0.229 5.008* 0.200
Note. † ICC(1) >= .12. ‡ ICC(2) >= .85. ** E <= .58 indicates independence * p <= .05.
Overall, none of the five dependent variables met all four criteria for group effects. Therefore, we
concluded that the assumption that observations were independent of building was met. Further analyses
proceeded by combining all cases in one group, ignoring building effects.
The regression models were hierarchically structured.
For all analyses, demographic control variables were
entered on Step One as a block. These comprised sex (coded 0 or 1 for male or female), age (five
categories entered as a continuous variable), and job type. Job type had four categories entered as three
dummy codes: Admin where 1 = administrative, 0 = other; Prof where 1 = professional, 0 = other; and,
Mgr, where 1 = managerial and 0 = other. The fourth category was technical.
3.2.3 Hierarchical regression models.
The order of entry of the other variables in each analysis was determined based on theoretical
considerations and on guidance from the literature. Our first interest was in gross descriptors of the
IRC RR-154 19
Workstation and Physical Condition Effects on Environmental Satisfaction
workstation: workstation area, partition height, enclosure, and presence of a window. These are likely to
be the most salient characteristics for occupants. Therefore, we examined the effects of workstation
characteristics on all dependent variables.
The workstation characteristics were correlated with physical conditions. These relationships
were examined separately and reported by Newsham et al. (2003a). For the regressions on satisfaction
outcomes, we decided to control first for these workstation characteristics before adding measured
physical conditions to the models. Thus, we controlled for the most salient workstation characteristics
before looking to see whether the physical conditions themselves predicted additional variance. The
models differed for each subset of environmental satisfaction. Thus, for Sat_Priv we looked for additional
variance explained by acoustic conditions. For Sat_Light we looked at lighting conditions. For Sat_Vent
we looked at ventilation and IAQ conditions. We looked at all of the physical condition models as
predictors of OES and JobSatis.
3.3 Predicting Satisfaction with Privacy
Satisfaction with privacy was first regressed on workstation
characteristics. For this analysis the sample size was 757
after removing cases with missing data and univariate outliers. After the control variables, workstation
area (SQRTAREA) was entered as the next step, followed by enclosure (MONPH_NOOPEN and
PANELS_CAT). WINDOW was the final step. Table 11 summarizes the result of this hierarchical
regression.
3.3.1 Workstation characteristics.
Table 11. Summary table for Sat_Priv regressed on workstation characteristics.
AGE_COMBINE
D
-.107** -.139*** -.135*** -.131***
GENDER .066 .057 .055 .056
ADMIN .127** .081 .084 .084
MGR .009 -.008 -.007 -.004
PROF .043 -.002 .000 -.001
SQRTAREA .145*** .149** .178**
MINPH_NOOPEN .014 .007
PANELS_CAT -.029 -.035
WINDOW -.053
R2 change .020* .018*** .001 .002
Total R2 .020* .038*** .038*** .041***
Adjusted R2 .013* .030*** .028*** .029***
Note. N=757. * p<=.05. **p<=.01. ***p<=.001.
The overall model was statistically significant at all steps. Of the control variables, age persisted
as a significant predictor, with younger people reporting greater satisfaction with privacy. In the first step,
administrators also showed greater satisfaction with privacy, but this variable dropped out when the
workstation characteristics were added. Workstation area was a significant predictor in all steps and
uniquely explained 1.8% of the variance. As expected, larger workstations predicted greater satisfaction
with privacy. Neither of the enclosure variables predicted satisfaction with privacy, but it would be
premature to conclude that enclosure is not important because of their high correlation with workstation
area (Table 8). Area is a strong predictor in this equation and could be carrying the variance for both the
size and degree of enclosure.
We next looked for additional predictive power from the three
physical aspects of the acoustic environment: ambient noise
(LNOISEA), speech intelligibility (SII), and the relative spectral properties of the ambient noise
3.3.2 Acoustic conditions.
IRC RR-154 20
Workstation and Physical Condition Effects on Environmental Satisfaction
(LOHI_DBA).
For LNOISEA and LOHI_DBA, we thought it possible that the relationships might take a
quadratic rather than a linear shape, with an intermediate value being optimal (i.e., neither too loud nor
too soft, neither too rumbly nor too hissy). However, individual regressions of Sat_Priv against the
quadratic shape for either LNOISEA or LOHI_DBA (always controlling for the five demographic
variables first), showed no evidence of a quadratic trend. Therefore we proceeded with linear terms only.
The model entered SII first, after the control variables and the workstation characteristics.
Conversations from others are among the most frequent noise-related complaints in open-plan offices
(Navai & Veitch, 2003), and we expected this to be the strongest predictor. The overall level of
background noise followed, and lastly the indicator of its spectral properties. The sample size was 694
after removing cases with missing data and outliers; as noted above, some of the acoustic variables had a
large amount of missing data. Table 12 summarizes the result.
Table 12. Summary table for Sat_Priv regressed on workstation characteristics and acoustic conditions.
 
AGE_COMBINE
D
-.111** -.135*** -.128*** -.126** -.128***
GENDER .072 .063 .065 .066 .065
ADMIN .152** .108* .103* .103 .102
MGR .032 .020 .015 .011 .010
PROF .067 .023 .028 .029 .031
SQRTAREA .175** .184** .194** .207***
MINPH_NOOPEN .012 .002 .015 .017
PANELS_CAT -.023 -.047 -.036 -.022
WINDOW -.072 -.070 -.080 -.091*
SII -.071 -.038 -.012
LNOISEA .046 .055
LOHI_DBA -.046
R2 change .022** .022** .004 .001 .001
Total R2 .022** .044*** .048*** .049*** .050***
Adjusted R2 .015** .031*** .034*** .034*** .034***
Note. N=694. * p<=.05. **p<=.01. ***p<=.001.
The final model was statistically significant and explained more variance than the model with
workstation characteristics only. Age and workstation area remained statistically significant predictors. In
this model, with acoustic conditions controlled, presence of a window also significantly predicted
satisfaction with privacy, in an inverse direction: those with windows were less satisfied. The acoustic
conditions themselves were not statistically significant predictors of satisfaction with privacy.
In our data set, satisfaction with privacy was
influenced by age, workstation area, and the
presence of a window. Overall, only 5% of the variance in satisfaction with privacy was explained, which
is a small effect size (Cohen, 1988).
3.3.3 Discussion: Satisfaction with privacy.
Having a window reduced satisfaction with privacy to a small degree. This relationship emerged
only when acoustic predictors were added to the model. It is possible that visual privacy might be
compromised by a window, depending on the surroundings and the availability of blinds to control the
view in. Alternatively, it is possible that a window provides a hard reflective surface that could allow
more speech transmission from one workstation to another, thereby reducing speech privacy. This is
consistent with the physics of sound transmission, but we know of no other satisfaction study to report
such an effect.
It was not surprising to find that workstation area significantly predicted satisfaction with
privacy. Larger workstations place occupants farther apart, reducing the number of people available to
IRC RR-154 21
Workstation and Physical Condition Effects on Environmental Satisfaction
overhear conversations and the number of sources of unwanted sound. The finding is consistent with
other investigations, in which workplace satisfaction was greater when workstation areas were larger
(Brill et al., 1984; O'Neill & Carayon, 1993; Sundstrom, Town, Brown, Forman, & McGee, 1982; Sutton
& Rafaeli, 1987). Other investigations have reported separately for different job types, finding differences
between them; our regression model controls for job type, yet finds the relationship nonetheless.
However, we did not find that enclosure, whether in the form of partition height or number of
panels, predicted satisfaction with privacy. In this we failed to replicate previous findings that have
focused on privacy as an outcome (O'Neill & Carayon, 1993; Sundstrom et al., 1980; Sundstrom et al.,
1982). Two reasons might explain this. First, the high correlation between workstation area and partition
height probably obscured the relationship. Second, there was little variance in the two-level categorical
variable for degree of enclosure (PANELS_CAT). However, it is also possible that any partition that is
not a full wall, regardless of its height, has the same effect on satisfaction with privacy (Kupritz, 2003a).
The finding that age negatively predicted satisfaction with privacy is an intriguing one with
parallels in other research using qualitative methods. Kupritz (2003a) found that older workers associated
having a larger office with talking privately with people, whereas younger workers did not. They also
associated office location with minimizing disruptions. Privacy needs appeared to change with age,
although experience rather than age per se might provide the better explanation. Given the prevalence of
open-plan offices, it is possible that younger employees have had less exposure to more enclosed
workplaces than older ones, and have therefore formed difference associations and expectations.
3.4 Predicting Satisfaction with Lighting
For this analysis, the sample size was 758 cases after
excluding cases with missing data and outliers. Table 13
shows the result of the analysis.
3.4.1 Workstation characteristics.
Table 13. Summary table for Sat_Light regressed on workstation characteristics.
AGE_COMBINE
D
.017 .004 -.003 -.027
GENDER .055 .051 .048 .043
ADMIN .086 .068 .067 .069
MGR .045 .038 .030 .015
PROF .018 .000 .006 .009
SQRTAREA .059 .163** .021
MINPH_NOOPEN -.135** -.097*
PANELS_CAT -.034 -.001
WINDOW .259***
R2 change .007 .003 .012** .053***
Total R2 .007 .010 .022* .075***
Adjusted R2 .000 .002 .012* .064***
Note. N=758. * p<=.05. **p<=.01. ***p<=.001.
The final model was statistically significant and explained 7.5% of the variance, which is a small-
to-medium effect size (cf. Cohen, 1988). At the final step, with WINDOW added, two predictors were
statistically significant: WINDOW and MINPH_NOOPEN. Both were in the expected directions: People
with a window were more satisfied with their lighting, and people with lower partitions were more
satisfied. Lower partition heights allow more daylight penetration to interior workstations (Reinhart,
2002) and improve electric light distribution (Newsham & Sander, 2003), so these effects are internally
consistent.
At the intermediate step, before the addition of WINDOW, both workstation area and partition
height were statistically significant predictors. Workstation area dropped out at the final step, and the
IRC RR-154 22
Workstation and Physical Condition Effects on Environmental Satisfaction
regression weight for partition height also became smaller. These changes reflect the power of the
WINDOW variable and the intercorrelations between the three variables. Workstation area is correlated
with both partition height and presence of a window. Presence of a window has the strongest relation to
satisfaction with lighting.
As for the acoustic conditions, we repeated the above analysis with
additional steps for measured lighting conditions. We conducted
three variations. First, we examined the result for all the workstations considered together. In this analysis
we used the variable NO_DL_WI instead of the simple window/no window coding of the WINDOW
variable, to account for variations in the amount of daylight that might reach a workstation adjacent to,
but not having, a window.
3.4.2 Lighting conditions.
Next, we pulled out a subset consisting only of interior workstations in which no daylight could be
present (those more than 15’ or 5 m from a window, NO_DL_WI = 0), and repeated the regression model
(omitting, of course, NO_DL_WI as a predictor). We finally looked at the outcome for those workstations
with either daylight or a window. The purpose of the regressions on the subsets was to determine whether
the effects of various physical conditions model would change for occupants with no window access or
daylight, relative to those with. In addition, the no-daylight model has the closest relation to physical
conditions predicted by the COPE software (Newsham & Sander, 2003).
Table 14 shows the result for the overall regression, with all workstations and controlling for
workstation characteristics. Although other workstation characteristics analyses had WINDOW entered
with workstation size and partition height, in this case we entered it last and use the NO_DL_WI variable
instead. In this analysis we first wanted to see what effect the measured lighting conditions would have,
regardless of whether or not a window or daylight were present. The order of entry was determined on
theoretical grounds. Each variable was a separate step. We considered that reflected images in the VDT
screen might be most detrimental to satisfaction, following results obtained by Veitch and Newsham
(2000). The illumination level, indexed by CUBEDAYT, entered next, as an indicator of the adequacy of
the amount of light available to see. Uniformity was the third lighting variable, its importance being
reflected in codes and standards (e.g., CIBSE, 1994). Directionality was the fourth variable, added
because previous NRC research has suggested that it influences satisfaction with lighting (Newsham,
Marchand, Svec, & Veitch, 2002). NO_DL_WI was the fifth, and last, lighting variable in this overall
model. (Entering it last also facilitated comparisons with the subsample models, discussed below.)
Table 14. Summary table for Sat_Light regressed on workstation characteristics and lighting conditions.
AGE_COMBINE
D .027 .009 .009 .008 .014 .008 -.009
GENDER .045 .038 .038 .032 .030 .036 .037
ADMIN .083 .065 .063 .055 .063 .067 .082
MGR .037 .023 .022 .016 .017 .009 .002
PROF .000 -.012 -.015 -.026 -.027 -.028 -.011
SQRTAREA
.170*** .177*** .127* .126* .122* .046
MINPH_NOOPEN
-.139** -.135** -.089 -.089 -.068 -.078
PANELS_CAT
-.040 -.042 -.033 -.030 -.024 -.010
VDT_CAT
-.093** -.080* -.092* -.101** -.104**
CUBEDAYT
.123** .119** .086* -.008
UNIFDAYT
-.081* -.113** -.108**
EH2V
-.098* -.013
NO_DL_WI
.281***
R2 change .008 .017** .009** .013** .006* .007* .049***
Total R2 .008 .024* .033** .046*** .052*** .059*** .108***
Adjusted R2 .001 .014* .021** .033*** .038*** .044*** .092***
Note. N = 740. * p<=.05. **p<=.01. ***p<=.001.
IRC RR-154 23
Workstation and Physical Condition Effects on Environmental Satisfaction
The control variables alone (step one) were not significant predictors of satisfaction with lighting,
but every other step achieved statistical significance. Workstation area was a significant predictor until
the last step. Of the lighting variables, each added statistically significant amounts of explained variance
on the step in which they entered. Each appears to have a role in explaining the variance in satisfaction
with lighting. Lower levels of reflected images in computer screens, higher average global light levels,
and greater desktop uniformity are all associated with higher satisfaction with lighting. Lower ratios of
horizontal to vertical illuminance (EH2V) are associated with higher satisfaction (i.e., relatively more
vertical than horizontal). However, at the final step with NO_DL_WI added, only VDT_CAT and
uniformity remain along with NO_DL_WI as significant predictors. This final variable explains the
largest amount of variance and has the largest standardized weight. Having daylight or having a window
each improve satisfaction with lighting. Relatively smaller beneficial effects are associated with lower
VDT glare and greater uniformity (recall that the variable UNIFDAYT is reverse-scored, so that lower
values are more uniform lighting).
Next, we separated the sample into two groups. Descriptive statistics for the three groups on the
variables in this analysis are shown in Table 15. Between-group differences are apparent. Peripheral
workstations are somewhat larger, have higher illuminance levels, and lower ratios of horizontal to
vertical illuminance, than central workstations. The illuminance level and directionality differences are
consistent with having windows providing daylight.
Table 15. Descriptive statistics for full sample and lighting subgroups.
Full Sample Central WS Peripheral WS
M SD N M SD N M SD N
SAT_LIGHT 4.75 1.20 740 4.40 1.16 312 5.03 1.14 427
AGE_COMBINED 2.62 .95 740 2.50 .99 312 2.72 .91 427
GENDER 1.52 .50 740 1.52 .50 312 1.52 .50 427
ADMIN .27 .44 740 .28 .45 312 .27 .44 427
MGR .09 .28 740 .05 .21 312 .11 .32 427
PROF .39 .49 740 .39 .49 312 .38 .49 427
SQRTAREA 8.88 2.02 740 8.55 1.99 312 9.16 1.98 427
MINPH_NOOPEN 60.79 9.46 740 61.38 9.89 312 60.46 9.06 427
PANELS_CAT 1.74 .44 740 1.73 .44 312 1.75 .43 427
VDT_CAT 1.92 .88 740 1.90 .86 312 1.93 .90 427
CUBEDAYT 241.53 149.72 740 168.94 67.58 312 296.76 172.05 427
UNIFDAYT .44 .20 740 .43 .21 312 .44 .19 427
EH2V 2.32 .82 740 2.68 .79 312 2.06 .74 427
NO_DL_WI .98 .91 740
WINDOW .70 .46 427
Note. Central workstations had NO_DL_WI = 0. There were 330 of these in the full COPE sample. Peripheral
workstations had NO_DL_WI = 1 or 2. There were 449 of these in the full COPE sample.
Table 16 reports the result of the regression analysis for the Central workstations. Although the
model was statistically significant for steps 2-6, only one step added significantly to the explained
variance, and only one variable was itself a statistically significant predictor. For those workstations
without any daylight, the ratio of horizontal to vertical illuminance was a significant predictor of
satisfaction with lighting. The direction of the effect differed from the overall analysis. In this case, higher
ratios were more satisfactory, indicating a preference for higher horizontal illuminance than vertical.
IRC RR-154 24
Workstation and Physical Condition Effects on Environmental Satisfaction
Table 16. Central workstations’ summary table for Sat_Light regressed on workstation characteristics and lighting
conditions.
AGE_COMBINE
D -.006 -.010 -.013 -.013 -.007
GENDER .023 .006 .005 .009 -.002
ADMIN .072 .102 .109 .111 .111
MGR .034 .045 .045 .053 .058
PROF -.113 -.060 -.048 -.060
.000
.001
.115
.054
-.057 -.060
SQRTAREA
-.053 -.046 -.068 -.091 -.099
MINPH_NOOPEN
-.146 -.144 -.117 -.113
PANELS_CAT
-.163
.098 .094 .092 .097 .067
-.072 -.087 -.072
VDT_CAT -.078
CUBEDAYT
.092 .056 .006
UNIFDAYT
-.086 -.042
EH2V
.179**
R2 change .027 .023 .005 .008 .005 .020**
2.027 .050* .055* .063* .069* .088**
Adjusted R2 .011 .025* .027* .032* .034* .052**
Total R
Note. N = 312. * p<=.05. **p<=.01. ***p<=.001.
For the Peripheral workstations we repeated the order of entry that was used for the Central
workstations, then entered WINDOW as a final step. This provided a contrast between actually having a
window in the workstation (as was the case for 70% of peripheral workstations), and having daylight but
no window. The results are shown in Table 17.
Table 17. Peripheral workstations’ summary table for Sat_Light regressed on workstation characteristics and
lighting conditions.
AGE_COMBINE
D -.020 -.041 -.035 -.035 -.032 -.037 -.040
GENDER .060 .026 .027 .025 .021 .041 .041
ADMIN .097 .060 .048 .047 .057 .064 .067
MGR .049 .033 .027 .026 .022 .007 .009
PROF .087 .072 .052 .052 .049 .036 .038
SQRTAREA
.196** .200** .196** .215** .209** .174*
MINPH_NOOPEN
-.045 -.042 -.038 -.040 -.019 -.030
PANELS_CAT
-.116* -.113* -.112* -.108 -.110* -.100
VDT_CAT
-.109* -.107* -.121* -.139** -.136**
CUBEDAYT
.012 .018 -.047 -.061
UNIFDAYT
-.121* -.157** -.158**
EH2V
-.149** -.138*
WINDOW
.071
R2 change .007 .023* .012* .000 .013* .015** .003
Total R2 .007 .030 .042* .042 .055* .070** .073**
Adjusted R2 -.005 .012 .021* .019 .030* .043** .044**
Note. N = 427. * p<=.05. **p<=.01. ***p<=.001.
For people with access to daylight or a window, although overall somewhat less variance was
explained the result is more interpretable than for the Central workstations. At the end of the sixth step,
without WINDOW, the model for peripheral workstations is the same as that for central workstations.
Here, workstation size, the number of panels, VDT glare, uniformity and directionality were all
statistically significant predictors. Satisfaction with lighting increased with larger workstations, fewer
IRC RR-154 25
Workstation and Physical Condition Effects on Environmental Satisfaction
panels, lower VDT glare, greater uniformity, and lower horizontal-to-vertical illuminance ratios. The
addition of the WINDOW variable on the following step did not add significantly to the explained
variance and changed the pattern of predictor significance only for one variable: the number of panels was
no longer statistically significant. This suggests that for people with some daylight, it is the physical
properties of the luminous environment that principally influence satisfaction with lighting, rather than
the qualities that are specific to the window, such as the view of outside that it affords.
The effect size for satisfaction with lighting falls
between the ranges of small and medium-sized
effects, with 10.8% of the variance explained when all workstations were considered. Slightly less
variance was explained for the central (8.8%) and peripheral (7.3%) workstations considered separately.
3.4.3 Discussion: Satisfaction with lighting.
The dominant finding here is the importance of a window to satisfaction with lighting. In the
workstation characteristics regression, presence of a window accounted for 5% of the variance in
satisfaction with lighting, which is half of the total explained. In the form of a continuous variable that
included the availability of daylight, it accounted for 5% over and above other workstation characteristics
and physical measurements of lighting conditions (in the full sample regression with lighting
characteristics). Having a window, or having access to daylight, improves satisfaction with lighting. Other
researchers, with other dependent measures, have also found that windows are desirable to occupants
(e.g., Finnegan & Solomon, 1981; Heerwagen & Heerwagen, 1986), and that people believe that working
under natural daylight is better for health and well-being than electric light (Veitch & Gifford, 1996;
Veitch, Hine, & Gifford, 1993).
Satisfaction with lighting was also a function of reflected images in VDT screens (higher values
being worse), which is what lighting research and common sense would both predict (Veitch &
Newsham, 1998; Veitch & Newsham, 2000). Its failure to predict satisfaction with lighting for the central
workstations might have been caused by the lower incidence of high-glare workstations in that subsample
(see Table 15). For both the peripheral workstations and the full sample, this categorical variable was an
important predictor, explaining over 1% of the variance (out of a total of 7.3% for the peripheral
workstations, and 10.8% for the full sample).
Uniformity predicted satisfaction with lighting for the full sample and the peripheral
workstations; people preferred more uniformity. This might reflect a desire among those with daylight to
avoid very nonuniform areas or very high contrasts between direct sunlight and shadow. We know of no
studies of desktop uniformity in windowed spaces. However, Bernecker, Davis, Webster and Webster
(1993) found that the luminance of horizontal and vertical surfaces, rather than desktop uniformity,
predicted visual comfort, a variable that would be expected to correlate highly with satisfaction. Boyce
and Slater (1990) found that few people found a nonuniform desk surface to be unacceptable. Uniformity
did not predict satisfaction for the central workstations in the present study, although that might have been
related to the smaller sample size.
The finding that directionality expressed as the ratio of horizontal to vertical illuminance
predicted satisfaction with lighting is new; to our knowledge only one report, a pilot study, has previously
used this ratio (Newsham et al., 2002). The change in direction from peripheral to central workstations is
very intriguing. It appears that for central workstations, satisfaction increases as the horizontal component
increases; whereas for peripheral workstations satisfaction increases as the vertical component increases.
It might be the case that when daylight is available through a window (a vertical source for all of our
buildings), people prefer that as the principal light source. It is unclear why the preferred directionality
would change for windowless workstations, unless it is the case that when the light source is more
directly down there is less possibility of reflections in the VDT screen.
For peripheral workstations only, workstation area was positively related to satisfaction with
lighting. Perhaps a larger workstation also means a larger window, which some have found to be
preferred for lighting and view (Cuttle, 1983; Keighley, 1973a, 1973b; Roche, Dewey, & Littlefair,
2000).
IRC RR-154 26
Workstation and Physical Condition Effects on Environmental Satisfaction
3.5 Predicting Satisfaction with Ventilation
As for satisfaction with privacy and satisfaction with
lighting, we first examined the effects of workstation
characteristics on satisfaction with ventilation. The results of this regression analysis are shown in Table
18. The sample size was 757 after cases having missing data and outlying cases were excluded. The
model was statistically significant at all steps, and the overall percentage of variance showed a medium-
sized effect of 9.6% explained variance. Although workstation size contributed significantly when it was
entered on the second step, it was not statistically significant in the final model. This indicates that
partition height, to which it was strongly correlated, is the more important predictor of satisfaction with
ventilation. Higher partitions are associated with lower satisfaction with ventilation, as is the presence of
a window. Women tend to have lower satisfaction with ventilation than men.
3.5.1 Workstation characteristics.
Table 18. Summary table for Sat_Vent regressed on workstation characteristics.
AGE_COMBINE
D
-.054 -.017 -.026 -.017
GENDER .192*** .204*** .202*** .203***
ADMIN -.057 -.004 -.006 -.006
MGR .005 .026 .017 .023
PROF .004 .058 .062 .061
SQRTAREA -.171*** -.068 -.018
MINPH_NOOPEN -.142** -.156***
PANELS_CAT -.020 -.032
WINDOW -.093*
R2 change .051*** .025*** .012** .007*
Total R2 .051*** .077*** .089*** .096***
Adjusted R2 .045*** .069*** .079*** .085***
Note. N=757. * p<=.05. **p<=.01. ***p<=.001.
We report here our tests of the additional variance explained
by ventilation conditions using the measured values. We also
converted the physical conditions to derived indices for thermal comfort and draught, and report that
regression in Appendix A. Table 19 shows the regression results for ventilation and IAQ conditions added
after workstation characteristics. This analysis had 721 cases after outliers and cases having missing data
were excluded. Only linear effects appear in this model; preliminary regressions of ventilation conditions
alone revealed no evidence of quadratic relations, although these might have been expected.
3.5.2 Ventilation/IAQ conditions.
IRC RR-154 27
Workstation and Physical Condition Effects on Environmental Satisfaction
Table 19. Summary table for Sat_Vent regressed on workstation characteristics and ventilation conditions.
AGE_COMBINE
D
-.050 -.012 -.021 -.021 -.012
GENDER .179*** .191*** .220*** .218*** .228***
ADMIN -.055 .003 -.023 -.023 -.023
MGR .022 .041 .018 .025 .006
PROF .005 .067 .055 .056 .062
SQRTAREA -.032 -.055 -.071 -.066
MINPH_NOOPEN -.161*** -.182*** -.192*** -.168***
PANELS_CAT -.022 -.032 -.030 -.034
WINDOW -.096* -.108** -.111** -.119**
AIR_V_H -.098** -.103** -.111**
RTD_H -.144*** -.139*** -.121**
REL_HUMID .052 .053 .096
DL_OUT -.057 -.062
FDCO2 -.124***
POLLUT -.047
R2 change .046*** .049*** .029*** .003 .016**
Total R2 .046*** .096*** .125*** .127*** .143***
Adjusted R2 .039*** .084*** .110*** .111*** .125***
Note. N=721. * p<=.05. **p<=.01. ***p<=.001.
The results show statistically significant models at all steps, with only diffuser location failing to
add to the explained variance. In the final model, three of the physical conditions were statistically
significant predictors, over and above the three previously identified. As seen above, men were more
satisfied with ventilation conditions than women, and people with lower partitions or without a window
were also more satisfied. In addition, lower air velocity (AIR_V_H) and lower temperatures (RTD_H)
were associated with higher satisfaction with ventilation, as was lower carbon dioxide concentration.
The regressions for satisfaction with
ventilation explained a total of 14.3% of the
variance when both workstation characteristics and physical conditions were in the model, bringing it into
the range of a medium-sized effect. Gender was a significant predictor, with women being less satisfied
than men. This is typical of research in this area (Hedge et al., 1992; Molhave, Jensen, & Larsen, 1991;
O'Neill, 1992; Zweers, Preller, Brunekreef, & Boleij, 1992).
3.5.3 Discussion: Satisfaction with ventilation.
Even controlling for these demographic variables, workstation characteristics were significant
predictors of satisfaction with ventilation. Lower partition heights increased satisfaction with ventilation.
This finding is intriguing, given that engineering research has found no relationship between partition
heights and ventilation effectiveness (Haghighat, Huo, Zhang, & Shaw, 1996; Shaw, MacDonald,
Galasiu, Reardon, & Won, 2003). Perhaps the effect on satisfaction is an inference from what is
observable, rather than a reflection of physical conditions.
The presence of a window also negatively affected satisfaction with ventilation, and this
relationship strengthened (as indicated by a larger weight) when physical conditions, including
temperature, were controlled. Although the temperature at the time of measurement was controlled in the
regression equation, it seems possible that the satisfaction rating would be influenced by the overall
experience of more variable temperatures near a window, with heat gain during times of direct sun and
the possibility of radiant cooling during winter months. This would be consistent with observations in UK
buildings by Roche, Dewey, and Littlefair (2000), who found that the upper limit of acceptable window
size was reached when it led to problems with thermal regulation.
Over and above workstation characteristics, physical conditions predicted satisfaction with
ventilation. Both air movement and temperature showed negative relationships to satisfaction with
IRC RR-154 28
Workstation and Physical Condition Effects on Environmental Satisfaction
ventilation, indicating that people neither want to be too hot nor to experience high air movement. Within
the range of these variables that we observed, these relationships are consistent with expectations
(ASHRAE, 1997; Fanger, 1982). Both variables would be expected to show quadratic relationships with
an optimum middle level, but we appear not to have sampled the full range needed to demonstrate this
shape.
We also found that people are sensitive to carbon dioxide concentration, with satisfaction with
ventilation increasing as carbon dioxide levels dropped. Although the direction of the effect was
predicted, it is most interesting that it was observable at the CO2 concentrations measured here. Current
North American recommendations cite 1000 ppm as the permissible limit, above which satisfaction would
be expected to decline (ASHRAE, 2001), but in our sample we detected evidence of declining satisfaction
even at lower concentrations; the median value in our sample was 639 ppm (Table 7, above).
3.6 Predicting Overall Environmental Satisfaction
We had previously validated a model in which the
individual components of environmental satisfaction
predicted overall environmental satisfaction (OES). This suggests that workstation characteristics and
physical conditions would influence OES indirectly; nonetheless, we also looked for direct effects of
workstation characteristics and physical conditions on OES. The regression model for OES regressed on
workstation characteristics is shown in Table 20. There were 733 cases in this regression, after excluding
cases having missing data and outliers.
3.6.1 Workstation characteristics.
Table 20. Summary table for OES regressed on workstation characteristics.
AGE_COMBINE
D
-.033 -.033 -.040 -.038
GENDER .052 .052 .050 .051
ADMIN .064 .063 .064 .064
MGR -.040 -.040 -.048 -.047
PROF -.080 -.081 -.075 -.075
SQRTAREA .003 .101 .110
MINPH_NOOPEN -.129** -.132**
PANELS_CAT -.026 -.029
WINDOW -.018
R2 change .017* .000 .011* .000
Total R2 .017* .017* .028** .028*
Adjusted R2 .010* .009* .017** .016*
Note. N=733. * p<=.05. **p<=.01. ***p<=.001.
The direct effect of workstation characteristics on OES was small, with only 2.8% of the variance
explained at the final step. Partition height (MINPH_NOOPEN) was the only statistically significant
predictor, and it accounted independently for only 1.1% of the explained variance.
The regression of OES on workstation characteristics and acoustic
conditions is summarized in Table 21. The sample size, after
dropping cases with missing data and outliers, was 671. There was no evidence of direct effects of
acoustic conditions on OES; none of the regression steps had a significant model. Indeed, in this analysis
even partition height was not predictive. The reduced sample size might explain this.
3.6.2 Acoustic conditions.
IRC RR-154 29
Workstation and Physical Condition Effects on Environmental Satisfaction
Table 21. Summary table for OES regressed on workstation characteristics and acoustic conditions.
AGE_COMBINE
D -.026 -.027 -.028 -.025 -.024
GENDER .053 .053 .053 .053 .053
ADMIN .096 .093 .093 .093 .094
MGR -.012 -.018 -.018 -.022 -.021
PROF -.052 -.050 -.050 -.049 -.053
SQRTAREA .075 .074 .086 .068
MINPH_NOOPEN -.098 -.097 -.082 -.085
PANELS_CAT -.008 -.007 .008 -.010
WINDOW -.036 -.036 -.048 -.033
SII .005 .045 .009
LNOISEA .057 .045
LOHI_DBA .061
R2 change .016 .006 .000 .001 .003
Total R2 .016 .022 .022 .023 .026
Adjusted R2 .009 .009 .007 .007 .008
Note. N = 671.
We conducted three regression analyses for OES involving
workstation characteristics and lighting conditions. As before, we
looked first at the whole sample and then at the effects on central and peripheral workstations.
3.6.3 Lighting conditions.
Table 22 summarizes the results for the full sample, of which 706 remained after excluding cases
having missing data and outliers. The model achieved statistical significance at each step, and added
slightly to the percentage of explained variance over the workstation characteristics model. Only partition
height was a statistically significant predictor at any step. When lighting conditions were controlled,
higher partitions were still associated with lower overall environmental satisfaction.
Table 22. Summary table for OES regressed on workstation characteristics and lighting conditions.
 
AGE_COMBINE
D -.045 -.049 -.049 -.050 -.048 -.048 -.050
GENDER .051 .050 .050 .047 .046 .047 .047
ADMIN .061 .063 .063 .059 .063 .064 .064
MGR -.049 -.056 -.056 -.058 -.058 -.058 -.059
PROF -.084 -.076 -.077 -.081 -.082 -.082 -.081
SQRTAREA .091 .092 .073 .072 .072 .066
MINPH_NOOPEN -.134** -.134** -.115* -.115* -.113* -.114*
PANELS_CAT -.023 -.023 -.020 -.019 -.018 -.017
VDT_CAT -.013 -.008 -.013 -.014 -.014
CUBEDAYT .049 .047 .045 .037
UNIFDAYT -.039 -.042 -.041
EH2V -.008 -.001
NO_DL_WI .023
R2 change .019* .011* .000 .002 .001 .000 .000
Total R2 .019* .031** .031** .033** .034* .034* .035*
Adjusted R2 .012* .020** .018** .019** .019* .018* .017*
Note. N = 716. * p<=.05. **p<=.01. ***p<=.001.
We again split the sample into central and peripheral workstations. The between-groups
descriptive statistics are shown in Table 23. There are small differences from the groups in the satisfaction
with lighting analyses because of small differences in sample size. Although OES was slightly higher in
IRC RR-154 30
Workstation and Physical Condition Effects on Environmental Satisfaction
peripheral workstations than central ones, we would not expect it to be a significant difference, given that
WINDOW was not a significant predictor in the workstation characteristics regression, nor was
NO_DL_WI in the full sample model for workstation conditions with lighting conditions.
Table 23. OES descriptive statistics for full sample and lighting subgroups.
Full Sample Central WS Peripheral WS
M SD N M SD N M SD N
OES 4.05 1.31 716 3.95 1.27 303 4.15 1.33 412
AGE_COMBINED 2.62 .96 716 2.49 .99 303 2.72 .92 412
GENDER 1.53 .50 716 1.53 .50 303 1.53 .50 412
ADMIN .27 .44 716 .27 .44 303 .26 .44 412
MGR .09 .28 716 .05 .21 303 .11 .32 412
PROF .40 .49 716 .40 .49 303 .40 .49 412
SQRTAREA 8.87 2.02 716 8.53 2.00 303 9.15 1.98 412
MINPH_NOOPEN 60.67 9.55 716 61.26 9.96 303 60.35 9.17 412
PANELS_CAT 1.74 .44 716 1.73 .45 303 1.75 .43 412
VDT_CAT 1.92 .89 716 1.90 .87 303 1.94 .90 412
CUBEDAYT 242.30 150.83 716 169.29 68.17 303 298.26 173.47 412
UNIFDAYT .44 .20 716 .43 .22 303 .44 .19 412
EH2V 2.32 .82 716 2.66 .79 303 2.06 .74 412
NO_DL_WI .97 .91 716
WINDOW .70 .46 412
Note. Central workstations had NO_DL_WI = 0. There were 330 of these in the full COPE sample. Peripheral
workstations had NO_DL_WI = 1 or 2. There were 449 of these in the full COPE sample.
For the central workstations, the OES regression is summarized in Table 24. Although the model
as a whole did not reach statistical significance on any step, there is one independent variable that was a
significant predictor throughout. Professionals had lower OES than non-professionals in the central
workstations. We suspect that this is an indication that people in this job class expect to have an office
with a window, or at least access to daylight. The average illuminance on the cube (CUBEDAYT) added
significantly to the percentage of variance explained on the step when it was added, but the overall
equation still did not reach statistical significance.
Table 24. Central workstations’ summary table for OES regressed on workstation characteristics and lighting
conditions.
AGE_COMBINE
D .012 .017 .017 .017 .016 .016
GENDER .030 .028 .028 .032 .034 .034
ADMIN -.057 -.034 -.035 -.033 -.034 -.034
MGR -.054 -.049 -.049 -.040 -.040 -.040
PROF -.191** -.160* -.162* -.178* -.179* -.179*
SQRTAREA
.043 .042 .012 .017 .017
MINPH_NOOPEN
-.106 -.107 -.068 -.070 -.070
PANELS_CAT
-.046 -.045 -.047 -.049 -.049
VDT_CAT
.011 .002 .005 .005
CUBEDAYT
.129* .137* .137
UNIFDAYT
.020 .020
EH2V
.016
R2 change .028 .010 .000 .015* .000 .003
Total R2 .028 .039 .039 .054 .055 .058
Adjusted R2 .012 .013 .009 .022 .019 .019
Note. N = 303. * p<=.05. **p<=.01. ***p<=.001.
IRC RR-154 31
Workstation and Physical Condition Effects on Environmental Satisfaction
The results for peripheral workstations showed a different pattern from the central workstations.
On step one, with only the control variables in the model, administrators show higher OES than non-
administrators. This could be the converse to the effect in the central workstations: administrators might
have higher OES because they do not expect to have peripheral workstations. In addition, older
participants showed lower OES than younger ones, but only in this analysis. At the final step we see an
interesting effect of adding WINDOW to the model. For people in peripheral workstations, the presence
of a window leads to lower OES, whereas larger workstations lead to higher OES. The final model
explained 6.3% of the variance in OES, a small effect but larger and more interpretable than when all the
workstations were considered together.
Table 25. Peripheral workstations’ summary table for OES regressed on workstation characteristics and lighting
conditions.
 
AGE_COMBINE
D -.118* -.118* -.117* -.117* -.116* -.117* -.110*
GENDER .063 .063 .063 .064 .063 .067 .066
ADMIN .151* .144* .141* .142* .147* .149* .141*
MGR .002 -.006 -.007 -.007 -.008 -.012 -.017
PROF .012 .009 .005 .005 .004 .002 -.004
SQRTAREA .060 .061 .064 .073 .071 .162*
MINPH_NOOPEN -.100 -.099 -.103 -.105 -.099 -.069
PANELS_CAT -.011 -.011 -.011 -.010 -.010 -.035
VDT_CAT -.022 -.025 -.029 -.034 -.042
CUBEDAYT -.011 -.009 -.023 .011
UNIFDAYT -.049 -.058 -.053
EH2V -.034 -.062
WINDOW -.186**
R2 change .032* .007 .000 .000 .002 .001 .021*
Total R2 .032* .038* .039 .039 .041 .042 .063*
Adjusted R2 .020* .019* .017 .015 .015 .013 .032*
Note. N = 412. * p<=.05. **p<=.01. ***p<=.001.
Table 26 shows the summary of the regression of OES on
workstation characteristics and ventilation conditions, which
included data from 697 cases. The model is statistically significant at all steps, although only two
independent variables were statistically significant predictors. As expected, lower partition heights were
associated with higher OES. In addition, people without an air supply diffuser in the workstation reported
lower OES (DL_OUT).
3.6.4 Ventilation/IAQ conditions.
IRC RR-154 32
Workstation and Physical Condition Effects on Environmental Satisfaction
Table 26. Summary table for OES regressed on workstation characteristics and ventilation conditions.
AGE_COMBINE
D
-.026 -.030 -.034 -.033 -.029
GENDER .045 .045 .051 .048 .053
ADMIN .072 .078 .060 .061 .061
MGR -.028 -.032 -.042 -.028 -.037
PROF -.083 -.072 -.077 -.073 -.070
SQRTAREA .109 .101 .071 .071
MINPH_NOOPEN -.141** -.132* -.149** -.138*
PANELS_CAT -.039 -.047 -.042 -.044
WINDOW -.026 -.030 -.035 -.039
AIR_V_H -.020 -.030 -.034
RTD_H -.060 -.050 -.041
REL_HUMID .051 .052 .069
DL_OUT -.107** -.111**
FDCO2 -.064
POLLUT -.015
R2 change .018* .013 .006 .009** .004
Total R2 .018* .031** .037** .047** .051**
Adjusted R2 .011 .019 .020 .028 .030
Note. N=697. * p<=.05. **p<=.01. ***p<=.001.
The regressions for OES in general
explain less variance than the
regressions for satisfaction in specific domains (privacy, lighting, and ventilation), probably reflecting a
mediated model. However, some of the significant predictors did not appear in the other regressions, so
that the OES regressions extend our knowledge.
3.6.5 Discussion: Overall environmental satisfaction.
One particularly intriguing finding is the negative relationship between partition height and OES.
Acoustic considerations had led us to predict that higher partitions would improve satisfaction with
privacy (which it did not, see above) and, by extension, overall environmental satisfaction. Here, we
found instead that lower partitions improved satisfaction. This finding is consistent with some
investigations, but not others. Oldham (1988) found that office satisfaction (a comparable construct to our
OES) increased for people who moved from a totally open (no partitions) office to an office with
partitions varying between 48 and 72 inches; this finding is similar to the pattern observed by Mercer
(1979). Conversely, Sundstrom et al. (1982) did not have a continuous measurement of partition height,
but found that the number of enclosed sides with height greater than 72 inches was a significant predictor
of workspace satisfaction for secretaries, bookkeepers, and accountants. In the original BOSTI study,
Brill et al. (1984) reported that with workstation area held constant, enclosure (height and number of
partitions together) was positively related to environmental satisfaction. However, Marans and Yan
(1989) found that, in a model that first entered occupants’ subjective ratings of workstation attributes,
actual floor area and enclosure did not add significantly to the prediction of environmental satisfaction for
occupants of open-plan offices. Moreover, subjectively rated noise and conversational privacy were
relatively unimportant predictors of environmental satisfaction. Rated adequacy of the space and rated
lighting quality were more important for open-plan office occupants.
Our findings are more similar to those of Marans and Yan (1989), in that the model with acoustic
conditions added did not add to the prediction of OES over the workstation characteristics alone.
Moreover, the partition height variable was not predictive of OES when lighting conditions were added.
With ventilation conditions added, it remained predictive, along with the presence of an air supply in the
workstation.
The descriptive statistics for the acoustic variables showed that speech intelligibility was
generally high (see Table 5). If everyone had relatively less privacy than they had wanted, perhaps a small
IRC RR-154 33
Workstation and Physical Condition Effects on Environmental Satisfaction
increase in partition height was not enough to improve overall environmental satisfaction. Lower
partitions would, however, have provided greater access to daylight and might have led to perceptions of
better air movement.
The effect of a window on OES for those peripheral workstations is noteworthy. For these people,
OES was lower if the window was present in the workstation (as opposed to being within 15’ or 5 m), but
this model controlled for lighting conditions. This might provide an explanation for the finding that for
people with a window, having a window is less important than for those who do not have a window
(Boubekri & Haghighat, 1993). Although having a window provides a view of outdoors and lots of
daylight, it can bring with it thermal problems. Controlling for the lighting effects of the window might be
what allowed this dissatisfaction to become apparent. This explanation is consistent with the effects
observed for satisfaction with ventilation.
We also found that having an air supply in the workstation was associated with higher
environmental satisfaction, even when the physical conditions were also in the model. Thus, the effect is
not a matter of the physical conditions being influenced by the location of the diffuser. It appears most
likely that this is a psychological judgement in which environmental satisfaction increases when one can
see that there is a direct supply of fresh air into the workstation. The literature does not appear to include
other investigations that have considered this variable, so this finding awaits replication.
3.7 Predicting Job Satisfaction
As for OES, we looked for direct effects of workstation
characteristics on job satisfaction. These, analysed with a
sample of 739, are shown in Table 27. The model was statistically significant at all steps, revealing
interesting relationships. In every step, age is a significant predictor of job satisfaction: Younger workers
are more satisfied with their jobs. This is a small effect, but explains half of the explained variance for the
entire model. In step two, workstation area was a significant predictor of job satisfaction, although in an
odd direction: Smaller workstations predicted greater job satisfaction. However, this effect disappeared
after partition height entered the equation in step 3, indicating that of these two highly correlated
variables, partition height is the more important predictor of job satisfaction. Lower partitions were
associated with greater job satisfaction.
3.7.1 Workstation characteristics.
Table 27. Summary table for JobSatis regressed on workstation characteristics.
AGE_COMBINE
D
-.127*** -.104** -.108** -.113**
GENDER -.007 -.001 -.004 -.005
ADMIN -.063 -.028 -.027 -.027
MGR .002 .016 .009 .006
PROF -.032 .002 .007 .007
SQRTAREA -.111** -.029 -.055
MINPH_NOOPEN -.104* -.097*
PANELS_CAT -.032 -.026
WINDOW .049
R2 change .021** .011** .007 .002
Total R2 .021** .031*** .039*** .041***
Adjusted R2 .014** .023** .028*** .029***
Note. N = 739. * p<=.05. **p<=.01. ***p<=.001.
As always, the sample size for job satisfaction regressions including
the acoustic conditions had a smaller sample size (N=679) than the
other models. For this analysis, the model was statistically significant at every step, but only age was a
significant predictor (Table 28). The fact that neither partition height nor workstation area reached
3.7.2 Acoustic conditions.
IRC RR-154 34
Workstation and Physical Condition Effects on Environmental Satisfaction
statistical significance in this regression suggests that the smaller sample size reduced statistical power to
detect these small effects. In a larger sample we might have observed effects of acoustic conditions as
well as workstation characteristics; indeed, some of the standardized regression weights (e.g., LNOISEA)
were relatively large here, although not statistically significant.
Table 28. Summary table for JobSatis regressed on workstation characteristics and acoustic conditions.
AGE_COMBINE
D -.129*** -.112** -.109** -.104** -.103*
GENDER -.010 -.007 -.006 -.005 -.005
ADMIN -.037 -.003 -.006 -.007 -.007
MGR .038 .040 .038 .030 .030
PROF -.015 .022 .024 .027 .026
SQRTAREA -.078 -.073 -.049 -.056
MINPH_NOOPEN -.074 -.080 -.047 -.049
PANELS_CAT -.022 -.036 -.006 -.014
WINDOW .060 .060 .035 .041
SII -.041 .042 .027
LNOISEA .119 .114
LOHI_DBA .025
R2 change .020* .019* .001 .005 .000
Total R2 .020* .039** .040** .045*** .046**
Adjusted R2 .013* .026** .026** .030*** .029**
Note. N= 679. * p<=.05. **p<=.01. ***p<=.001.
Once again we examined both the models for the full sample, and
for two subgroups. The results for the regression of job satisfaction
on workstation characteristics and lighting conditions using the full sample are shown in Table 29. The
model is statistically significant and explains a total of 4.2% of the variance, but in the final step the only
statistically significant predictor is age. Partition height was statistically significant in steps two and three,
but its variance spread over other variables in subsequent steps, leaving none sufficiently powerful to
reach statistical significance at the final step.
3.7.3 Lighting conditions.
Table 29. Summary table for JobSatis regressed on workstation characteristics and lighting conditions.
   
AGE_COMBINE
D -.127*** -.107** -.107** -.108** -.109** -.111** -.115**
GENDER -.002 .001 .001 -.001 .000 .002 .002
ADMIN -.062 -.027 -.027 -.029 -.032 -.030 -.028
MGR -.008 .000 .000 -.001 -.001 -.004 -.005
PROF -.034 .004 .004 .001 .001 .001 .004
SQRTAREA
-.040 -.040 -.053 -.053 -.054 -.070
MINPH_NOOPEN
-.099* -.099* -.086 -.086 -.079 -.082
PANELS_CAT
-.021 -.021 -.019 -.020 -.018 -.015
VDT_CAT -.007 -.003 -.001 -.004 -.005
CUBEDAYT
.033 .034 .024 .004
UNIFDAYT .020 .009 .010
EH2V
-.031 -.014
NO_DL_WI
.059
R2 change .020** .018** .000 .001 .000 .001 .002
Total R2 .020** .038*** .038*** .039** .039** .040** .042**
Adjusted R2 .013** .027*** .026*** .025** .024** .024** .024**
Note. N = 721. * p<=.05. **p<=.01. ***p<=.001.
IRC RR-154 35
Workstation and Physical Condition Effects on Environmental Satisfaction
Overall and subgroup descriptive statistics for the variables in this analysis are shown in Table
30. The independent variables show small differences from the subgroup values for the satisfaction with
lighting and OES regressions because of small differences in sample size. The small difference in job
satisfaction between the groups would not be expected to be statistically significant, given the fact that
NO_DL_WI was not a significant predictor in the overall regression model (Table 29), nor was
WINDOW in the regression model for job satisfaction with workstation conditions alone (Table 27).
Table 30. JobSatis descriptive statistics for full sample and lighting subgroups.
Full Sample Central WS Peripheral WS
M SD N M SD N M SD N
JOBSATIS 5.14 .98 721 5.09 1.00 303 5.18 .97 417
AGE_COMBINED 2.63 .95 721 2.51 .98 303 2.72 .92 417
GENDER 1.52 .50 721 1.51 .50 303 1.53 .50 417
ADMIN .27 .45 721 .28 .45 303 .27 .44 417
MGR .08 .28 721 .05 .21 303 .11 .31 417
PROF .38 .49 721 .39 .49 303 .38 .49 417
SQRTAREA 8.87 2.02 721 8.53 2.00 303 9.15 1.98 417
MINPH_NOOPEN 60.70 9.47 721 61.19 9.96 303 60.45 9.03 417
PANELS_CAT 1.74 .44 721 1.72 .45 303 1.75 .43 417
VDT_CAT 1.92 .88 721 1.88 .86 303 1.94 .90 417
CUBEDAYT 241.80 149.92 721 169.76 67.05 303 296.38 172.79 417
UNIFDAYT .44 .20 721 .43 .21 303 .44 .19 417
EH2V 2.32 .82 721 2.68 .80 303 2.06 .74 417
NO_DL_WI .98 .91 721
WINDOW 5.14 .98 721 .70 .46 417
Note. Central workstations had NO_DL_WI = 0. There were 330 of these in the full COPE sample. Peripheral
workstations had NO_DL_WI = 1 or 2. There were 449 of these in the full COPE sample.
The results for the central workstations show the effect of the smaller sample size and reduced
statistical power. Although the percentage of explained variance is comparable to other models (4.2% in
total), neither the model as a whole nor any of the independent variables is statistically significant. For
people in central workstations it does not appear that lighting conditions directly influence job
satisfaction.
IRC RR-154 36
Workstation and Physical Condition Effects on Environmental Satisfaction
Table 31. Central workstations’ summary table for JobSatis regressed on workstation characteristics and lighting
conditions.
AGE_COMBINE
D -.130* -.110 -.112 -.112 -.112 -.110
GENDER -.024 -.031 -.033 -.033 -.034 -.035
ADMIN -.038 -.004 -.001 -.001 -.001 -.002
MGR .021 .030 .030 .028 .028 .030
PROF .002 .056 .062 .064 .065 .064
SQRTAREA
-.085 -.081 -.077 -.078 -.080
MINPH_NOOPEN
-.034 -.033 -.039 -.039 -.052
PANELS_CAT
-.056 -.059 -.058 -.057 -.065
VDT_CAT
-.041 -.039 -.040 -.036
CUBEDAYT
-.022 -.025 -.037
UNIFDAYT
-.006 .005
EH2V
.045
R2 change .020 .020 .002 .000 .000 .001
Total R2 .020 .040 .041 .042 .042 .043
Adjusted R2 .003 .014 .012 .009 .006 .003
Note. N = 303. * p<=.05. **p<=.01. ***p<=.001.
For the peripheral workstations, the overall model was statistically significant at each step.
However, the lighting conditions did not add any information to what was known from the regression of
workstation characteristics alone. In the early steps, age and partition height were statistically significant
predictors. Partition height ceased to be predictive as more variables were added into the model.
Table 32. Peripheral workstations’ summary table for JobSatis regressed on workstation characteristics and
lighting conditions.
 
AGE_COMBINE
D -.144** -.114** -.114* -.114* -.114* -.116* -.116*
GENDER .004 .024 .024 .023 .023 .030 .030
ADMIN -.089 -.055 -.056 -.056 -.058 -.056 -.056
MGR -.024 -.016 -.017 -.017 -.016 -.021 -.021
PROF -.066 -.044 -.045 -.045 -.045 -.049 -.049
SQRTAREA
-.066 -.065 -.068 -.072 -.073 -.081
MINPH_NOOPEN
-.129* -.129* -.127* -.126* -.120 -.122
PANELS_CAT
.016 .017 .017 .016 .015 .018
VDT_CAT
-.005 -.004 -.001 -.006 -.006
CUBEDAYT .008 .007 -.013 -.016
UNIFDAYT
.023 .011 .011
EH2V
-.045 -.043
WINDOW
.016
R2 change .029* .027** .000 .000 .000 .001 .000
Total R2 .029* .056** .056** .056** .057** .058* .058*
Adjusted R2 .017* .038** .035** .033** .031** .030* .028*
Note. N = 417. * p<=.05. **p<=.01. ***p<=.001.
The regression of job satisfaction on workstation
characteristics and ventilation conditions had 703 cases after
the exclusion of cases with missing data and outliers. The results are shown in Table 33. In addition to the
effects of age and partition height, this model shows that air quality is a significant predictor of job
3.7.4 Ventilation/IAQ conditions.
IRC RR-154 37
Workstation and Physical Condition Effects on Environmental Satisfaction
satisfaction. Together the two variables that indicated the presence of air quality problems explained 2.1%
of the variance, and both were statistically significant predictors in the final model, with higher pollutant
levels leading to poorer job satisfaction. Moreover, the standardized regression weights show that they
have comparable effects on predicted job satisfaction to the other significant predictors in the model (the
Beta weights are all of the same order of magnitude, around .100).
Table 33. Summary table for JobSatis regressed on workstation characteristics and ventilation conditions.
AGE_COMBINE
D
-.132*** -.121** -.120** -.120** -.113**
GENDER -.008 -.005 .002 .000 .007
ADMIN -.055 -.021 -.019 -.019 -.022
MGR .031 .033 .029 .037 .013
PROF -.021 .018 .016 .018 .024
SQRTAREA -.047 -.051 -.067 -.048
MINPH_NOOPEN -.102* -.123* -.134* -.107*
PANELS_CAT -.021 -.019 -.017 -.020
WINDOW .050 .048 .045 .030
AIR_V_H -.018 -.023 -.034
RTD_H -.024 -.019 -.002
REL_HUMID -.029 -.028 .054
DL_OUT -.062 -.061
FDCO2 -.121**
POLLUT -.106*
R2 change .022** .019** .001 .003 .021***
Total R2 .022** .041*** .043** .046** .067***
Adjusted R2 .015** .029*** .026** .028** .046***
Note. N= 703. * p<=.05. **p<=.01. ***p<=.001.
Job satisfaction had been expected to show few direct
relationships with workstation characteristics or physical
conditions; we had expected the relationships to be indirect ones. Indeed, few such relationships were
observed. Nonetheless, some of the relationships were of the same magnitude (5% variance explained) as
for the domain-specific analyses above.
3.7.5 Discussion: Job satisfaction.
In the same manner as for overall environmental satisfaction, we were surprised to find that
partition height showed a small, but statistically significant, negative relationship to job satisfaction.
Lower partition heights were associated with higher job satisfaction, even after controlling for individual
demographic characteristics. This finding is contrary to previous research, in which job satisfaction
increased with the degree of enclosure (Oldham, 1988; Oldham & Brass, 1979; Sundstrom et al., 1980;
Sundstrom et al., 1982). Because none of the earlier studies used measured partition height as a
continuous independent variable, direct connections to our results are limited. Oldham and Fried (1987)
found that enclosure did not influence job satisfaction in a main effect; rather, it interacted with the
judged darkness of the workspace. Small, dark work spaces led to low job satisfaction. Interestingly, in
our analysis, the effect of partition height disappeared when lighting characteristics were added to the
model.
We did not find that acoustic conditions added to the prediction of job satisfaction. This is
consistent with other investigations both in offices (Leather, Beale, & Sullivan, 2003) and in factories
(Melamed, Fried, & Froom, 2001). However, other investigations have found that workplace noise levels
interact with other job characteristics and stressors to influence job satisfaction. Our investigation did not
include measurement of these variables; therefore we are unable to replicate the other results.
Surprisingly, the concentrations of both carbon dioxide and other pollutants were significant
predictors of job satisfaction, even when workstation characteristics and other ventilation conditions were
IRC RR-154 38
Workstation and Physical Condition Effects on Environmental Satisfaction
held constant. This finding is new, and worthy of further research attention.
Across all models we found that age was negatively related to job satisfaction. Younger people
were more satisfied with their jobs than older people. Given the high mean value for job satisfaction, this
finding does not mean that older people were dissatisfied, only that they were less satisfied. The literature
on job satisfaction shows all manner of relationships to age, with relatively little support for simple linear
effects (Bernal, Snyder, & McDaniel, 1998; Hochwarter, Ferris, Perrewe, Witt, & Kiewitz, 2001). It is
possible that our sample represented the negatively sloped portion of a U-shaped curve (Hochwarter et al.,
2001).
3.8 Cumulative Risk Factors
We developed a set of heuristics to identify workstation
characteristics and physical conditions that would lead to a
higher incidence of low satisfaction, based partly on the regression results reported here and partly on
other literature. These are reported elsewhere (Newsham et al., 2003b). Based on these heuristics we
developed two new variables to track the cumulative risks in each workstation - that is, the total number
of potentially adverse conditions experienced by each respondent. One variable counted only those
conditions that will be predicted in the COPE software (RISK_COPE), and the other included all the
identified potential risks (RISK_ALL). The criteria used to define these variables are shown in Table 34
together with the descriptive statistics for each variable. Only those cases that had data on all of the
variables needed for the cumulative risk calculation were included. Each case received one point for each
physical characteristic that met the criterion values. Thus, higher scores indicate that the case had more
risk for poor satisfaction.
3.8.1 Cumulative risk variables.
Both variables were well distributed. There was a slight negative skew to both variables, but both
met the criteria for normal distributions that had been previously set. Histograms of both variables are
shown in Figure 2. Values for RISK_ALL are of course higher because there were more criteria used to
calculate it.
Table 34 also shows the bivariate correlations for the two variables with OES and JobSatis. These
dependent variables were used in subsequent analyses of the effects of cumulative risk because the
individual domains of satisfaction had been the source of the criteria for their definition. The correlations
were expected to be negative (lower risk -- higher satisfaction). Although they are low, the correlations
for OES are in the same range as some that had shown interpretable results in the earlier regression
analyses.
Figure 2. Histograms for cumulative risk variables.
012345678910
Cumulative Risk (COPE)
0
50
100
150
200
250
Count
0.0
0.1
0.2
0.3
Proportion per Bar
012345678910
Cumulative Risk (ALL)
0
50
100
150
200
Count
0.0
0.1
0.2
Proportion per Bar
IRC RR-154 39
Workstation and Physical Condition Effects on Environmental Satisfaction
Table 34. Criteria and descriptive statistics for cumulative risk variables.
RISK_COPE RISK_ALL
SII >= 0.5 >= 0.5
MINPH_NOOPEN <= 54 OR>= 66 <= 54 OR>= 66
SQRTAREA <= 8 <= 8
LNOISEA <= 44 OR >= 50 <= 44 OR >= 50
DESKILLUM <= 300 <= 300
VDT_CAT >= 2 >= 2
WINDOW = 0 = 0
RTD_H <= 21.5 OR >= 23.5
AIR_V_H >= 0.10
FDCO2 >= 650
Minimum 0 0
Maximum 7 10
M 3.10 4.34
SD 1.49 1.88
Median 3 4
N 729 729
OES bivariate correlation (N = 696) -.04 -.05
JobSatis bivariate correlation (N = 717) -.01 .00
We conducted hierarchical regression
analyses to determine whether or not the
cumulative risk scores predicted overall environmental satisfaction and job satisfaction (next section),
again controlling for demographic characteristics on Step one. Separate analyses were conducted, first for
RISK_COPE and then for RISK_ALL.
3.8.2 Predicting overall environmental satisfaction.
Table 35 shows the summary result for RISK_COPE, and Table 36 shows the result for
RISK_ALL. Neither regression model was statistically significant at any step, nor was either risk variable
a significant predictor of OES.
Table 35. Summary table for OES regressed on RISK_COPE.
AGE_COMBINE
D
-.026 -.030
GENDER .046 .047
ADMIN .085 .076
MGR -.010 -.014
PROF -.053 -.056
RISK_COPE -.037
R2 change .014 .001
Total R2 .015 .015
Adjusted R2 .007 .007
Note. N= 689. * p<=.05. **p<=.01. ***p<=.001.
IRC RR-154 40
Workstation and Physical Condition Effects on Environmental Satisfaction
Table 36. Summary table for OES regressed on RISK_ALL.
AGE_COMBINE
D -.027 -.033
GENDER .045 .050
ADMIN .084 .070
MGR -.010 -.018
PROF -.054 -.063
RISK_ALL -.060
R2 change .014 .003
Total R2 .014 .018
Adjusted R2 .007 .009
Note. N= 688. * p<=.05. **p<=.01. ***p<=.001.
Tables 37 and 38 show the corresponding analyses for
regressions in which JobSatis was the dependent variable.
The overall models were statistically significant, but the only statistically significant predictor was age,
which again showed a small, negative relationship to job satisfaction, explaining approximately 2% of the
variance.
3.8.3 Predicting job satisfaction.
Table 36. Summary table for JobSatis regressed on RISK_COPE.
AGE_COMBINE
D -.124*** -.126***
GENDER -.006 -.006
ADMIN -.048 -.052
MGR .018 .016
PROF -.022 -.024
RISK_COPE -.124 -.018
R2 change .019* .000
Total R2 .019* .019*
Adjusted R2 .011* .010*
Note. N= 696. * p<=.05. ** *p<=.001.
Table 38. Summary table for JobSatis regressed on RISK_ALL.
AGE_COMBINE
D -.124*** -.126***
GENDER -.007 -.006
ADMIN -.048 -.053
MGR .018 .015
PROF -.023 -.026
RISK_ALL
-.022
R2 change .019* .000
Total R2 .019* .019*
Adjusted R2 .011* .010*
Note. N= 695. * p<=.05. **p<=.01. ***p<=.001.
Although there was adequate variability in the scores for
cumulative risk, neither variable was a statistically
significant predictor of either overall environmental satisfaction or job satisfaction. There are several
possible explanations. First, it is possible that the relationship is not a linear one. Perhaps low levels of
cumulative risks may be tolerated, but a higher level might not. It is also possible that the various risks are
3.8.4 Discussion: Cumulative risk.
IRC RR-154 41
Workstation and Physical Condition Effects on Environmental Satisfaction
not additive in this simple way; rather, each might influence a different outcome, and some might do so in
ways than cancel out (e.g., having a window improves satisfaction with lighting but decreases satisfaction
with ventilation). Whatever the reason, one implication of this null result is that the risk markers in the
COPE software do not add up to an overall score for the satisfaction consequence of any workstation
design.
3.9 Ranked Order of Importance of Workstation Features
Seven workstation characteristics were ranked for their relative importance
to the individual. In this instance, lower numbers reflect greater
importance; thus, a ranking of 1 would indicate the most important element. The responses are
summarized in Table 39. Using the mean and median rankings as a guide, the elements sort in the order:
Air Quality & Ventilation; Privacy; Noise Levels; Temperature; Lighting; Size; and, Window Access.
3.9.1. Frequencies.
Table 39. Ranked importance of seven workstation features.
Rank 1 2 3 4 5 6 7 N M SD Median
Lighting 65 93 138 113 119 115 64 707 4.03 1.78 4.00
Air Quality &
Ventilation
163 106 105 109 98 83 43 707 3.42 1.91 3.00
Temperature 80 129 110 98 102 103 85 707 3.94 1.93 4.00
Noise Levels 94 124 97 125 118 84 65 707 3.79 1.86 4.00
Privacy 155 118 103 86 91 97 57 707 3.51 1.99 3.00
Size of Workstation 64 66 101 103 106 132 135 707 4.50 1.93 5.00
Window Access 86 71 53 73 73 93 258 707 4.82 2.20 5.00
We examined these using the Goodman-Kruskal Gamma non-
parametric test to investigate whether the rank assigned to each
environmental aspect was influenced by three workstation characteristics: workstation area, partition
height, and presence of a window. We also examined the influence of age and job category, because of
the effects that these variables showed in some of the regression analyses.
3.9.2. Nonparametric analyses.
Gamma assesses the relationship between two ordered categorical variables. It may take on values
between -1 and +1, and is a proportional reduction of error statistic, meaning that its size tells us the
percentage of the cases in which our classification on one variable is improved by knowing the other
variable (Wilkinson, Blank, & Gruber, 1996).
As this test can only be conducted using categorical data, we formed categories for the continuous
variables, SQRTAREA and MINPH_NOOPEN. Table 40 shows the definitions of the category cutpoints
and the number of workstations in each category that resulted. We chose cutpoints that made practical
sense: integer numbers of feet for SQRTAREA, and values than seemed to correspond to commonly
available systems furniture heights for partition height.
Table 40. Category definitions for continuous variables.
SQRTAREA <=
4
4 - 5 5 - 6 6 - 7 7 - 8 8 -9 9 - 10 10 -
11
11 -
12
12 -
13
>= 13
SQRA_CAT 1 2 3 4 5 6 7 8 9 10 11
# 6 81 110 116 103 111 135 51 54 12
MINPH_NOOPEN <=
30
30 -
48
48 -
57
57 -
63
63 -
68
68 -
74
>= 74
MINPHCAT 1 2 3 4 5 6 7
# 1 99 150 124 274 127 4
3.9.3 Crosstabulations by workstation area. Table 41 summarizes the results of the
crosstabulations of the ranks assigned to each
IRC RR-154 42
Workstation and Physical Condition Effects on Environmental Satisfaction
feature, by workstation area. The gamma values in our data show that the relative importance of four
features varied by workstation area, all by a small amount (11 to 13%). The relative importance of
lighting went up as workstation area increased, but window access declined in importance. The relative
importance of air quality and ventilation declined. Noise levels became more important as size increased.
Table 41. Summary of crosstabulations for feature importance ranks by workstation area.
Feature Goodman-Kruskal Gamma
Lighting .11**
Air Quality & Ventilation -.12**
Temperature .02
Noise Levels .11**
Privacy .00
Size of Workstation .02
Window Access -.13**
Note. ** p <= .01
Table 42 summarizes the results of the
crosstabulations of the ranks assigned to each
feature, by partition height. Only two features showed significant relationships, and these were small. The
importance of air quality and ventilation declined as partition height increased. The importance of
workstation size increased as partition height increased.
3.9.4 Crosstabulations by partition height
Table 42. Summary of crosstabulations for feature importance ranks by partition height.
Feature Goodman-Kruskal Gamma
Lighting .07
Air Quality & Ventilation -.10*
Temperature .02
Noise Levels .04
Privacy .00
Size of Workstation .10*
Window Access -.12**
Note. * p <= .05. ** p <= .01
Table 43 summarizes the results of the
crosstabulations of the ranks assigned to each
feature, by the variable NO_DL_WI, which codes for both the presence of a window and access to
daylight. These results are intriguing: lighting increases in importance across the change from no
daylight, to daylight access, to having a window. Having access to a window, however, declines in
importance.
3.9.5 Crosstabulations by windows and daylight
Table 43. Summary of crosstabulations for feature importance ranks by access to daylight.
Feature Goodman-Kruskal Gamma
Lighting .16**
Air Quality & Ventilation .01
Temperature -.01
Noise Levels .06
Privacy -.02
Size of Workstation .05
Window Access -.23**
Note. ** p <= .01
3.9.6 Crosstabulations by age Table 44 summarizes the results of the crosstabulations of the
ranks assigned to each feature, by age. Only the importance of
IRC RR-154 43
Workstation and Physical Condition Effects on Environmental Satisfaction
air quality and ventilation varied by age, becoming less important for older employees.
Table 44. Summary of crosstabulations for feature importance ranks by age.
Feature Goodman-Kruskal Gamma
Lighting .07
Air Quality & Ventilation -.16**
Temperature .04
Noise Levels .02
Privacy .03
Size of Workstation -.02
Window Access .02
Note. ** p <= .01
Table 45 summarizes the results of the
crosstabulations of the ranks assigned to each feature,
by job category. Job category was coded on an assumed gradient of responsibility, where 1 =
administrative; 2 = technical; 3 = professional; 4 = managerial. Thus, the results indicate that people in
the professional and managerial categories place more importance on temperature, whereas administrative
and technical ranks place more importance on noise levels and access to a window. (There was no
relationship between job category and having a window or daylight, gamma = .08, n.s.)
3.9.7 Crosstabulations by job category
Table 45. Summary of crosstabulations for feature importance ranks by job category.
Feature Goodman-Kruskal Gamma
Lighting .01
Air Quality & Ventilation .03
Temperature .20***
Noise Levels -.11**
Privacy -.04
Size of Workstation .03
Window Access -.11**
Note. ** p <= .01. *** p <= .001
The overall pattern of importance rankings
is somewhat surprising: Although size and
window access (or daylight) were important predictors of some of the satisfaction outcomes, they did not
receive the highest importance rankings. Kupritz (2003a) also found windows to be relatively low-
ranking, but in her list “having a large personal office” was the top-ranked item for both younger and
older employees.
3.9.8 Discussion: Ranked importance of features.
The crosstabulation of our data show that the relative importance of these features depended on
the state of other features. Some of these relationships have parallels in the literature. Boubekri and
Haghighat (1993) also found that people without windows rated window access as more important than
did people whose cubicles had windows.
The results of the importance rankings seem to suggest that the relative importance of a
workstation feature is greater when that feature might be less than optimal. Thus, people with windows
rate window access as less important, but lighting, which should include glare control, as more important.
This pattern is consistent with the regression analyses for satisfaction with lighting. Professionals
experienced lower noise levels than others (Table 8), and rated noise as less important. Those who,
traditionally have a lower likelihood of having window access (administrators) place more importance on
having it. However, this finding does not hold for all features: lower partition height was associated in the
regressions with increased satisfaction with ventilation, but in the rank analysis with more importance for
air quality and ventilation.
IRC RR-154 44
Workstation and Physical Condition Effects on Environmental Satisfaction
One relationship that was not present was a connection between age and the importance of
privacy. This is consistent with Kupritz (2001) but not with her more recent work (Kupritz, 2003a)*, in
which older workers placed more importance on workplace features than would provide more privacy,
than did younger workers.
3.10 Open-Ended Comments
At the conclusion of each workstation visit, the research team left a paper questionnaire inviting
written comments to three questions, together with a stamped enveloped in which the participant could
mail it back to NRC. The response was poor, with 108 completed questionnaires returned from 779
workstations visited (13.9%). The responses to each question were typed verbatim, then categorised.
Table 45 shows the top three categories of response to each of the three questions. The number of
comments exceeds the number of questionnaires because many people gave more than one response to
each question.
The dominant concern of these open-plan office workers is functionality. They want their
workstations to have the space and features necessary to the performance of their work. In one area,
employees who worked with large building plans complained that there was not enough desk area to
unroll and support the entire sheet. Others commented more favourably, “I have the tools (computer,
desk) that I need to do my job well.” The quantitative questionnaire did not address this issue directly so
there is no easy comparison to the other results.
Privacy appears to follow functionality. Although a large number of favourable responses
concerned privacy, a larger number of respondents complained about a lack of privacy, or wanted to make
changes in order to improve privacy. This might be considered an extension of functionality: freedom
from distractions to enable the work to be done.
On the positive side, and a feature that some would like to change, was availability of a window.
Those who mentioned it, wanted a window. No one said they wanted to give up a window.
Finally, air quality and thermal comfort issues merited both complaints and desire for change.
Temperature fluctuations over the day from too hot to too cold were among the complaints; others
focused on the cold problem. Some complained that the air was stuffy, others that there were smells. We
found no favourable comments in this category.
Indeed, there were fewer comments provided in response to “Things I like most” than to the other
two questions, suggesting that people were more likely to respond negatively than positively. Perhaps the
favourable characteristics are less salient than the features that cause discomfort or annoyance.
Table 45. Summary of top three categories of open-ended comments.
THINGS I LIKE MOST THINGS I LIKE LEAST THINGS I WOULD CHANGE
FEATURE # COMMENTS FEATURE # COMMENTS FEATURE # COMMENTS
Total 215 Total 247 Total 235
Functionality
(enough space,
good equipment)
76 Lack of privacy,
many
distractions,
noisy
77 Functionality 71
Window & view 59 Lack of
functionality
73 Privacy,
distraction, noise
67
Privacy, little
distraction
24 Poor IAQ &
thermal comfort
41 IAQ, thermal
comfort
22
Window & view 21
* See also Kupritz (2003b) for a discussion of these data as they apply to a variety of routine work activities.
IRC RR-154 45
Workstation and Physical Condition Effects on Environmental Satisfaction
4.0 Conclusions
Taken together, the results of this cross-sectional field study provide modest evidence that the
physical environment within the workstation influences its occupant’s satisfaction in several ways. The
most substantial effects were found for two specific environmental domains: satisfaction with lighting
and satisfaction with ventilation, with regressions explaining on the order of 10-14% of the variance in
these outcomes. Satisfaction with privacy, overall environmental satisfaction, and job satisfaction effects
were smaller in size.
Variables that were not measured in this study probably account for more of the variance in these
outcomes. For instance, personality variables such as stimulus screening interact with workstation design
characteristics; people who are less able to screen irrelevant stimuli are more affected by enclosure and
workstation area than those whose screening skills are high (Oldham et al., 1991; Oldham & Fried, 1987).
These relationships may be complex: Block and Stokes (1989) observed that sex,
introversion/extroversion, and task type interacted with office type (open or enclosed) to influence task
performance. Job complexity, which is not fully accounted for by the gross categorization of job type that
we controlled for, also interacts with workstation characteristics and physical conditions such as noise to
influence job satisfaction (Leather et al., 2003; Melamed et al., 2001; Oldham et al., 1991). Adverse
environmental conditions are tolerable by those with less complex jobs, but not for those with complex or
demanding jobs. Our investigation, with its strict focus on varieties of satisfaction and limits to the
questionnaire length, was not able to incorporate measurements of these variables.
Another important finding is that although conditions occur across a wide range, satisfaction
levels - particularly for satisfaction with lighting and job satisfaction - are relatively high. The physical
conditions, too, compare well against recommended practice. There were few cases of ventilation or
thermal conditions outside the recommended ranges; similarly, noise levels were not particularly high.
Speech intelligibility levels, when calculated with the “normal speech” level, were poor (mean SII=.51),
but there is evidence that people in open-plan offices speak more quietly (Warnock & Chu, 2002). A
recalculation with this lower speech level resulted in an average SII = .20 (J. Bradley, personal
communication, July 22, 2003), which is the target value adopted by acousticians (American National
Standards Institute, 1997). Although there were workstations with potentially problematic reflected
images in the VDT screen, the overall lighting levels were well within recommended practice. Given the
low frequency of very poor conditions by commonly accepted standards, the small percentage of variance
explained is not surprising.
We found some evidence that the effects on overall environmental satisfaction and job
satisfaction are indirect. That is, workstation physical characteristics and physical conditions have larger
direct effects on the three subscales (satisfaction with privacy, satisfaction with ventilation, and
satisfaction with lighting), than on overall environmental satisfaction or job satisfaction. This is consistent
with the satisfaction model previously developed (Charles et al., 2003), and with previous theory in
environmental satisfaction (Marans & Spreckelmeyer, 1982; Marans & Yan, 1989). The indirect nature of
the relationship probably accounts for the finding that the cumulative risk factors did not predict either
overall environmental satisfaction or job satisfaction. Another contributing element is the fact that some
of the risk factors have two faces: For instance, having a window is a good thing for satisfaction with
lighting, but a bad thing for satisfaction with ventilation.
Nonetheless, the results provide guidance for improving satisfaction with open-plan offices that
designers may use even when such details about occupants are not known. It appears from these
regressions that the single most influential design change that can improve satisfaction is to provide a
window, or at least access to daylight within 5 m (15 ft) of the workstation. However, one must also
simultaneously provide a means of glare control and take steps to moderate the thermal effects of the
window - so that it is neither too hot when there is direct sun, nor too cold in winter. Another positive step
for employees is to increase workstation size, which was positively related to satisfaction with privacy,
satisfaction with lighting, and satisfaction with ventilation in some (although not all) analyses.
IRC RR-154 46
Workstation and Physical Condition Effects on Environmental Satisfaction
It is counterintuitive that lower partition heights were associated with higher overall
environmental satisfaction and higher job satisfaction. One possible explanation is the greater daylight
penetration that lower partitions afford. The relationship disappeared when lighting conditions were
controlled, and did not occur at all for central workstations. The separate model for peripheral
workstations showed greater environmental satisfaction for larger workstations, no relationship to
partition height, and greater environmental satisfaction with no window (but access to daylight). The
importance of daylight, as opposed to a window with a view, merits further study.
As regards ventilation, the results suggest that the presence of a supply diffuser for each
workstation can improve satisfaction. Moreover, greater research attention should be paid to the
concentrations of various pollutants, as it appears that occupants may be more sensitive to low
concentrations than was previously thought. This might explain the high ranking for importance given to
indoor air quality ventilation issues.
The open-ended responses provided a reminder of the importance of the fit between workstation
design and task demands. Functionality was the element most often mentioned in these comments, both as
an element that worked and as an element that individuals would want to change. The COPE field study
was not designed to identify the specific space and furnishings needs of each occupation, and therefore no
recommendations of this kind are possible. Privacy, which was ranked as important both in the open-
ended responses and in the importance rankings, might be considered an extension of functionality.
Further discussion of both of these issues is available in the COPE report by Marquardt, Veitch, and
Charles (2002).
This field study is the only one, to our knowledge, to combine such a detailed set of physical
measurements of workstation conditions with psychometrically valid measurements of satisfaction in all
of its aspects. The next steps for research in this area should include a wider range of variables relating to
occupants, their work, and their organizations, to enable a finer-grained analysis and prescriptions for
workplace design that are tailored to individuals and their specific requirements.
IRC RR-154 47
Workstation and Physical Condition Effects on Environmental Satisfaction
5.0 References
American National Standards Institute. (1997). Methods for calculation of the speech intelligibility index
(ANSI S3.5-1997). New York: Acoustical Society of America.
American Society of Heating Refrigerating and Air Conditioning Engineers (ASHRAE). (1992). Thermal
environmental conditions for human occupancy (ANSI/ASHRAE Standard 55-1992). Atlanta,
GA: ASHRAE.
American Society of Heating Refrigerating and Air Conditioning Engineers (ASHRAE). (1997). Thermal
comfort. In R. A. Parsons, A. J. Brandstrom, & C. D. Tate (Eds.), ASHRAE Handbook
Fundamentals (p. 8.1-8.28). Atlanta, GA: ASHRAE.
American Society of Heating Refrigerating and Air Conditioning Engineers (ASHRAE). (2001).
Ventilation for acceptable indoor air quality (ANSI/ASHRAE Standard 62-2001). Atlanta, GA:
ASHRAE.
Bartko, J. J. (1976). On various intraclass correlation reliability coefficients. Psychological Bulletin, 83,
762-765.
Bernal, D., Snyder, D., & McDaniel, M. (1998). The age and job satisfaction relationship: Does its shape
and strength still evade us? Journals of Gerontology: Series B: Psychological Sciences & Social
Sciences, 53B(5), P287-P293.
Bernecker, C. A., Davis, R. G., Webster, M. P., & Webster, J. P. (1993). Task lighting in the open office:
A visual comfort perspective. Journal of the Illuminating Engineering Society, 22(1), 18-25.
Block, L. K., & Stokes, G. S. (1989). Performance and satisfaction in private versus nonprivate work
settings. Environment and Behavior, 21(3), 277-297.
Boubekri, M., & Haghighat, F. (1993). Windows and environmental satisfaction: A survey study of an
office building. Indoor Environment, 2, 164-172.
Brennan, A., Chugh, J. S., & Kline, T. (2002). Traditional versus open office design: A longitudinal field
study. Environment and Behavior, 34(3), 279-299.
Brill, M., Margulis, S. T., Konar, E., & BOSTI. (1984). Using office design to increase productivity (Vol.
1). Buffalo, NY: Workplace Design and Productivity, Inc.
Brill, M., Weidemann, S., & BOSTI Associates. (2001). Disproving widespread myths about workplace
design. Jasper, IN: Kimball International.
Brookes, M. J., & Kaplan, A. (1972). The office environment: Space planning and affective behavior.
Human Factors, 14(5), 373-391.
Charles, K. E. (2003). Fanger’s thermal comfort and draught models (DRAFT) (NRC/IRC Client Report
B3205.XX). Ottawa, ON: National Research Council of Canada, Institute for Research in
Construction.
Charles, K. E., & Veitch, J. A. (2002). Environmental satisfaction in open-plan environments: 2. Effects
of workstation size, partition height and windows (IRC-IR-845). Ottawa, ON: National Research
Council of Canada, Institute for Research in Construction.
Charles, K. E., Veitch, J. A., Farley, K. M. J., & Newsham, G. R. (2003). Environmental satisfaction in
open-plan environments: 3. Further scale validation (NRC/IRC Client Report B3205.21).
Ottawa, ON: National Research Council of Canada, Institute for Research in Construction.
Chartered Institution of Building Services Engineers (CIBSE). (1994). Code for interior lighting. London,
UK: CIBSE.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.
Cuttle, C. (1983). People and windows in workplaces. In Proceedings of the People and Physical
Environment Research Conference (pp. 203-212). Wellington, New Zealand: Ministry of Works
and Development.
Dansereau, F., Alutto, J. A., & Yammarino, F. J. (1984). Theory testing in organizational behavior: The
variant approach. Englewood Cliffs, NJ: Prentice-Hall.
Fanger, P. O. (1982). Thermal comfort. Malabar, FL: Robert E. Krieger Publishing Company.
IRC RR-154 48
Workstation and Physical Condition Effects on Environmental Satisfaction
Fanger, P. O., Melikov, H., Hanzawa, H., & Ring, J. (1988). Air turbulence and sensation of draught.
Energy and Buildings, 12, 21-39.
Finnegan, M. C., & Solomon, L. Z. (1981). Work attitudes in windowed vs windowless environments.
Journal of Social Psychology, 115, 291-292.
Griffith, J. (1997). Test of a model incorporating stress, strain, and disintegration in the cohesion-
performance relation. Journal of Applied Social Psychology, 27, 1489-1526.
Haghighat, F., Huo, Y., Zhang, J. S., & Shaw, C. Y. (1996). The influence of office furniture, workstation
layouts, diffuser types and location on indoor air quality and thermal comfort conditions at
workstations. Indoor Air, 6(3), 188-203.
Hedge, A. (1982). The open-plan office: A systematic investigation of employee reactions to their work
environment. Environment and Behavior, 14(5), 519-542.
Hedge, A., Erickson, W. A., & Rubin, G. (1992). Effects of personal and occupational factors on sick
building syndrome reports in air-conditioned offices. In J. C. Quick, L. R. Murphy, & J. J. Hurrell
(Eds.), Work and Well Being: Assessment and Interventions for Occupational Mental Health (pp.
286-298). Washington, D.C.: American Psychological Association.
Heerwagen, J. H., & Heerwagen, D. R. (1986). Lighting and psychological comfort. Lighting Design +
Application, 16(4), 47-51.
Hochwarter, W. A., Ferris, G. R., Perrewe, P. L. , Witt, L. A., & Kiewitz, C. (2001). A note on the
nonlinearity of the age-job-satisfaction relationship. Journal of Applied Social Psychology, 31(6),
1223-1237.
Illuminating Engineering Society of North America (IESNA). (1993). American national standard
practice for office lighting (ANSI/IESNA-RP-1-1993). New York: IESNA.
International Organization for Standardization (ISO). (1984). Moderate thermal environments:
Determination of the PMV and PPD indices and specification of the conditions for thermal
comfort (ISO 7730). Geneva, Switzerland: ISO.
James, L. R. (1982). Aggregation bias in estimates of perceptual agreement. Journal of Applied
Psychology, 67, 219-229.
Keighley, E. C. (1973a). Visual requirements and reduced fenestration in offices: a study of multiple
apertures and window area. Building Science, 8, 321-331.
Keighley, E. C. (1973b). Visual requirements and reduced fenestration in offices: a study of window
shape. Building Science, 8, 311-320.
Kerlinger, F. N., & Lee, H. B. (2000). Foundations of behavioral research (4th ed.). Fort Worth, TX:
Harcourt College.
Kline, R. B. (1997). Principles and practice of structural equation modeling. New York: Guilford Press.
Kupritz, V. (2003a). Accommodating privacy to facilitate new ways of working. Journal of Architectural
and Planning Research, 20(2), 122-135.
Kupritz, V. (2003b). The effects of physical design on routine work activities. Journal of Architectural
and Planning Research, 20(2), 110-121.
Kupritz, V. (2001). Aging worker perceptions about design and privacy needs for work. Journal of
Architectural and Planning Research, 18(1), 13-22.
Leather, P., Beale, D., & Sullivan, L. (2003). Noise, psychosocial stress and their interaction in the
workplace. Journal of Environmental Psychology, 23(2), 213-222.
Marans, R. W., & Spreckelmeyer, K. F. (1982). Evaluating open and conventional office design.
Environment & Behavior , 14(3), 333-351.
Marans, R. W., & Yan, X. (1989). Lighting quality and environmental satisfaction in open and enclosed
offices. Journal of Architectural & Planning Research, 6(2), 118-131.
Marquardt, C. J. G., Veitch, J. A., & Charles, K. E. (2002). Environmental satisfaction with open-plan
office furniture design and layout (NRCC/IRC Client Report B3205.14). Ottawa, ON: National
Research Council of Canada, Institute for Research in Construction.
Melamed, S., Fried, Y., & Froom, P. (2001). The interactive effect of chronic exposure to noise and job
IRC RR-154 49
Workstation and Physical Condition Effects on Environmental Satisfaction
complexity on changes in blood pressure and job satisfaction: A longitudinal study of industrial
employees. Journal of Occupational Health Psychology, 6(3), 182-195.
Mercer, A. (1979). Office environments and clerical behaviour. Environment and Planning B, 6, 29-39.
Molhave, L., Jensen, J. G., & Larsen, S. (1991). Subjective reactions to volatile organic compounds as air
pollutants. Atmospheric Environment, 25A(7), 1283-1293.
Navai, M., & Veitch, J. A. (2003). Acoustic satisfaction in open-plan offices: Review and
recommendations (NRC/IRC Client Report B3205.17). Ottawa, ON: National Research Council
of Canada, Institute for Research in Construction.
Newsham, G. R., Marchand, R. G., Svec, J. M., & Veitch, J. A. (2002). The effect of power constraints on
occupant lighting choices and satisfaction: A pilot study. In Proceedings of the IESNA Annual
Conference, Salt Lake City, UT, August 4-6, 2002 (pp. 115-131). New York: Illuminating
Engineering Society of North America.
Newsham, G. R., & Sander, D. M. (2003). The effect of office design on workstation lighting: A
simulation study. Journal of the Illuminating Engineering Society, 32 (2), 52-73.
Newsham, G. R., Veitch, J. A., Charles, K. E., Bradley, J. S., Shaw, C.-Y., Reardon, J. T., et al. (2003a).
Environmental satisfaction in open-plan environments: 4. Relationships between physical
variables (NRC/IRC Client Report B3205.22). Ottawa, ON: National Research Council of
Canada, Institute for Research in Construction.
Newsham, G. R., Veitch, J. A., Charles, K. E., Marquardt, C. J. G., Geerts, J., & Sander, D. M. (2003b).
Environmental satisfaction in open-plan environments: 6. Satisfaction algorithms for software
(NRC/IRC Client Report B3205.24). Ottawa, ON: National Research Council of Canada, Institute
for Research in Construction.
O'Neill, M. J. (1992). Effects of workspace design and environmental control on office workers'
perceptions of air quality. Proceedings Of The Human Factors Society 36th Annual Meeting,
Atlanta GA (part 2, pp. 890-894). Santa Monica, CA: The Human Factors Society.
O'Neill, M. J., & Carayon, P. (1993). The relationship between privacy, control, and stress responses in
office workers. Proceedings of the Human Factors and Ergonomics Society 37th Annual
Meeting (pp. 479-483). Santa Monica, CA: Human Factors and Ergonomics Society.
Oldham, G. R. (1988). Effects of changes in workspace partitions and spatial density on employee
reactions: A quasi-experiment. Journal of Applied Psychology, 73(2), 253-258.
Oldham, G. R., & Brass, D. J. (1979). Employee reactions to an open-plan office: A naturally-occurring
quasi-experiment. Administrative Science Quarterly, 24, 267-284.
Oldham, G. R., Kulik, C. T., & Stepina, L. P. (1991). Physical environments and employee reactions:
Effects of stimulus-screening skills and job complexity. Academy of Management Journal, 34,
929-938.
Oldham, G. R., & Rotchford, N. L. (1983). Relationship between office characteristics and employee
reactions: A study of the physical environment. Administrative Science Quarterly, 28, 542-556.
Oldham, G. R., & Fried, Y. (1987). Employee reactions to workspace characteristics. Journal of Applied
Psychology, 72(1), 75-80.
Pedhazur, E. J. (1997). Multiple regression in behavioral research: Explanation and prediction (3rd ed.).
Fort Worth, TX: Harcourt Brace.
Reinhart, C. F. (2002). Effects of interior design on the daylight availability in open plan offices. In
Proceedings of the 2002 Summer Study of the American Commission for an Energy Efficient
Environment (ACEEE). Washington, D.C.: ACEEE.
Roche, L., Dewey, E., & Littlefair, P. (2000). Occupant reactions to daylight in offices. Lighting Research
and Technology, 32(3), 119-126.
Shaw, C.-Y., MacDonald, M. A., Galasiu, A. D., Reardon, J. T., & Won, D. Y. (2003). Experimental
investigation of ventilation performance in a mock-up open-plan office (NRC/IRC Client Report
B3205.19). Ottawa, ON: National Research Council of Canada, Institute for Research in
Construction.
IRC RR-154 50
Workstation and Physical Condition Effects on Environmental Satisfaction
Slater, A. I., & Boyce, P. R. (1990). Illuminance uniformity on desks: Where is the limit? Lighting
Research and Technology, 22( 4), 165-174.
"Space planning". (2003, September). Space planning and interior design report: How much smaller can
we go? Canadian Facility Management & Design, pp. 18-20,22.
Spreckelmeyer, K. F. (1993). Office relocation and environmental change: A case study. Environment
and Behavior, 25(2), 181-204.
Sundstrom, E. (1982). Privacy in the office. Environment and Behavior, 14(3), 382-389.
Sundstrom, E., Town, J. P., Brown, D. W., Forman, A., & McGee, C. (1982). Physical enclosure, type of
job, and privacy in the office. Environment and Behavior, 14(5), 543-559.
Sundstrom, E. (1987). Work environments: Offices and factories. In D. Stokols & I. Altman (Eds.),
Handbook of Environmental Psychology (pp. 733-782). New York: Wiley.
Sundstrom, E., Burt, R. E., & Kamp, D. (1980). Privacy at work: Architectural correlates of job
satisfaction and job performance. Academy of Management Journal, 23(1), 101-117.
Sutton, R. I., & Rafaeli, A. (1987). Characteristics of work stations as potential occupational stressors.
Academy of Management Journal, 30(2), 260-276.
Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics (4th ed.). Boston, MA: Allyn &
Bacon.
Veitch, J. A., Bradley, J. S., Legault, L. M., Norcross, S. G., & Svec, J. M. (2002). Masking speech in
open-plan offices with simulated ventilation noise: Noise-level and spectral composition effects
on acoustic satisfaction (IRC-IR-846). Ottawa, Ontario, Canada: National Research Council
Canada, Institute for Research in Construction.
Veitch, J. A., Farley, K. M. J., & Newsham, G. R. (2002). Environmental satisfaction in open-plan
environments: 1. Scale validation and method (IRC-IR-844). Ottawa, ON: National Research
Council Canada, Institute for Research in Construction.
Veitch, J. A., & Gifford, R. (1996). Assessing beliefs about lighting effects on health, performance, mood
and social behavior. Environment and Behavior, 28, 446-470.
Veitch, J. A., Hine, D. W., & Gifford, R. (1993). End-users' knowledge, beliefs, and preferences for
lighting. Journal of Interior Design, 19(2), 15-26.
Veitch, J. A., & Newsham, G. R. (1998). Lighting quality and energy-efficiency effects on task
performance, mood, health, satisfaction and comfort. Journal of the Illuminating Engineering
Society, 27(1), 107-129.
Veitch, J. A., & Newsham, G. R. (2000). Exercised control, lighting choices, and energy use: An office
simulation experiment. Journal of Environmental Psychology, 20(3 ), 219-237.
Warnock, A. C. C., & Chu, W. T. (2002). Voice and background noise levels measured in open offices
(IRC-IR-837). Ottawa, Ontario, Canada: National Research Council of Canada, Institute for
Research in Construction.
Wilkinson, L., Blank, G., & Gruber, C. (1996). Desktop data analysis with SYSTAT. Upper Saddle River,
NJ: Prentice-Hall.
Woods, J. E., Drewry, G. M., & Morey, P. R. (1987). Office worker perceptions of indoor air quality
effects on discomfort and performance. Indoor Air '87: The 4th International Conference on
Indoor Air Quality and Climate (pp. 464-468).
Yammarino, F. J., & Markham, S. E. (1992). On the application of within and between analysis: Are
absence and affect really group-based phenomena? Journal of Applied Psychology, 77, 168-176.
Zalesny, M. D., & Farace, R. V. (1987). Traditional versus open offices: A comparison of sociotechnical,
social relations, and symbolic meaning perspectives. Academy of Management Journal, 30, 240-
259.
Zweers, T., Preller, L., Brunekreef, B., & Boleij, J. S. M. (1992). Health and indoor climate complaints of
7043 office workers in 61 buildings in the Netherlands. Indoor Air, 2, 127-136.
IRC RR-154 51
Workstation and Physical Condition Effects on Environmental Satisfaction
Acknowledgements
This investigation forms part of the Field Study sub-task for the NRC/IRC project Cost-effective
Open-Plan Environments (COPE) (NRCC Project # B3205), supported by Public Works and Government
Services Canada, the Building Technology Transfer Forum, Ontario Realty Corp., USG Corp., British
Columbia Buildings Corp., Natural Resources Canada, and Steelcase, Inc. COPE is a multi-disciplinary
project directed towards the development of a decision tool for the design, furnishing, and operation of
open-plan offices that are satisfactory to occupants, energy-efficient, and cost-effective. Information
about COPE is available at http://irc.nrc-cnrc.gc.ca/ie/cope/.
The authors are grateful to the following individuals: Chantal Arsenault, John Bradley, Marcel
Brouzes, Natalie Brunette, Raymond Demers, Ryan Eccles, Tim Estabrooks, Brian Fitzpatrick, Ralston
Jaekel, Judy Jennings, Roger Marchand, Emily Nichols, and Scott Norcross (data collection); Louise
Legault (research design advice); Kelly Farley (data analysis); Gordon Bazana and Cara Duval (data
management). We also thank the management and employees in the nine buildings for their participation.
IRC RR-154 52
Workstation and Physical Condition Effects on Environmental Satisfaction
Appendix A: Derived Thermal Indices Regressions
Derived Thermal Indices
There exist derived indices for thermal comfort (predicted mean vote, PMV, and predicted
percentage dissatisfied, PPD) (Fanger, 1982) and discomfort caused by draught (Fanger, Melikov,
Hanzawa, & Ring, 1988). Further details of these indices can be found in Charles (2003). We calculated
these using the standard formulae, using common assumptions for the values that we had not measured.
PMV and PPD were calculated using a computer program provided in ISO Standard 7730
(International Organization for Standardization (ISO), 1984). For the variables that were not measured in
the COPE field study, the assumptions shown in Table XXX were used. The PMV index provides a
thermal comfort score ranging from –3 (cold) to +3 (hot). As a variable in this form would be difficult to
interpret in the regression analyses, we used PMV’s related index, the predicted percentage dissatisfied
(PPD) for these regressions.
Table A1. Assumptions for PMV and PPD calculations.
Variable Units Source
Metabolism W/m2 Standard value for office work: 70 (equivalent to 1.2 met) (ISO 1984;
ASHRAE, 1992)
External Work W/m2 Standard value for office work: 0 (ISO, 1984; ASHRAE, 1992)
Clothing m2.oC /W Standard value by season: spring 0.11, summer 0.08, fall 0.11, winter 0.16
(equivalent to clo values of 0.7, 0.5, 0.7, 1.0 respectively) (ISO, 1984)
The draught index was calculated using the formula provided in ASHRAE Standard-55
(American Society of Heating Refrigerating and Air Conditioning Engineers (ASHRAE), 1992), using an
assumed turbulence intensity of 35% (as recommended in this Standard). The draught index was
calculated at both the head and ankle heights, these being the areas most susceptible to draught discomfort
(American Society of Heating Refrigerating and Air Conditioning Engineers (ASHRAE), 1992).
However, to be consistent with the previous regression analyses using the raw variables, we used only the
draught index calculated at head height (DRAUGHT_H) as a regression predictor.
We then repeated the analyses that we had earlier conducted using the raw measurements of
ventilation conditions, using these derived indices in their places.
Predicting Satisfaction with Ventilation
Table A2 shows the results for the regression using derived thermal indices in place of the raw
measurements (shown in Table 19). The percentage of variance explained is slightly lower (12% instead
of 14%). Whereas both air movement and temperature were significant predictors in the raw
measurements model, only draught is significant here. The magnitude and direction of the other effects
are consistent with the raw measurements model.
IRC RR-154 53
Workstation and Physical Condition Effects on Environmental Satisfaction
Table A2. Summary table for Sat_Vent regressed on workstation characteristics and derived ventilation indices.
AGE_COMBINE
D
-.054 -.015 -.020 -.020 -.010
GENDER .184*** .196*** .208*** .207*** .216***
ADMIN -.058 .000 -.004 -.005 -.005
MGR .008 .028 .020 .028 .014
PROF .001 .062 .053 .055 .059
SQRTAREA -.040 -.058 -.080 -.084
MINPH_NOOPEN -.157*** -.170***
WINDOW
-.185*** -.174***
PANELS_CAT -.022 -.013 -.009 -.013
-.095* -.105** -.110** -.111**
PPD .079* .086* .060
DRAUGHT_H -.086* -.092* -.091*
DL_OUT -.076 -.082*
FDCO2 -.116**
POLLUT .000
R2 change .048*** .050*** .010* .005 .012**
Total R2 .048*** .098*** .109*** .114*** .126***
Adjusted R2 .041*** .087*** .095*** .098*** .108***
Note. N= 713. * p<=.05. **p<=.01. ***p<=.001.
Predicting Overall Environmental Satisfaction
Overall environmental satisfaction results are shown in Table A3. the comparable analysis for the
raw measurements is shown in Table 26. The percentage of variance explained in the two models was the
same, but the significant predictors are not the same. Although neither temperature nor humidity were
statistically significant predictors in the raw measurements model, PPD is significant here (Table A3). As
expected, conditions leading to higher PPD predictions are associated with higher OES. This is a very
small effect, with PPD and draught together contributing only 0.7% to the explained variance.
Table A3. Summary table for OES regressed on workstation characteristics and derived ventilation indices.
AGE_COMBINE
D
-.026 -.029 -.032 -.031 -.027
GENDER .043 .042 .049
-.086
.047 .050
ADMIN .065 .071 .063 .063 .064
MGR -.041 -.045 -.052 -.038 -.043
PROF -.090 -.078 -.082 -.080
SQRTAREA .097 .081 .043 .044
-.131* -.137** -.159** -.158**
PANELS_CAT -.034 -.025
WINDOW -.024 -.033 -.042 -.044
PPD .084* .096* .089*
FDCO2 -.040
POLLUT -.006
R2 change .019* .011 .007 .013** .002
Total R2 .019* .030* .037** .050*** .051***
Adjusted R2 .012* .017* .022** .033*** .032***
MINPH_NOOPEN
-.020 -.021
DRAUGHT_H -.047 -.058 -.058
DL_OUT -.126** -.127**
Note. N= 689. * p<=.05. **p<=.01. ***p<=.001.
IRC RR-154 54
Workstation and Physical Condition Effects on Environmental Satisfaction
Predicting Job Satisfaction
Results for the regressions involving job satisfaction are shown in Table A4 for the derived
indices, and Table 33 for the raw measurements. They explain similar amounts of variance (6.7 % and
6.9%), but in the raw measurements model both carbon dioxide and other pollutants were statistically
significant predictors. The other predictors are the same for both models.
Table A4. Summary table for JobSatis regressed on workstation characteristics and derived ventilation indices.
AGE_COMBINE
D
-.135*** -.120** -.120** -.120** -.107**
GENDER .001 .003 .005 .004 .013
ADMIN
.026
-.053 -.069 -.055
-.052 -.012 -.009 -.010 -.006
MGR .029 .030 .036 .015
PROF -.022 .022 .022 .024 .030
SQRTAREA -.057
MINPH_NOOPEN -.092 -.107* -.113*
PANELS_CAT
-.095
-.040 -.042 -.040 -.037
WINDOW .050 .053 .049 .035
PPD -.030 -.024 -.029
DRAUGHT_H -.010 -.015 -.015
DL_OUT -.058 -.059
FDCO2
-.079
R2 change .023** .001 .003 .021***
Total R2 .023** .044*** .045*** .048*** .069***
Adjusted R2 .016** .032*** .030***
-.121**
POLLUT
.022**
.031*** .050***
Note. N= 695. * p<=.05. **p<=.01. ***p<=.001.
Summary
The derived indices rely on several assumptions for their calculation from field data such as this,
which one would expect would reduce their reliability. Overall, the results for the analyses with the raw
data seem to be more useful, given that the raw measurements allow more precise statements about the
causes of effects and that more predictors reached statistical significance in those models.
IRC RR-154 55
... Some studies have reported reverberation times (T in seconds) [2,5], and psychoacoustic metrics [9] (e.g., loudness, etc.) [6,10,11]. The other major group of related metrics include those derived from octave/one-third octave band spectra such as noise rating, preferred noise criterion, balance between spectral regions, etc [2,6,[10][11][12][13][14]. ...
... The surface area in Table 1 represents the portion of the entire floor plate consisting of OPOs only and does not include areas for building services (elevator lobbies, plant rooms), kitchens, enclosed rooms for meetings, personnel, etc., as long as these were not within the office perimeter. This may partly account for smaller office areas reported in Table 1 than some previous studies, e.g., [5,14,16]; also see Appendix A. ...
... Those based on octave/one-third octave band filtering. Besides giving an averaged spectral representation, the filtered data can be used to derive certain metrics used in previous studies [6,10,14]. These metrics include the noise rating (NR) [21], room criterion Mark II (RC) [22], balanced noisecriterion (NCB) [23], and the difference between the Aweighted SPL averages of the low (16-63 Hz) and high (1000-4000 Hz) one-third octave band decibel levels (Lo-Hi) [24]. ...
Article
Open-plan offices (OPOs) have been around for more than half a century now, chronicling the vicissitudes of workplace topography amongst other factors. This paper addresses one such factor – the sound environment in occupied OPOs in relation to several objective workplace parameters, using measurements in contemporary OPOs and comparisons with studies over the last 50 years. Omnidirectional and binaural sound measurements were conducted in 43 offices during typical working hours. The results describe variation in several acoustic and psychoacoustic metrics, and present statistical models that predict these metrics as a function of the number of workstations in offices. LA,eq of 53.6 dB is typical for occupied OPOs, with spectral slope of approximately −4 dB/octave. LA,eq values do not vary much over the workplace parameters studied (e.g., floor plate area, work activity, etc), except for −2.7 dB and −4.1 dB differences between offices with/without carpeting, and offices with ceiling absorption but with/without carpeting, respectively; most likely from reduced floor impact noise leading to speech level reduction. Sound fluctuation, as characterised by the metric Noise Climate (NCl: LA10 – LA90) and the psychoacoustic Fluctuation Strength (FS), decreases significantly with increasing number of workstations in OPOs. This suggests lesser auditory distraction in larger offices, which needs further investigation. In terms of historical trends, OPOs have become quieter over the years, especially background noise quantified as LA90, although there are several subtleties. Overall, current findings can inform several OPO design perspectives including policy documents, provide values for laboratory simulations of OPO acoustic environments, help interpret subjective impressions of OPO occupants, etc.
... Generally, open-plan office layouts are believed to facilitate communication and interaction among co-workers by reducing environmental distance and promoting team-work and creativity (Duval, Veitch, & Charles, 2002;Kim & de Dear, 2013;O'Neill, 2008;Veitch, Charles, Farley, & Newsham, 2007;Veitch, Charles, Newsham, Marquardt, & Geerts, 2003). One basic reason for the creation of open-plan office arrangements was to increase social interaction and communication among team members and co-workers. ...
... Social density refers to the number of individuals inhabiting a place, and spatial density refers to the size of the place or the amount of usable space per employee (Brennan et al., 2002;Duval , 2002). Some researchers have suggested that increased social density and closeness in the workplace lead to improved opportunities for friendship and facilitate interpersonal contact, information exchange, and communication and promote job satisfaction (Kim & de Dear, 2013;Maher & Von Hippel, 2005;Veitch et al., 2003). On the other hand, some researchers have reached the opposite conclusion, and they suggested that satisfaction with the work environment may decrease in places with high social and spatial density because of the lack of privacy and the uncontrolled social contact (Baldry & Barnes, 2012;Duval et al., 2002;Jahncke, 2012;Rashid, Wineman, & Zimring, 2009). ...
... Previous studies have indicated that some environmental variables such as the arrangements or design of the spaces, flexibility, size, and proximity of the workstation (which is more common and possible in share and open-plan offices) can facilitate the probabilities or limit the social occurrence (Candido et al., 2019;Duval et al., 2002;Kim & de Dear, 2013;O'Neill, 2008;Veitch et al., 2007;Veitch et al., 2003). As mentioned previously, people's behavior and actions are affected by the physical design of the building. ...
Article
Full-text available
This study aims to investigate the influence of the work environment in open‐plan workplaces on employees’ work‐related behavior (social interaction and communication) and their outcome (creativity). A quantitative methodology implying a cross‐sectional survey was used to fulfill the study's purposes in addition to the literature review. The Partial Least Square of Structural Equation Model (PLS‐SEM) was used to analyze data. The respondents were 117 staff employees working as programmers and video editors in creative multimedia companies in Tehran. The findings of this study revealed that although employees in an open‐plan office do not have the ability to control environmental features, a great deal of work‐related behavior and creativity is called for in their daily work. In fact, open‐plan spaces improve social interaction among co‐workers by increasing adjacency and reflect a more democratic working culture. Managers and space designers can use the findings of this study to boost workplace facilities in order to positively affect employees’ work behavior and outcome.
... Previous studies in environmental behavior indicated that satisfaction with the work environment is a key sign of employee well-being and performance at work (Mike O'Neill, 2008;Schakib-Ekbatan, Wagner, & Lussac, 2010;Tanabe, Haneda, & Nishihara, 2015;Van der Voordt, 2004;Veitch, Charles, Farley, & Newsham, 2007;Veitch, Charles, Newsham, Marquardt, & Geerts, 2003). Some factors are related to individual well-being and satisfaction (both job and environment) in the work environment. ...
... Previous studies have indicated the psychosocial effect of the work environment on employees' health and well-being (Briner, 2000;Edwards, Cable, Williamson, Lambert, & Shipp, 2006;George & Brief, 1992). Furthermore, prior studies support the finding of this study and suggested that environmental satisfaction is considered as a key indicator of employees' well-being and performance at work ( Van der Voordt, 2004;Veitch et al., 2007;Veitch et al., 2003;Yee, Yeung, & Cheng, 2008). Previous studies suggested that personal control over the workstation is directly related to group efficiency, teamwork, and collaboration (Hua, 2007;Hua, Loftness, Heerwagen, & Powell, 2011;Lee & Brand, 2005;Lee & Brand, 2010). ...
Article
Full-text available
Recently the modern focus of management, human resource, and health research were on office design and reducing or minimizing workplace problems and overall cost for office design. Thus, the concepts of open-plan office design have been defined as providing at least a basic solution to many of these notable and current challenges in current working systems and organizations. In fact, open workspace are often suggested since they offer more flexibility for dynamic organizational changes and fast technological growth in the modern working style. Today because of the spread of Corona-virus disease 2019 (COVID-19) there is a need to focus even on new open-plan office designs to prevent and minimize health risks. The fact is the world needs to be ready for new viruses and or perhaps to unknown diseases in the future. The role of the workplace and the physical environment that affects employees’ perceptions about the work environment has always been challenging. In fact, managers and space designers need to pay careful attention to designing the work environment with a supportive workspace is a beneficial activity to promote individuals’ perceptions about the work environment, satisfaction, and outcome. In this condition, employees feel valued by the organization and management. Consequently, the role of environmental features in affecting employees’ perceptions about the workplace, environmental satisfaction, and overall outcome is remarkable and must be understood by managers and space designers.
... We developed two items to capture whether prison officers perceive that they have appropriate material resources to do their job effectively (α = .92). Althoug