University of Southern California Dissertations and Theses

Considering occupants: comprehensive POE research on office environment of Southern California

(USC Thesis Other)

Considering occupants: comprehensive POE research on office environment of Southern California

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Considering Occupants:

Comprehensive POE Research on Office Environment of Southern California

By

Jehyun Moon

Presented to the
FACULTY OF THE
SCHOOL OF ARCHITECTURE
UNIVERSITY OF SOUTHERN CALIFORNIA
In partial fulfillment of the
Requirements of degree
MASTER OF BUILDING SCIENCE

AUGUST 2016

1

COMMITTEE

Joon-Ho Choi, Ph.D., LEED AP
Assistant Professor
USC School of Architecture
joonhoch@usc.edu
213-740-4576

John V. Mutlow
Professor
USC School of Architecture
mutlow@gmail.com
213-821-0749

Marc Schiler
Professor
USC School of Architecture
marcs@usc.edu
213-740-4591

2

ABSTRACT
Rapidly developing technologies and growing demands of people for a better
quality of life have enhanced the collaboration between design research and practice. In
this context, Post Occupancy Evaluation (POE) and its research are one of the most well-
developed examples of the research-practice feedback system. However, the weakness of
the POE is often pointed out its excessive reliance on the survey and general-targeted
solution. This weakness may lead to irrelevant modification and occupants’ dissatisfaction
with the indoor environmental quality (IEQ). As a solution, comprehensive POE is
suggested. Quantifiable data such as IEQ measurements and occupants’ satisfaction is
collected from over 400 workstations in modern buildings of Southern California through
on-site measurement and satisfaction surveys on the occupants. Statistical analyses are
conducted on collected data within specified categories of building types, building
attributes, and human factors. The analysis not only correlates the influence on one another
but also allows sorting the factors by the intensity of the impact. The results verified the
necessity of the modified POE regarding practice-based approach and specific-targeted
solution. Based on findings, specific recommendations and strategic design guidelines are
suggested to help enhance the environmental conditions in working environment.

HYPOTHESIS
Within the working environment of an office setting, identifying the impact of
individual IEQ components on occupants’ environmental satisfaction by human and spatial
factors can help maximize outcomes within limited resources such as time and budget.
3

TABLE OF CONTENTS
LIST OF FIGURES .......................................................................................................... 5
LIST OF TABLES ............................................................................................................ 7
Chapter 1: Introduction ................................................................................................... 8
1.1 Indoor Environmental Quality (IEQ) Research ................................................................ 8
1.2 Post Occupancy Evaluation (POE) as IEQ Research Tool .......................................... 9
1.2.1 Importance of POE ...................................................................................................... 9
1.2.2 Limitation in POE ..................................................................................................... 10
1.2.3 POE in Office Environment ................................................................................. 11
1.3 Terms .................................................................................................................................................. 12
1.3.1 Indoor Environment Quality ................................................................................ 12
1.3.2 Post Occupancy Evaluation ................................................................................. 13
1.3.3 Human Comfort in Indoor Environment ....................................................... 13
1.4 Scope of the Research ................................................................................................................ 15
1.5 Structure of the Thesis ................................................................................................................ 16

Chapter 2: Background and Literature Review .......................................................... 17
2.1 Comprehensive POE Research .............................................................................................. 17
2.1.1 Integrated IEQ Study ............................................................................................... 18
2.1.2 Conventional Research Goals of POE ............................................................ 21
2.1.3 Impact of POE on Occupants in Workspace Environment ................... 24
2.2 IEQ Indicators ................................................................................................................................ 27
2.2.1 Visual Quality ............................................................................................................. 27
2.2.2 Thermal Quality ......................................................................................................... 29
2.2.3 Indoor Air Quality Indicators .............................................................................. 31
2.2.4 Acoustic Quality Indicators ................................................................................. 32

Chapter 3: Research Objectives and Methodologies ................................................... 34
3.1 Research Objectives .................................................................................................................... 35
3.2 Methodologies ................................................................................................................................ 36
3.2.1 Building Attributes Observation ........................................................................ 36
3.2.2 IEQ Measurement ..................................................................................................... 36
4

3.2.3 Occupants’ Satisfaction Survey ......................................................................... 37
3.2.4 Statistical Analysis ................................................................................................... 40

Chapter 4: Data Description .......................................................................................... 44
4.1 Data Distribution ........................................................................................................................... 44
4.1.1 Data Distribution by Human Factors ............................................................... 44
4.1.2 Data Distribution by Spatial Factors ................................................................ 45
4.2 Measured and Surveyed Data ................................................................................................. 46
4.2.1 IEQ Measurement ..................................................................................................... 46
4.2.2 Satisfaction Survey ................................................................................................... 54

Chapter 5: Data Analysis and Findings ........................................................................ 57
5.1 Statistical Analysis ....................................................................................................................... 57
5.1.1 Identifying Impact Questions .............................................................................. 57
5.1.2 Correlation between IEQ and Occupants’ Satisfaction .......................... 64
5.1.3 Analyses by Different Occupant Group ......................................................... 66
5.2 Impact of Factors on Occupants’ Satisfaction ................................................................ 71
5.2.1 Impact of Gender....................................................................................................... 71
5.2.2 Impact of Age ............................................................................................................. 72
5.2.3 Impact of Building Type ....................................................................................... 74
5.2.4 Impact of Workstation Location ........................................................................ 76
5.3 In-depth Analysis Using an Advanced Data Mining Tool ........................................ 78

Chapter 6: Conclusion .................................................................................................... 85
6.1 Discussion ........................................................................................................................................ 85
6.2 Limitations and Future Directions ........................................................................................ 86
6.3 Conclusion ....................................................................................................................................... 87

Appendix: Satisfaction Survey by Question ................................................................. 89

Bibliography .................................................................................................................. 102

5

LIST OF FIGURES
Figure 1: Methodologies of the Research ......................................................................... 34
Figure 2 Distribution of Work Surface Illuminance ......................................................... 46
Figure 3 Distribution of Screen Illuminance .................................................................... 47
Figure 4 Distribution of Reading Zone Illuminance ......................................................... 48
Figure 5 Distribution of Average Luminance ................................................................... 49
Figure 6 Distribution of UGR ........................................................................................... 49
Figure 7 Distribution of Ambient Temperature ................................................................ 50
Figure 8 Distribution of Air Velocity at 1.1m Height ...................................................... 51
Figure 9 Distribution of Air Velocity at 0.1m Height ...................................................... 52
Figure 10 Distribution of CO2 Level ................................................................................ 52
Figure 11 Distribution of Acoustic Decibel ...................................................................... 53
Figure 12 Rose Chart of Survey Questions....................................................................... 55
Figure 13 Rose Chart of Survey Questions by IEQ Criteria ............................................. 56
Figure 14 Interval Plot of Data by Different Point Scale.................................................. 66
Figure 15 Interval Plot of Data by Human Factor ............................................................ 68
Figure 16 Interval Plot of Data by Spatial Factor ............................................................. 69
Figure 17 Interval Plot of Average Luminance Level by Gender .................................... 71
Figure 18 Interval Plot of Air Velocity at 1.1m by Gender .............................................. 72
Figure 19 Interval Plot of Work Surface Illuminance by Age Group............................... 73
Figure 20 Interval Plot of Air Velocity at 1.1m by Age Group ........................................ 74
Figure 21 Interval Plot of Work Surface Illuminance by Building Type ......................... 75
Figure 22 Interval Plot of Air Velocity at 1.1m by Building Type .................................. 76
Figure 23 Interval Plot of Reading Zone Illuminance by Workstation Location ............. 77
Figure 24 Interval Plot of Air Velocity at 0.1m by Workstation Location ....................... 78
Figure 25 Decision Tree (J48) from Entire Survey Questions ......................................... 80
Figure 26 Decision Tree (J48) from 5 Selected Survey Questions ................................... 81
Figure 27 Decision Tree (J48) from Selected Survey Questions and IEQ Measurements 82
6

Figure 28 Decision Tree (J48) from Selected Survey Questions and Impact Factors ...... 83
Figure 29 Decision Tree (J48) from Combined Data ....................................................... 84
Figure 30 Distribution of Answers to Q6 and Box Plot.................................................... 89
Figure 31 Distribution of Answers to Q7 and Box Plot.................................................... 90
Figure 32 Distribution of Answers to Q8 and Box Plot.................................................... 90
Figure 33 Distribution of Answers to Q9 and Box Plot.................................................... 91
Figure 34 Distribution of Answers to Q10 and Box Plot.................................................. 91
Figure 35 Distribution of Answers to Q11 and Box Plot.................................................. 92
Figure 36 Distribution of Answers to Q12 and Box Plot.................................................. 92
Figure 37 Distribution of Answers to Q13 and Box Plot.................................................. 93
Figure 38 Distribution of Answers to Q14 and Box Plot.................................................. 93
Figure 39 Distribution of Answers to Q15 and Box Plot.................................................. 94
Figure 40 Distribution of Answers to Q16 and Box Plot.................................................. 94
Figure 41 Distribution of Answers to Q17 and Box Plot.................................................. 95
Figure 42 Distribution of Answers to Q18 and Box Plot.................................................. 95
Figure 43 Distribution of Answers to Q19 and Box Plot.................................................. 96
Figure 44 Distribution of Answers to Q20 and Box Plot.................................................. 96
Figure 45 Distribution of Answers to Q21 and Box Plot.................................................. 97
Figure 46 Distribution of Answers to Q22 and Box Plot.................................................. 97
Figure 47 Distribution of Answers to Q23 and Box Plot.................................................. 98
Figure 48 Distribution of Answers to Q24 and Box Plot.................................................. 98
Figure 49 Distribution of Answers to Q25 and Box Plot.................................................. 99
Figure 50 Distribution of Answers to Q26 and Box Plot.................................................. 99
Figure 51 Distribution of Answers to Q27 and Box Plot................................................ 100
Figure 52 Distribution of Answers to Q28 and Box Plot................................................ 100
Figure 53 Distribution of Answers to Q29 and Box Plot................................................ 101

7

LIST OF TABLES
Table 1: Type of Measured IEQ Data ............................................................................... 37
Table 2 Occupants’ Satisfaction Survey Questions .......................................................... 38
Table 3 Sensational Scale for Occupants’ Environmental Satisfaction Survey ............... 40
Table 4 Data Distribution by Human Factors ................................................................... 44
Table 5 Data Distribution by Spatial Factors .................................................................... 45
Table 6 Stepwise Regression on Q29 vs. Q6 - 28 ............................................................. 58
Table 7 Stepwise Regression on Q29 vs. Q15, Q26 - 28 .................................................. 59
Table 8 Stepwise Regression on Q27 vs. Q18 – 22: Lighting .......................................... 60
Table 9 Stepwise Regression on Q29 vs. Q18 – 22: Lighting .......................................... 60
Table 10 Stepwise Regression on Q26 vs. Q13 – 14, Q25: Temperature and Air Quality ... 61
Table 11 Stepwise Regression on Q15 vs. Q13 – 14, Q25: Temperature and Air Quality ... 61
Table 12 Stepwise Regression on Q29 vs. Q13 – 14, Q25: Temperature and Air Quality ... 61
Table 13 Stepwise Regression on Q28 vs. Q16 – 17: Acoustic Quality........................... 62
Table 14 Stepwise Regression on Q29 vs. Q16 – 17: Acoustic Quality........................... 62
Table 15 Stepwise Regression on Q29 vs. Q13, Q16, Q22, Q25 ..................................... 63
Table 16 Correlation Analysis between IEQ Measurement and Satisfaction Survey ...... 65
Table 17 IEQ Measurement and Survey Combinations Analyzed ................................... 70

8

Chapter 1: Introduction
This chapter contains introductory information on indoor environmental quality
(IEQ) research and post-occupancy evaluation (POE) to determine the starting point of the
research. The terms used in the research and the scope of the research are explained. In the
end, the structure of the thesis is briefly observed to help understand the reading.

1.1 Indoor Environmental Quality (IEQ) Research
Indoor Environmental Quality (IEQ) simply can be described as the conditions
inside the building. This includes comfort in lighting, thermal, air quality, and acoustics.
The satisfaction of building occupants can be increased through improving these various
aspects of IEQ taken into consideration multilaterally. Improving IEQ of the building has
been proven to have a significant impact on building occupants’ health and comfort by
taking account of indoor air quality and thermal, visual, and acoustic quality. Several
researchers have identified that high-quality indoor environments enhance productivity,
decrease absenteeism, and improve the value of the building [1,2]. Moreover, reaching
higher quality in indoor environment of the building is likely to have an indirect effect of
energy saving and reduction in CO2 emissions, achieved by adapting energy efficient
products and state-of-the-art technologies [3]. Because people spend 87% of their time in
enclosed buildings [4], it is evident that consideration of IEQ in building design is
inevitable.

Some researches in the topic of IEQ can be reached from the diverse domain and
fields due to its complexity and variety of relevant factors. Regarding air quality, for
9

instance, building occupants may be exposed to a variety of contaminants from office
machines, cleaning facilities, activities, carpets, furnishings, cigarette smoke, water-
damaged materials, fungal, mold, insects, and outdoor pollutants [5]. Each of listed factors
has its impact on air quality that numerous researches are done on the topic. Similarly, the
impact of other factors such as lighting, temperatures, relative humidity, ventilation levels,
and acoustics on IEQ are also well-known topics on indoor environment, which shows the
significance of research in IEQ.

1.2 Post Occupancy Evaluation (POE) as IEQ Research Tool
1.2.1 Importance of POE
To understand and improve the Indoor Environmental Quality (IEQ) of the space,
Post Occupancy Evaluation (POE) has been used as the primary methodology for
assessment. Most commonly used definition of POE is “a process of obtaining feedback
on a building’s performance that allows designers, developers, owners, and occupants to
identify objectively flawed building services and design features, and enhance successful
design” [6]. Compared to the previous trend of focus on measuring and regulating the
resource efficiency of buildings, POE takes into consideration of building occupants.

According to Preiser, one of the pioneers of the POE, the main goal of carrying out
POE in the building is (1) improvement of the comfort levels of the occupants and (2)
maintenance of the HVAC systems for better indoor conditions, while reducing energy
costs. The occupant’s comfort level is comprised of indoor lighting, air quality, thermal,
and acoustics. The objective also includes understanding the occupant’s perspective about
10

the building and its performance. Building performance is also a key factor that should be
taken into account in an evaluation process [7].

According to Meir, methods used in POE include plan analysis, monitoring of IEQ,
and surveys including walkthroughs, observations, user satisfaction questionnaires, and
interviews. POE was introduced in 1960, in response to significant problems experienced
in building performance with particular emphasis on the building occupant’s perspective.
POE is a way of providing subjective and objective feedback that can inform planning and
practice throughout the building’s lifecycle from the initial design to the occupation. Meir
describes the benefits from POE in the short, medium and long term. The short-term
benefits include users’ feedback on problems in building and the identification of solutions.
Medium-term benefits include feed-forward of the positive and negative lessons learned
from the next building cycle. Long-term benefits are creating databases and the update and
generation of planning and design protocols and paradigms [8].

1.2.2 Limitation in POE
Post occupancy evaluation is a useful tool for the built environment professions.
By finding how users are making use of spaces, both the strengths and weaknesses of
implemented proposal can be determined. It can then build on the strengths and work to
limit the deficiencies. However, practitioners have identified several potential problems in
carrying out post-occupancy evaluations. This includes the balance of potential costs and
potential problems of liability. Additional barriers include: (1) occupants and users
changing the way the space is used in unexpected ways, making it difficult to judge its
11

performance against the original brief, (2) the fact that post-occupancy evaluation by itself
is insufficient to ensure a good performance for it is too late if the problems have already
occurred, (3) the designers of public spaces have often moved on to new projects and so
are not always available to benefit from the feedback loop arising from a post-occupancy
evaluation, and (4) clients believe that feedback on a past project will not help them, but
rather it will be of greatest assistance to the future customers of the design team [9].

Once again, Meir explains that POE researchers are often regarded with suspicion
and even hostility since the research may cause friction between different stakeholders such
as architects, consultants, clients, owners, managers, and users. This institutional and
professional split of authorities, perspectives and liabilities have hindered the use of POE
as a self-evident tool for the design and construction industry [8].

The biggest challenge, however, is to collect, analyze and correctly interpret
occupants’ feedback in a systematic and useful manner. As a matter of a fact, weakness of
the POE is often pointed out in its credibility due to its excessive reliance on subjective
survey data. Also, POE often focuses on generalized target that consideration on impact
factors such as human factor or spatial factor is neglected. Therefore, enhancing its
objectivity in POE and identifying impact factors are necessary.

1.2.3 POE in Office Environment
In the words of office design consultant Francis Duffy, "The office building is one
of the great icons of the twentieth century. As the most evident index of economic, social,
12

technological, and financial activities, buildings have become a symbol of what this
century has been about." The quote is right because the office buildings are the most
tangible reflection of a profound change in employment patterns that has occurred in last
hundred years. In the United States, almost 50 percent of the working population is
employed in office settings which shows significant growth as compared to 5 percent of
the population at the beginning of the 20th century [10].

Meanwhile, the USGBC (United States Green Building Council) has created the
most commonly used system for rating green buildings, LEED (Leadership in Energy and
Environmental Design) system. Until 2014, over 13,000 buildings have received Platinum
or Gold certification worldwide. With the generalization of the rating system, there is
agreement that LEED has brought sustainability into the consideration of building design
and construction of modern buildings. However, it is less clear on whether it has reduced
the adverse effect of buildings on the environment or improved indoor environments for
the occupants. A comprehensive post-occupancy evaluation (POE) study in commercial
buildings is useful to establish the success and shortcomings of the current rating system,
and to guide future editions [11].

1.3 Terms
1.3.1 Indoor Environment Quality
Indoor Environmental Quality (IEQ) is most essentially portrayed as the conditions
inside the building. It incorporates access to sunlight and perspectives, thermal quality, air
quality, acoustic conditions, and inhabitant control over lighting and warm solace. It might
13

likewise incorporate the practical parts of space, for example, whether the format gives
simple access to instruments and individuals when required and whether there is adequate
space for occupants [12]. The fulfillment of occupants can be expanded by considering
integrated IEQ instead of merely concentrating on temperature or air quality alone.

1.3.2 Post Occupancy Evaluation
Post Occupancy Evaluation (POE) is the process of obtaining feedback on a
performance of buildings in use. The value is being increasingly recognized and becoming
mandatory on public projects. It is valuable in all building construction sectors, especially
healthcare, education, commercial buildings and general housing. Poor building
performance may impact running costs, occupant well-being, and business efficiency. POE
is significantly different from conventional surveys and market researches. It is based on
the direct experiences of building users for evaluating how a building is operated for
intended use. POE can be conducted for many purposes, including calibrating new
buildings, developing new facilities and managing problematic buildings. Organizations
also find its value when establishing maintenance, replacement, purchasing or supply
policies; preparing for refurbishment; or selecting accommodation for purchase or rent [13].

1.3.3 Human Comfort in Indoor Environment

As an importance of IEQ on occupants’ comfort and welfare grows, many of
current sustainable building designs take the issue of IEQ and comfort into consideration.
To achieve quality indoor environment, IEQ is focused on comfort from four different
perspectives: visual, thermal, indoor air, and acoustic quality.
14

As visual comfort is part practical and part aesthetic, it employs such strategies as
artificial lighting, day lighting and creating visually interesting environments. Visual
comfort is commonly measured by illumination levels and its distribution. Based on
information provided by Autodesk Sustainability Workshop, this measure not only
includes the brightness of light sources but also how well light is spread around spaces and
the colors in the light. The goal of achieving visual comfort is to illuminate without using
too much energy or glare. Good lighting design achieves visual comfort by adequate
daylight and artificial light [14].

The current design for thermal comfort is based on ASHRAE Standard 55. In
commercial buildings, occupants’ comfort should frequently be assessed, using
ASHRAE’s Performance Measurement Protocols for Commercial Buildings. Based on
introduction from building science research group, there are six factors to consider in
evaluating the conditions for thermal comfort. The analysis includes measuring of air
temperature, relative humidity, surface temperatures, occupants’ metabolic rates, the
amount of clothing worn, and air speed across body surfaces. These comfort criteria are
made based on ASHRAE Standard 55 [15].

Indoor Air Quality (IAQ) is a state of the air within space. Indoor space with good
air quality is low in toxins, contaminants, and odors. Good air quality can be reached when
spaces are well ventilated with outdoor air and protected from pollutants. Strategies such
as bringing in 100% outside air, maintaining proper exhaust systems complying with
ASHRAE Standard 62.1 are used to create good IAQ. Also, utilizing high-efficiency
15

MERV filters in the heating ventilation and air conditioning (HVAC) system, prohibiting
smoking, installing walk-off mats at entryways, providing indoor plants, and using non-
toxic materials and green housekeeping products are recommended [12].

Acoustical control is also an important method of enhancing occupants’ comfort in
indoor space. People tends to be more productive and happy when they are not distracted
by noises from the outside or surrounding environment. Acoustic comfort is especially
important for office buildings. Even though, how humans perceive sounds and loudness is
a subjective measure, the comfortable environment can be created by controlling objective
measures like decibel level, reverberation time, and the sound reflection and damping
properties of materials [16].

1.4 Scope of the Research
The research focuses on the factors that affect human comfort at workstations
within an office environment in Southern California. Comprehensive POE is conducted to
collect data including building attributes, IEQ measurement, and occupants’ satisfaction
survey. Statistical analysis is carried out to identify the degree of impact on one another.
Human comfort is observed in four different aspects: visual, thermal, indoor air quality and
acoustic comfort. The research only used the general meaning of office environment from
the dictionary, “a room or building that are used as a place for commercial, professional,
or governmental work”. Thus, the age of people who participated in research ranges
between 18 and 69. As a spatial factor, type of buildings and workstation location is used,
16

but it does not consider detailed settings such as floor plan of office, the number of furniture
and device within the workstation, or exact location of the workstation within the building.

1.5 Structure of the Thesis
Chapter 1 is the introduction to indoor environmental quality (IEQ) research and
the post-occupancy evaluation (POE) as a tool for the research, especially in office
environments. Chapter 2 provides the background and literature review on comprehensive
POE research from three different aspects. Also, IEQ indicators are explained based on
four different human comfort criteria. Chapter 3 describes the research objectives and
corresponding methodologies: building attributes observation, IEQ measurement,
occupants’ satisfaction survey, and statistical analysis. Chapter 4 shows collected data by
IEQ measure and each question used in a satisfaction survey. Distribution of the data is
also described by human factors and spatial factors of the data. In Chapter 5, the result of
multiple statistical analysis and findings are explained. In the end, decision-making tool is
suggested as a tool based on the findings. Finally, Chapter 6 concludes the thesis with
discussing limitations and future improvements for the research.
17

Chapter 2: Background and Literature Review
This chapter list and summarize the background and literature reviewed to explain
fundamentals of comprehensive POE research and IEQ indicators that are used to
determine and relate human comfort with IEQ measurement.

2.1 Comprehensive POE Research
Research on Indoor Environmental Quality (IEQ) has proven its significant
influence on occupants’ health and productivity regarding air quality, lighting, acoustic,
and thermal comfort [1,2]. To understand and to improve the IEQ of the space, post-
occupancy evaluation (POE) has been used as the primary methodology in assessment
[8,17]. POE is a systematic evaluation process on the performance of building after it is
built and occupied for some time [7]. However, weakness of the POE is often pointed out
by researchers on its excessive reliance on subjective surveys and solution on generalized
occupants [18]. In order to supplement deficiencies, various types of comprehensive POE
research have been developed and suggested by researchers.

Background research on comprehensive POE research is approached from three
different directions: (1) POE research with integrated IEQ study through actual
measurement of the indoor environment and occupants’ satisfaction survey, (2)
conventional research goals of POE, and (3) impact of POE on occupants’ workspace
environment. Comprehensive POE investigated using observations and walkthroughs,
monitoring, survey questionnaires and plan analysis.
18

2.1.1 Integrated IEQ Study
One of the current trends in comprehensive POE research is to strengthen its
objectivity by incorporating IEQ measurement within POE. Buildings are very complex
systems, and its interaction with occupants may cause complex interrelations and
malfunctions. Therefore, it is compulsory that study of building post occupancy should be
based on a multi-level system of checks and tests. The system should involve thermal
comfort along with heating, ventilate, and air-conditioning. Also, physiological and
psychological comfort should be considered since these issues will affect energy
consumption and human comfort. Following are some of the exemplary projects done by
various researchers on integrated IEQ study.

Liang compared monitored IEQ variables with occupants’ satisfaction surveys to
improve indoor environmental quality in green office buildings in Taiwan. The research
explained that the green building certification system in Taiwan evaluates the performance
of buildings in lighting, ventilation, acoustics, and decoration but not in the performance
of delivering thermal comfort. The study investigated and compared the green conventional
office buildings in middle Taiwan on aspects of IEQ during a period of the active cooling
season. Among the monitored variables, the levels of noise, illumination, and CO2 in both
building types were meeting the international or Taiwan's regulatory standards. However,
volatile organic compounds were not. The degrees of overall IEQ satisfaction and the
proportion of occupants voted were both greater in the green buildings compared to the
conventional buildings. The research has found statistically significant difference between
the mean score of satisfaction in the green buildings and of conventional buildings [19].
19

Morhayim conducted POE in office and laboratory university buildings in Israel
with the IEQ measurements and survey. He finds it problematic as Sick Building Syndrome
(SBS) was identified and has been investigated for 30 years, yet buildings recently
designed and constructed seem to disregard the phenomenon. A POE study done in a
building in Israel which combined functions of office and laboratory. The research used a
review of building plans to shos weakness in decision-making and design caused by lack
of POE. The results of measurements, surveys, and interviews are the outcome of an
educational exercise taken with a group of graduate students [20].

Meir suggested a comprehensive methodology for post occupancy evaluation in a
hot, dry climate. The methods used in POE include plan analysis and monitoring of Indoor
Air Quality and thermal performance. The surveys are conducted with walk-through
observations and user satisfaction questionnaires. The research reviews a rare case of
cooperation among the different stakeholders in a complex of different accommodation
and facilities, located in the Negev Desert Highlands in Israel. The first part of the research
conducts a lateral monitoring aimed at gaining a better understanding of the overall
performance of the project. A subsequent series of user surveys, observations, and
questionnaires are taken focusing on potential discrepancies between the need to ensure
thermal comfort and privacy issues. The second set of measurements was taken
subsequently focusing on the in-depth monitoring of one family unit representing the most
complex type of units and covering an eight-month period [21].

20

Kansara introduced the method of Post Occupancy Monitoring as an evolution from
Post Occupancy Evaluation of the built environment. The method is based on the
qualitative and quantitative aspirations of the stakeholders to apply it closing the feedback
system of the built environment. The research discusses previous attempts such as Plan of
Work for architects by the Royal Institute of British Architects (RIBA) as well as the soft
landings approach for handing over the building. The research points out the gap in the
literature in terms of continuously analyzing a building once it is occupied. Suitable
incentives are recommended to each of the stakeholders to participate in the process and
learn from past initiatives. The research suggests the need for energy monitoring, thermal
comfort analysis and documenting user satisfaction as a basis for all existing building rating
systems [22].

Choi suggests a basis for future IEQ standards and guidelines by conducting POE
on office buildings with on-site measurement, TABS, and questionnaire data. Based on the
need of indoor environmental conditions to support computer-intensive activities as well
as paper-based tasks, the research argues that the current standards and guidelines for
indoor environments were predominantly developed without consideration for these
modern office variables. The limitation may lead to an occupant’s dissatisfaction with the
indoor environmental quality (IEQ) as well as unnecessary energy use. The research
performed a broad range of post-occupancy evaluation studies in Federal office buildings
across the U.S. over seven years. Spot and continuous measurements are conducted to
consider IEQ condition. Occupants’ satisfaction with environmental attributes was
surveyed simultaneously. Statistical analyses have linked characteristics and
21

environmental qualities to occupants’ satisfaction. The results challenged the validity of
current IEQ standards and guidelines. Specific recommendations for improving current
standards and guidelines are outlined to help enhance environmental conditions in
workplaces for future design projects [23].

As seen from the collected background research, comprehensive POE has a various
model in its form. However, some similarities are found that combination of
multidisciplinary data increases the objectivity of the method. The methods and tools
employed are both quantitative and qualitative, and may be classified into three rough
categories by the information analyzed and assessed. (1) Measurements, monitoring,
sampling; (2) surveys, questionnaire, cohort studies, observations, task performance test;
and (3) document analysis, on-site observation.

2.1.2 Conventional Research Goals of POE
Common purposes of POE are (1) to highlight any immediate teething problems
that can be addressed and solved, (2) to identify any gaps in communication and
understanding that impact on the building operation, (3) to provide lessons that can be used
to improve design and procurement on future projects, and (4) to act as a benchmarking
aid to compare across projects and over time [13]. For instance, Meir sets a goal of post-
occupancy evaluation as developing a platform to allow the systematic study of buildings
when it is occupied. The POE provide lessons that will improve future design [21]. Some
of the research projects are observed to identify conventional research goals of POE.

22

Zagreus conducted a web ‐based indoor environmental quality survey as he defines
building occupants as a rich source of information about indoor environmental quality
which effect on comfort and productivity. The research has developed a web-based survey
and online reporting tools to collect, process and present the information. The survey was
conducted in more than 70 buildings to create a database. Survey questions ask occupant
satisfaction on IEQ areas such as lighting, thermal comfort, indoor air quality, acoustics,
office layout, furnishings, and building maintenance. The research has focused on three
case studies that explained different applications of the survey. The cases include pre and
post analysis of occupants in a new building, and a case which conducted physical
measurements in addition to survey to find out the impact of environmental factors on
occupants' comfort level and productivity. The research is an example of survey used to
establish how new buildings are meeting an occupant’s design objectives [24].

Frontczak conducted a questionnaire survey in Danish homes to investigate the
factors that influence occupants’ comfort. The questionnaire contained questions on
occupants’ behavior, their knowledge on controlling the indoor environment, and the ways
to achieve a comfort level. A total of 2499 survey questionnaires were sent to occupants of
the most common types of housing in Denmark. With a response rate of 26%, the results
show that the main indoor environmental parameters such as visual, acoustic and thermal
conditions, and air quality are considered by occupants as the most important parameters
for comfort level. Individual control of the indoor environment was indicated to be highly
preferred by the respondents, and only in the case of temperature did they accept both
manual and automatic control. The respondents stated that they were confident about how
23

the systems for controlling indoor environmental quality in their homes should be used.
54% of them reported having had at least one problem related to the indoor environment
at home. A majority of those respondents did not try to search for information on how to
solve the problem. The result was inferred to suggest that there is a need for increasing
people’s awareness regarding the consequences of a poor indoor environment on their
health and for improving people’s knowledge on how to ensure a healthy indoor climate
[25].

Langston researched into office design and its effect on employee satisfaction and
performance. However, most studies have tended to concentrate on the impact of the built
environment on human performance, ignoring the actual needs of employees working in
different organizational settings. The research aimed to investigate the nature and extent of
occupant satisfaction with the built environment in different organizational settings in
Australia for a range of climates. A survey was conducted in Australia between 2004 and
2005 in 41 buildings (6 government buildings, 14 educational buildings, and 21
commercial buildings). The research uses Kruskal-Wallis H test to explore whether there
are differences in the mean ranking of office environment satisfaction among three
different organizational settings. Meanwhile, Mann-Whitney U test was employed to test
whether there are differences in the average ranking of office environment satisfaction
between any two groups [29].

As seen from selected research projects, POE is conducted to provide feedback on
how successful the workplace is supporting the occupying organization and individual’s
24

requirements. The studies suggest that POE can be used to assess and also can be utilized
as part of the Evidence-based design process. To widen its usability, POE may involve
more objective measures such as environmental monitoring, space measurement and cost
analysis, in addition to usual feedback from the building occupants through questionnaires,
interviews, and workshops.

2.1.3 Impact of POE on Occupants in Workspace Environment
To point out the impact of POE on occupants’ workstation environment, POE
researches that were conducted especially in office or relevant setting were focused on.
From the description and results of selected projects, the necessity of POE is found in
workspace environment as space is occupied and used by the owner during most time of
the day.

Li argued in the research that people and building performance are intimately linked.
The research focuses on the issue of occupant behavior; principally, its impact, and the
influence of building performance on occupants in an office environment. It explains how
energy is consumed in buildings and identifies the range of occupant-interactive
opportunity. The issue of post-occupancy evaluation (POE) is covered, exposing the
concept of the energy performance gap and why discrepancies occur. Then, the research
focuses on building performance, particularly indoor environment of the office
environment, impact on work productivity, and how it is measured. It also discusses
occupant adaptation in achieving thermal comfort, in addition to, the role of energy
25

management systems, smart sensor networks, and data mining with occupant behavior as
the backdrop [26].

Abbaszadeh researched on the impact of indoor environmental quality on occupants’
satisfaction in green buildings. The research points that occupants in green buildings were
more satisfied with the thermal comfort and air quality in their workspace. However, the
average satisfaction scores in green building for lighting and acoustic quality were
comparable to the non-green buildings. Comparing complaints profiles of those dissatisfied
with lighting and acoustic quality, a higher percentage of occupants were dissatisfied with
light levels and sound privacy in green buildings. The results suggest a need for
improvements in controllability of lighting, and innovative strategies to accommodate
sound privacy needs in open plan or cubicle office layouts in both comparison groups [27].

Pfafferott conducted research on comparing low-energy office buildings in summer
using different thermal comfort criteria. Monitoring was carried out over 2 to 3 years in 12
low-energy office buildings which are located in three different summer climate zones in
Germany. The weather at the building site and the room temperatures in several office
rooms were monitored. The raw data were processed for data evaluation using a
sophisticated method to remove errors and outliers and to identify the time of occupancy.
The comfort of all office rooms in each building is evaluated separately. The result of
research in these 12 low-energy office buildings indicates that building with only natural
heat sinks for cooling provide good thermal comfort during summer periods in Germany.
However, long heat waves generated in such as during the extreme European summer of
26

2003 overstrain passively cooled buildings with air-driven cooling concepts in terms of
thermal comfort [28].

Langston identified significant differences in aspects of air, temperature, space
suitability, flexibility, usability, and controllability in different building types. Occupants
in commercial buildings were more satisfied with their physical work environment than
occupants in other building types. Occupants in educational settings showed the highest
satisfaction with most variables in the workspace design and management category.
Occupants of Government buildings showed a lower level of satisfaction with their
physical work environment and workspace design and management. In addition, occupants
of the government and educational buildings showed more similarity with each other, while
the occupants of commercial buildings displayed significant difference [29].

Wagner focused on the issues of comfort and workspace quality as buildings have
gained much importance with the European "Energy Performance of Buildings Directive"
of 2001. The 4-week field study on thermal comfort with 50 subjects in a naturally
ventilated office building in Karlsruhe, Germany in summer, shows that thermal sensation
votes do not correspond to calculated predicted mean votes. However, a very good
agreement can be seen with adaptive comfort models. A survey on workplace occupant
satisfaction in 16 office buildings in Germany revealed that the occupants' control of the
indoor climate. Moreover, the perceived effect of their interference strongly influences
their satisfaction with thermal indoor conditions [30].
27

2.2 IEQ Indicators
Lighting, thermal, indoor air and acoustic quality are commonly considered as an
IEQ indicator. The indicators can be represented by performance metrics relevant to their
potential impacts on occupant satisfaction and acceptance of indoor environments.

2.2.1 Visual Quality
According to the definition from Autodesk Sustainability Workshop, a measure of
the total perceived power of light is called luminous flux or luminous power. It is the
amount of light emitted from a particular source of all directions. It is measured in lumens
to compare how bright a light source is. Humans can perceive light within the visible
spectrum of wavelengths between about 390 nm and 700 nm. Some wavelengths of light
are perceived more strongly, and luminous flux is scaled to reflect this. Radiant flux is a
measure that quantifies the total power of the electromagnetic radiation from a source
which is measured in watts. The amount of light from the source to certain directions is
called luminous intensity and measured in candelas. One candela is same as light from the
single candle in all directions [31].

The amount of light that falls on a surface is called illuminance and is measured in
lux (lumen/m
2
) or foot-candles (lumen/ft2). One foot-candle equals to 10.764 lux.
Illuminance is the measurement used the most for optimizing visual comfort. It is because
current building regulations and standards use illuminance to specify the minimum light
level for specific tasks and spaces. This value does not change the material properties of
28

the surface being illuminated. However, it does rely on the color and reflectance of the
surfaces that surround it [31].

Lighting levels in offices need to be a quality that is easy to see and be able to
perform office tasks safely without eye strain. During typical working hours, lighting in
the office environment tends to rely on a combination of electric lighting and daylighting
from windows. People are known to prefer working by daylight and to enjoy an outdoor
view. Also, this mixture of lighting enables a degree of flexibility which is a useful
outcome. Windows can help with maintaining a strategic distance from or diminishing eye
strain by permitting a person to concentrate on inaccessible questions as opposed to a
delayed review of close protests, for example, PC screens. At the same time, the use of
windows needs to be carefully balanced with any unfriendly thermal effects or undesirable
lighting impacts, such as glare [14].

According to CIBSE, glare can be defined as a condition of vision in which
discomfort or a disturbance in the ability to see details or objects, caused by an improper
distribution or range of luminance, or to extreme contrasts. Although there are many
indices for visual discomfort due to glare in the literature, the most extended ones are
Unified Glare Rating (UGR) index and the Daylighting Glare Index (DGI). UGR is used
to measure the degree of glare in this research. The UGR index expresses the discomfort
glare occasioned by the presence of bright light sources, luminaires or windows. UGR
index values are within the range of 13-30, and moreover, the lower UGR values, the less
visual discomfort condition [32].
29

2.2.2 Thermal Quality
Human thermal comfort is a combination of how we feel (subjective sensation) and
several objective interactions with the environment such as heat transfer rates. In person-
related thermal comfort, it is related to deep body temperature which always remains close
to 37 º C. The Metabolic dissipation rate is the heat that must be evacuated by unit body
mass of 0.5 W/kg to 5 W/kg, depending on activity. Around 100 W is generated for an
adult in office work. Skin temperature is usually below 33 º C, allowing the heat evacuation,
but it depends on external conditions, clothing, and activity levels. Also, age and risk
groups such as babies, elders, ill people have a different rate. Personal habits such as
clothing difference among seasons and gender, previous accommodation, personal
preferences on comfortable temperature, and actual mood of the person may influence as
comfort is a psychological problem [33].

In an environment-related aspect, air temperature or water temperature, radiant
background temperature of walls, relative humidity, and wind speed are to be considered.
According to Sharma, not only average values but their gradients and transients matter too.
Non-thermal environmental variables like ambient light and noise may affect the thermal
sensation. The thermal comfort parameter that is the most difficult to measure is
background radiant temperature. Because, it varies by direct solar irradiance, wall solar
reflectance, sky temperature, wall temperature, and all the geometric view factors [33]
This is sometimes referred to as Mean Radiant Temperature, or MRT.

30

To measure the level of human thermal comfort, researchers tend to collect
individual responses within a few degrees of comfort. The most common scale used for
thermal comfort is the 7-point scale. Each measure of the scale represents: Cold (95% or
more of people complain of being cold), Cool (75% of individuals complain of being cold),
Slightly cool (25% of individuals complain of being cold), Comfortable (less than 5% of
individuals complain of being cool or warm), Slightly warm (25% of individuals being
hot), Warm (75% of individuals complain of being hot) Hot (more than 95% of individuals
complain of being hot). The common goal of thermal comfort analysis is finding an
appropriate function of the physical parameters [34].

The Predicted Mean Vote (PMV) is another system used to scale thermal comfort.
The scale runs from Cold (-3) to Hot (+3). Before being adopted as an ISO standard, it was
originally developed by Fanger. The original data was collected by subjecting a large
number of individuals to various conditions within a climate chamber and having them
select a position on the scale the best indicated their comfort sensation. A mathematical
model of the correlation between other environmental and physiological factors considered
are extracted from the data. The result produces the following sensation scale. According
to ASHRAE 55, recommended PMV range for thermal comfort in interior space is between
-0.5 and +0.5. Predicted Percentage of Dissatisfied (PPD) calculates the possible rate of
occupants that will be dissatisfied with the thermal conditions. The maximum number
possible of people dissatisfied with conditions is 100%. As it is impossible to satisfy all of
the people at once, the recommended acceptable PPD range for thermal comfort is less than
10% persons dissatisfied for an interior space in ASHRAE 55 [35].
31

2.2.3 Indoor Air Quality Indicators
Improper indoor air quality (IAQ) can be a significant health, environment, and
economic problem. IAQ measures must determine how well indoor air satisfies thermal
and respiratory requirements of occupants to prevents unhealthy accumulation of pollutants
and acceptance rate for a sense of comfort level. Researches have established the
occurrence of various building-related illnesses with identifiable and diverse causes. A
subset of these diseases also known as sick building syndrome (SBS) mainly includes
subjective symptoms such as mild irritation of nose, eyes and throat, headaches, and
lethargy. The symptoms are believed to arise from multiple sources which are associated
mainly with air-conditioning of office buildings [12].

A number of organizations in America have established indoor air quality standards,
such as the American Society of Heating, Refrigerating and Air- Conditioning Engineers
Standard (ASHRAE 189.1), the South Coast Air Quality Management District's
(SCAQMD) Rules, and the Sheet Metal and Air Conditioning Contractors' National
Association's (SMACNA) IAQ Guidelines [36].

One of the indicators of indoor air quality is a level of carbon dioxide (CO2). As
people consume oxygen and breathe out CO2, the air becomes stale. If the CO2
concentration is too high, occupants become tired and have difficulty concentrating.
Bringing in fresh air from outside with oxygen keeps people energetic and happier. To
prove its significance, the LEED rating system has credits for meeting ASHRAE standards
for fresh air. Mold and mildew are another common indoor air pollutant, especially in
32

humid climates. The simplest way to avoid them in the building is to prevent condensation.
To prevent condensation in the building, the best solution is to make the building breathe
so the moisture can escape [36].

2.2.4 Acoustic Quality Indicators
Sound is created by waves of compressed air that we perceive with our ear. Many
objective properties contribute to how we measure sound based on its wave nature. For
instance, the density of air impacts how fast a sound moves across space. The distance
between waves, both regarding a linear distance as well as time, gives a sound its pitch.
Sound also embodies energy, with the magnitude of air compression leading to an
experience of volume. The perception of sound by the human ear is determined by
thresholds of both frequency and magnitude. The range of pitches humans can hear lies
between 20 and 20000 Hz. As the body ages, however, the range of frequencies our ears
can sense begins to diminish, and the ear requires more energy (measured in higher decibels)
to hear each pitch. Thresholds of sound pressure also exist for the human ear. Measured in
decibels, we can hear sounds as quiet as 0 dB, and 130 dB stands at the commencement of
pain, though this can depend on the pitch. The combination gives us the phenomenological
experience of loudness [37]. In 1981, OSHA suggested a requirement to protect all
employees in general industry where employees are exposed to an average noise level of
85 dB or higher for more than 8 hours of work shift [38].
.

33

Good office acoustics is a significant contributor to work performance and well-
being in the workplace. The ability to find quiet times and places is essential to support
complex knowledge work while the ability to have planned or spontaneous interactions
without disturbing others is necessary for teamwork and relationship development. Having
speech privacy is needed for confidential interactions and work processes. “Acoustical
comfort” is achieved when the workplace provides appropriate acoustical support for
interaction, confidentiality, and concentrative work [39].

34

Chapter 3: Research Objectives and Methodologies
To conduct an extended POE, three types of dataset are measured and collected
from the selected workstations. The first set of data is the composition of building attributes
that are observed and recorded to recognize the spatial factors of the workstation. The
second dataset is collected through an actual measure of IEQ features (luminance,
illuminance, temperature, noise level, etc.) at the workstation. Lastly, a satisfaction survey
form is distributed to the occupants of each workstation to take into consideration the
human factor. With collected series of datasets mentioned above, statistical analysis
software is used to investigate the correlation between each factor and its impact on one
another, as illustrated in Figure 1.

Figure 1: Methodologies of the Research
35

3.1 Research Objectives
The main objectives are (1) to compare actual IEQ measurement and occupants’
environmental satisfaction at workstations of office environment, (2) to understand the
relationship between occupants’ environmental satisfaction and their human factors, (3) to
investigate the effects of building attributes on occupants’ environmental satisfaction and
IEQ performance of the building, (4) to compare and prioritize the specified factors by its
intensity of impact on occupants’ environmental satisfaction, and (5) to establish design
guidelines based on the significant findings from analysis of the database.

This approach can identify critical factors in the physical environment that impact
building occupants’ comfort and satisfaction. The outcome can contribute to (1) defining
correlations between occupant perception and measured data, (2) prioritizing the impact
factor to maximize occupants’ environmental satisfaction within limited project resources,
and (3) developing POE-based design guidelines in an office environment in Southern
California.

The main streamline of the research is to collect and analyze various data from
office facilities in Southern California regarding IEQ and human comfort. It began with
selecting office environments and buildings located in Southern California for a field study.
Then, it was carried out by comparing data collected from the facilities, using mixed
methods of qualitative and quantitative research methods. Specific methodologies for each
of research objectives are demonstrated below.

36

3.2 Methodologies
3.2.1 Building Attributes Observation
Physical attributes of the building and workstation are carefully recorded and
observed. The recording and observation are to identify the key attributes of the building
and workspace that might impact measured environmental conditions or occupants’
satisfaction in the indoor environment. Observed attributes include floor number (on which
floor is the workstation located), position (whether the workstation is located at the center
or perimeter of the building), area (net area, gross area), WWR (window-to-wall ratio),
building orientation, and building material. The data represents the spatial factor of the
workstation that could influence visual, thermal, and acoustic satisfaction of the occupant
and IEQ measurements.

3.2.2 IEQ Measurement
Actual IEQ measurements such as thermal, air quality, lighting, and acoustic
measurements are collected from each workstation. To collect more accurate IEQ data, two
different sets of data are measured, based on the duration of data measurement. Continuous
IEQ measurement is conducted with ‘IEQ cart’, developed by one of the academic advisors,
Dr. Joon-Ho Choi of University of Southern California. This cart is designed to collect
temperature data at 1.1m, 0.6m, and 0.1m height from the floor. Also, relative humidity
(RH), oxocarbon (CO, CO2), particulate matters (PM) and total volatile organic
compounds (TVOC) at 1.1m height, defined as the “breathing zone” by ASHRAE-129 [40],
can be measured. All the measurements from the IEQ cart are automatically transferred to
the database through data logger. The cart was placed at the position of the occupant’s chair
37

at each workstation for 15 minutes, with 15-second sensing intervals for recording. In
addition to continuous data from the IEQ cart, instant IEQ data is collected through data
loggers installed in corresponding spots of selected workspace. Hand-held sensors were
used to measure light levels on the monitor, work surface, and reading zone, as well as
radiant temperature, air velocity, and noise level of the workstation. List of measured IEQ
variables is identified in Table 1.

Table 1: Type of Measured IEQ Data

IEQ Data
Continuous Data Instant Data
Visual Quality
Illuminance,
Average Luminance Level
Illuminance
(monitor, workspace, reading zone)
Luminance, UGR
Thermal Quality
Temperature (0.1m / 0.6m / 1.1m),
Relative Humidity
Radiant Temperature
Indoor Air Quality CO, CO 2, PM, TVOC Air Velocity
Acoustic Quality Noise Level Noise Level

3.2.3 Occupants’ Satisfaction Survey
For the occupants’ satisfaction survey, a modified version of COPE environmental
satisfaction questionnaire was used. The original version of the survey was developed by
the National Research Council of Canada to support the Cost-effective Open-Plan
Environments (COPE) project in 2003 [41]. The survey form used in this research consists
of 2-page, 29-questionaire, which asks about the satisfaction level of the occupant on
current indoor environmental condition. 29 survey questions used in the survey about its
IEQ criteria is as follow in Table 2.
38

Table 2 Occupants’ Satisfaction Survey Questions
Category Question
Q01
Occupant
Information
How many years have you been working in this building?
Q02
In a week, how many numbers of hours do you spend at your
workstation?
Q03 What is your age?
Q04 What is your gender?
Q05 What is your job category?
Q06
Workspace
Condition
How satisfied are you with your job/institution?
Q07
How satisfied are you with the size of your personal workstation to
accommodate your work, materials, and visitors?
Q08 How satisfied are you with the level of privacy in your workspace?
Q09
How satisfied are you with your ability to alter physical conditions in
your work area? (e.g. operable windows, blinds)
Q10
How satisfied are you with your access to a view of outside from where
you sit?
Q11
How satisfied are you with the distance between you and other people
you work with?
Q12
How satisfied are you with the degree of enclosure of your work area
by walls, screens, or furniture?
Q13
Thermal
Quality
How satisfied are you with the current temperature at your workspace?
Q14 How satisfied are you with the thermostats? (e.g. operability)
Q15 IAQ
How satisfied are you with the current air quality in your workspace?
(i.e. stuffy/stale air, cleanliness, odors)
39

Q16
Acoustic
Quality
How satisfied are you with the amount of noise from other people's
conversations while you are at your workstation?
Q17
How satisfied are you with the amount of background noise you hear at
your workstation? (i.e. not speech, noise from mechanical systems)
Q18
Visual
Quality
How satisfied are you with the light for doing computer work?
Q19
How satisfied are you with the amount of reflected light or glare on the
computer screen?
Q20
How satisfied are you with the amount of direct glare from light
fixtures? (ex: unshielded light bulb visible from your position)
Q21 How satisfied are you with the amount of direct glare from daylight?
Q22
How satisfied are you with the quality of the lighting in your work
area?
Q23
Spatial
Factor
How satisfied are you with the general building and office layout?
(workstation layout)
Q24
How satisfied are you with the colors and textures of the flooring,
furniture, and surface finishes?
Q25
Satisfaction
How would you describe the air movement in your workspace?
Q26 How would you describe the ambient temperature in your workspace?
Q27 How would you describe the lighting conditions in your workspace?
Q28
How would you describe the noise level/acoustic conditions of your
workspace?
Q29
Overall, how satisfied are you with the indoor environment of your
workspace?

A 7-point scale has been used in multiple subjective survey questionnaires as a
typical tool due to its technical features that help prevent complex interpretations of the
answer [42]. Each scale represents different satisfaction level for each IEQ component: -
40

3–very dissatisfied, -2–dissatisfied, -1–slightly dissatisfied, 0–neutral, +1–slightly satisfied,
+2–satisfied, +3–very satisfied. However, the data are transformed into the lower level of
point scale (3-point scale and 2-point scale) throughout the analyses to illustrate better the
tendency and its correlation. A 3-point scale was used to combine the occupants of non-
neutral groups (negative, neutral, and positive), and a 2-point scale was used to simplify
the responses into negative and positive by considering neutral status as a positive response.
Scale points of each scale are described in Table 3.

Table 3 Sensational Scale for Occupants’ Environmental Satisfaction Survey
Sensation unpleasant/ dark/ cool/ quiet Neutral pleasant/ bright/ warm/ noisy
7-point Scale -3 -2 -1 0 +1 +2 +3
3-point Scale Negative Neutral Positive
2-point Scale Negative Positive

3.2.4 Statistical Analysis
With the collected dataset, cross-sectional analyses are performed based on the
collected data with a strategic grouping of data sets by human factor and spatial factors
such as gender, age group, and workstation location. The analyses on lighting quality are
based on IESNA environmental quality standards and guidelines [43] while ASHRAE-55
[44] and 62.1 [45] is used for thermal and air quality assessment. The total dataset of
attribute record, IEQ measure, and satisfaction survey is 411, which meaning enough
sample number to achieve normality for statistical analyses. Correlation analysis, two-
sample t-test, one-way analysis of variance (ANOVA), and an ordinal logistic regression
analysis are used in this research using Minitab computer software program. All analyses
41

were primarily performed at 95% statistical significance level to examine the relationship
between collected dataset.

(1) Stepwise Regression
Stepwise regression is a semi-automated mechanism of building a statistical model
by adding or removing variables based solely on the t-statistics of their estimated
coefficients. Properly used, the stepwise regression puts more power and information at
the ordinary multiple regression options, and it is especially useful for sifting through a
significant number of potential independent variables and fine-tuning a model [46]. In
stepwise regression, R-squared is a statistical measure representing how close the data are
to the fitted regression line. It is known as the coefficient of determination. R-squared is
always between 0 and 100%. 0% means that none of the variability of the response data is
around its mean while 100% indicates it is around its mean. However, R-squared does not
explain whether a regression model is adequate [47].

(2) Correlation Analysis
Correlation Analysis estimates a sample correlation coefficient, more specifically
the Pearson Product Moment correlation coefficient. The sample correlation coefficiency
denoted r, ranges between -1 and +1 and quantifies the direction and strength of the linear
association between the two variables. The correlation between two variables can be either
positive or negative [48]. However, it is important that correlation does not imply causation.
Only properly controlled experiments allow determining if a relationship is causal. A well-
known example of this concept states that ice cream sales are positively correlated with
42

shark attacks on swimmers. It would be a false statement to conclude that ice cream
consumption is causing sharks to attack [49].

(3) Two-sample T-test
The two-sample t-test is a hypothetical examination for solving questions about the
mean where the data are collected from two random samples of independent observations,
each from an underlying normal distribution. It is used to determine if two population
means are equal. A common application is to examine if a new treatment or process is
superior to a current process or treatment. Two-sample t calculates a confidence interval
and does a hypothesis test of the difference between two population means when standard
deviations are unknown, and samples are drawn independently from each other. This
procedure is based on the t-distribution. It works best for small samples if the data were
drawn from distributions that are normal or close to normal. Confidence in the results
increase as the sample sizes increase [50].

(4) Analysis of Variance
Analysis of variance (ANOVA) examines the hypothesis that assumes the means
of two or more populations are equal. ANOVAs assess the significance of factors by
comparing the response variable means at the different factor levels. The null hypothesis
states that all factor level means are equal while the alternative hypothesis states that at
least one is different. To perform ANOVA, the continuous response variable is required as
well as at least one categorical factor with two or more levels. ANOVA require data from
normally distributed populations with equal variances in factor levels. However, ANOVA
43

procedures work well even if the normality assumption is violated, unless one or more of
the distributions are highly skewed or if the variances are entirely different.
Transformations of the original data may correct these violations [51].

The reason not to compare groups with multiple t-tests instead of ANOVA is
because of a chance of making Type 1 error in a t-test. Type 1 error in statistical hypothesis
testing is the incorrect rejection of a true null hypothesis (a false positive). This error is
usually 5%. Running two t-tests on the same data will increase the chance of "making a
mistake" to 10%. The formula for determining the error rate for multiple t-tests is not as
simple as multiplying 5% by the number of tests. However, for a few multiple comparisons,
the results are very similar. As such, three t-tests would be 15% which is close to the actual
value of 14.3%. These are unacceptable errors. An ANOVA controls for these errors so
that the Type 1 error remains at 5% and result can be more confident that any significant
result found is not just down to chance [52].

(5) Decision Tree (J48)
A Decision Tree (J48) is a data mining model that decides the dependent variable
of a new sample based on various attribute values of the data. Data mining is the
computational process of pattern discovery in large data sets. The overall goal of the data
mining is to capture information from data and transform it into a simple format. Data
mining computer software Weka developed by machine learning group at the University
of Waikato in New Zealand [53] is used for analysis. The software automatically analyses
a large body of data and decide what information is most relevant.
44

Chapter 4: Data Description
In this chapter, the distinctive features of collected data and its distribution by
different factors are explained and illustrated. Three types of data are collected for
comprehensive POE: IEQ measurement, building attributes, and occupants’ satisfaction.
The data was collected from more than 400 workstations within office environment of 14
different modern buildings in Southern California.

4.1 Data Distribution
4.1.1 Data Distribution by Human Factors
Each dataset was collected from over 400 workstations within these buildings. To
clarify the distribution of the collected data, the entire dataset is categorized by some of the
key factors of the data. As illustrated in Table 4, there were 188 female occupants and 223
male occupants ranging in age of 18 to 69 years old. A total of 196 subjects were in the
junior group (18 to 29 years old), 165 in the mid-aged group (30 to 49 years old), and 50
in the senior group (50 to 69 years old).

Table 4 Data Distribution by Human Factors
Age Group Age Female Male Total
Junior 18-29 96 100 196
Subtotal 96 100 196
Mid-age 30-39 50 55 105
40-49 24 36 60
Subtotal 74 91 165
Senior 50-59 12 21 33
60-69 6 11 17
Subtotal 18 32 50
Total 188 223 411
45

4.1.2 Data Distribution by Spatial Factors
The data is collected from workstations with different spatial characteristics such
as building type and workstation location. There are two separate groups of buildings
where the data was collected. One group is commercial buildings located in Los Angeles.
The other group is university buildings from the University of Southern California. Data
measured from commercial buildings count 196 while data from university buildings are
215. From each workstation, IEQ data and building attribute information were measured
and collected while the occupant took the satisfaction survey. The datasets also can be
divided into two groups by workstation location. Defining perimeter workstation as the
workstation within 15ft from the exterior wall, there are 190 datasets from perimeter
workstation and 173 from the center workstation.

Table 5 Data Distribution by Spatial Factors
Location Commercial Buildings University Buildings Total
Center 69 104 173
Perimeter 87 103 190
Unidentified 40 8 48
Total 196 215 41 1

Distribution of the data gives evidence of its different features of each data, and
also directs analyses on data to be made by the different group. Besides the distribution of
the data, it is also helpful to see the entire data as a whole. It explains the overall condition
of measured buildings and satisfaction level of occupants. Outlook of measured and
surveyed data from IEQ measurement and occupant satisfaction survey are described in
next section.
46

4.2 Measured and Surveyed Data
4.2.1 IEQ Measurement
Various components of indoor environmental quality (IEQ) are measured to
diagnose the workstation status and to compare with satisfaction survey in statistical
analysis. Distribution of entire IEQ data shows the overall environmental condition of the
buildings participated in the study. The histogram is used to illustrate the frequency of data
by its measure. Box plot graph is used to identify the range of majority data distribution.
Measured IEQ data represents lighting, thermal, IAQ, and acoustic quality of the
workstation.

Figure 2 Distribution of Work Surface Illuminance

As a lighting quality of the workstation, distribution of measured work surface
illuminance data is illustrated in Figure 2. The majority of data are spread between 200 and
600 lux while mean value is 448.5, and the standard deviation is 325.1. According to
47

ANSI/IES RP-1-12 American National Standard Practice for Office Lighting, primary
work surface illuminance level is recommended between 200 and 500 lux. It is observed
that only 49% of the measured workstations are within recommended work surface
illuminance criteria. 21% of the data were measured under 200 lux while 30% is found at
over 500 lux.

Figure 3 Distribution of Screen Illuminance

Distribution of measured screen illuminance data is illustrated in Figure 3. The
majority of data are spread between 100 and 300 lux while mean value is 238 lux, and the
standard deviation is 184. According to ANSI/IES RP-1-12 American National Standard
Practice for Office Lighting, primary screen illuminance level is recommended between 50
and 100 lux. It is observed that only 9% of the measured workstations are within
recommended screen illuminance criteria. 2% of the data were measured under 50 lux
while 89% is found at over 100 lux.
48

Figure 4 Distribution of Reading Zone Illuminance

The distribution of measured reading zone illuminance data is illustrated in Figure
4. The majority of data are spread between 200 and 400 lux while mean value is 349.1 lux,
and the standard deviation is 241. According to ANSI/IES RP-1-12 American National
Standard Practice for Office Lighting, primary screen illuminance level is recommended
between 200 and 500 lux. It is observed that 57% of the measured workstations are within
recommended reading zone illuminance criteria. 27% of the data were measured under
200 lux while 16% is found at over 500 lux.

Distribution of average luminance level is illustrated in Figure 5. The majority of
data are spread between 20 and 70 cd/m2 while the mean value is 74.8 cd/m2 and the
standard deviation is 111. According to IESNA, the indoor office luminance inside the
visual field can be up to 850 cd/m2. All the measured average luminance level from the
workstation are observed to be within its recommended criteria.
49

Figure 5 Distribution of Average Luminance

Figure 6 Distribution of UGR

The distribution of the unified glare rating (UGR) is illustrated in Figure 6. The
majority of data are spread between 15 and 20 while mean value is 18, and the standard
600 500 400 300 200 100 0 -100
80
70
60
50
40
30
20
10
0
Mean 74.80
StDev 111.0
N 176
Avg Luminance cd/m2
Frequency
Histogram of Avg Luminance cd/m2
50

deviation is 6.16. According to International Commission on Illuminance (CIE), the UGR
should be lesser than or equal to 19. Setting up the recommended UGR to 13 through 19,
it is observed that 54% of the measured workstations are within recommended UGR criteria.
15% of the data were measured under UGR 13 while 31% is found at over UGR 19.

Figure 7 Distribution of Ambient Temperature

As a thermal and air quality of the workstation, distribution of ambient temperature
is illustrated in Figure 7. The majority of data are spread between 22 and 25°C while mean
value is 23.3°C, and the standard deviation is 1.34. According to ASHRAE 55 standards,
the ambient temperature of the indoor space should be between 20 and 25.6°C. It is
observed that 92% of the measured workstations are within recommended temperature.
Only 8% of the data were measured from above 25.6°C.

51

Figure 8 Distribution of Air Velocity at 1.1m Height

Distribution of air velocity at 1.1m height is illustrated in Figure 8. The majority of
data are spread between 0.0 and 0.2 m/s while mean value is 0.15 m/s and the standard
deviation is 0.18. According to ASHRAE standards, indoor air speed should be 0.2 m/s or
less. It is observed that 80% of the measured workstations are within recommended air
velocity. However, 20% of the data were measured at air velocity higher than 0.2 m/s.

Similarly, the distribution of air velocity at 0.1m height is illustrated in Figure 9.
The majority of data are spread between 0.0 and 0.1 m/s while mean value is 0.07 m/s and
the standard deviation is 0.08. Following the same ASHRAE standards to the air velocity
at 1.1 m/s, 94% of the measured workstations are observed within recommended air
velocity. Only 6% are found to be not meeting the standards.

52

Figure 9 Distribution of Air Velocity at 0.1m Height

Figure 10 Distribution of CO2 Level

Distribution of measured CO2 level from each workstation is illustrated in Figure
10. The majority of data are spread between 600 and 800 ppm while the mean value is 733
ppm, and the standard deviation is 170. According to the EPA (United States
53

Environmental Protection Agency), the permissible indoor CO2 level should be less than
800 ppm. Meanwhile, ASHRAE 62 regulates the permissible indoor CO2 level as 1000
ppm or 700 ppm above outdoor CO2 levels. For ASHRAE standards, a measure of outdoor
CO2 level is necessary. Following the EPA, 75% of the measured workstations are
observed within recommended CO2 level while 25% are found from out of EPA
recommendation.

Figure 11 Distribution of Acoustic Decibel

As an acoustic quality of the workstation, distribution of noise level is illustrated in
Figure 11. The majority of data are spread between 50 and 70 dB while mean value is 63
dB, and the standard deviation is 13.5. According to ASHRAE standards, acoustic decibel
should be less than 40 dB in open plan office space and 35 dB in a private office. Following
the recommendation of open space office, only 2% of the measured offices are within
54

ASHRAE acoustic standards. 98% of the data were found to be exceeding the
recommended acoustic decibel

4.2.2 Satisfaction Survey
Each of the questions from the occupant’s satisfaction survey has its corresponding
IEQ and spatial criteria. Single survey data only represents the perception of the individual.
However, combined data helps understand and figure the tendency within the data through
mean vote value. By counting a number of occupants who answered on each question, the
distribution of the answers can be shown with a simple bar chart. In addition, box plot
graph is generated in order to point out the mean, median value, and the distribution of data
by its preference. The series of graphs drawn for each survey question can be found at the
end of this document as an appendix.

Instead of listing all the single graphs by each question, the rose chart is used in this
section to effectively illustrate and compare the data by the question. The seven circles in
the chart represent the seven scale point used in the survey (-3: very dissatisfied, -2:
dissatisfied, -1: slightly dissatisfied, 0: neutral, +1: slightly satisfied, +2: satisfied, +3: very
satisfied). Each linear line starting from the center of the circle represents each question
made in the survey. The mean votes made by the occupants are marked with red dots on
the graph, where the circles and the linear lines intersect. The authentic form is generated
through connecting all the single red dots from each question, and this form can be
compared with the yellow circle, representing the overall environmental satisfaction of the
occupants, and the green circle representing the neutral status of each question.
55

Figure 12 Rose Chart of Survey Questions

In the rose chart of Figure 12, 16 representative questions from the survey are
compared with each other. The mean value of the votes of each question is marked at above
0 points, which means that occupants are generally satisfied or neutral with the
environment. The blue shape mostly matches with the yellow circle representing overall
satisfaction. However, there are some criteria that the mean vote didn’t reach the level of
overall satisfaction. However, some of the questions that are related to IAQ and acoustic
quality of the workstation are found to show relatively lower satisfaction level than overall
satisfaction.
56

Figure 13 Rose Chart of Survey Questions by IEQ Criteria

Recognizing the difference of the mean value of the votes in questions of different
IEQ criteria, rose chart in Figure 13 is drawn to illustrate the pattern. The graph is generated
based on the result of four representative questions of each IEQ criteria. This time, the blue
form is generated at around neutral level, which is slightly smaller than the previous blue
form. Interestingly, acoustic quality, in this time, shows relatively better satisfaction level
than the others. This may mean that selected questions of Figure 12 are not reflecting the
overall satisfaction of the occupants. Thus, the necessity of more intensive approach on
data such as statistical analysis is identified.

57

Chapter 5: Data Analysis and Findings
In this chapter, data collected from the previous chapter are analyzed by statistical
analysis software. Stepwise regression is used to identify the impact questions among the
survey while correlation analysis identifies the relation between IEQ measurement and
survey questions. Then, findings from Analysis of Variance (ANOVA) are illustrated by
its different impact factors. Finally, decision tree (J48) that reveals a hierarchy of the impact
factors are introduced and suggested for utilizing the research findings.

5.1 Statistical Analysis
5.1.1 Identifying Impact Questions
As a first step toward statistical analysis, stepwise regression is used on occupants’
satisfaction survey. Stepwise regression is a useful tool when there are many variables and
need to identify a useful subset of the predictors. Stepwise regression selects a model by
automatically adding or removing individual predictors, a step at a time, based on their
statistical significance. The result of the process is a single regression model, which shows
the clear and simple hierarchy of the factors. In Minitab software, it starts with no
predictors in the model and adds the most significant variable for each step. Stops are
recorded when all variables not in the model have p-values that are greater than the
specified Alpha-to-Enter value. In addition, Mallows' Cp is a value that helps to choose
between multiple regression models by importance. It compares the precision and bias of
the full model to models with a subset of the predictors. Smaller Mallows' Cp indicates
relatively precise model with small variance in estimating [54].

58

Table 6 Stepwise Regression on Q29 vs. Q6 - 28
Stepwise Regression
Step 1 Step 2 Step 3 Step 4
Coef P Coef P Coef P Coef P
Constant 0.5323 0.3048 0.2584 0.2320
Q23 0.3682 0.000 0.2664 0.000 0.2235 0.000 0.2326 0.000
Q22 0.2800 0.000 0.2601 0.000 0.2381 0.000
Q12 0.1179 0.004 0.1206 0.003
Q27 0.950 0.013
S 1.06876 1.00775 0.998310 0.991606
R-sq 22.42 % 31.20 % 32.66 % 33.74 %
Mallows’ Cp 70.24 21.04 14.53 10.26
Step5 Step 6 Step 7

Coef P Coef P Coef P
Constant 0.2693 0.1893 0.1816
Q23 0.2185 0.000 0.2069 0.000 0.2026 0.000
Q22 0.2281 0000 0.2136 0.000 0.1644 0.001
Q12 0.1125 0.006 0.0983 0.017 0.0866 0.037
Q27 0.1035 0.007 0.0882 0.024 0.0773 0.049
Q14 0.0700 0.037 0.0706 0.034 0.0672 0.044
Q6 0.0863 0.043 0.0861 0.043
Q18 0.0885 0.055
S 0.987223 0.983189 0.979710
R-sq 34.50 % 35.20 % 35.83 %
Mallows’ Cp 7.85 5.74 4.08

Table 6 is a result of stepwise regression analysis on Q29 (Satisfaction on overall
environment) versus rest of the questions on IEQ satisfaction (Q6 through Q28). The
reason for this regression analysis is to find out factors that are relevant to occupants’
satisfaction on the overall environment. According to the result, Q23 (satisfaction on
workstation layout), Q22 (satisfaction on quality of overall lighting), Q12 (satisfaction on
degree of enclosure), Q27 (satisfaction on lighting condition), Q14 (satisfaction on access
to thermostat), Q6 (satisfaction on job), and Q18 (satisfaction on lighting for computer
work) has shown significant relation in order on Q29. As 3 out of 7 selected questions are
59

about lighting satisfaction, it can be interpreted that occupants’ satisfaction on lighting may
have more significant impact on overall satisfaction compared to the other factors.

Table 7 Stepwise Regression on Q29 vs. Q15, Q26 - 28
Stepwise Regression
Step 1 Step 2
Coef P Coef P
Constant 0.6224 0.5694
Q15 0.2726 0.000 0.2677 0.000
Q27 0.1273 0.003
S 1.14032 1.12855
R-sq 12.09 % 14.12 %
Mallows’ Cp 8.13 1.09

Similarly, stepwise regression analysis on Q29 (satisfaction on overall environment)
versus satisfaction on each IEQ criteria (Q15 – air quality, Q26 – temperature, Q27 –
lighting condition, and Q28 – acoustic condition) is conducted to determine the most
relevant factors of each IEQ comfort. As shown in Table 7, Q15 and Q27 has demonstrated
the relevance to satisfaction on the overall environment. Interestingly, Q27 has shown its
impact on overall environment satisfaction again that it represents the importance of
lighting in occupants’ comfort. The analysis revealed the rough interpretation of the data
which suggest considerate analysis on air quality satisfaction and lighting satisfaction.

60

Table 8 Stepwise Regression on Q27 vs. Q18 – 22: Lighting
Stepwise Regression
Step 1 Step 2 Step 3
Coef P Coef P Coef P
Constant 0.2732 0.3086 0.2607
Q18 0.1943 0.000 0.2423 0.000 0.1891 0.003
Q20 -0.0849 0.136 -0.1149 0.053
Q22 0.1142 0.078
S 1.32886 1.32675 1.32313
R-sq 4.26 % 4.82 % 5.58 %
Mallows’ Cp 4.35 4.11 3.00

Table 9 Stepwise Regression on Q29 vs. Q18 – 22: Lighting
Stepwise Regression
Step 1 Step 2 Step 3
Coef P Coef P Coef P
Constant 0.3638 0.3235 0.3176
Q22 0.3920 0.000 0.2966 0.000 0.2375 0.000
Q20 0.1665 0.000 0.1243 0.007
Q18 0.1265 0.013
S 1.10871 1.08993 1.08293
R-sq 19.62 % 22.51 % 23.69 %
Mallows’ Cp 21.95 8.60 4.33

Stepwise regression analysis can also be used to point out the representing factor
of each IEQ satisfaction criteria (lighting, temperature, air quality, acoustic quality). For
instance, Q18 through 22 of the survey are about occupants’ satisfaction on lighting
condition (Q18 – light for computer work, Q19 – reflected light or glare on a computer
screen, Q20 – direct glare from a light fixture, Q21 – direct glare from daylight, and Q22
– quality of overall lighting). As shown in Table 8 and Table 9, these survey questions are
compared with Q27 (satisfaction on lighting condition) and Q29 (satisfaction on overall
environment) through stepwise regression. As a result, Q18, Q20, and Q22 are pointed out
as significant factors on occupants’ lighting satisfaction and visual comfort.
61

Table 10 Stepwise Regression on Q26 vs. Q13 – 14, Q25: Temperature and Air Quality
Stepwise Regression
Step 1 Step 2
Coef P Coef P
Constant 0.0333 -0.0239
Q25 0.1864 0.000 0.1798 0.000
Q13 0.0812 0.042
S 1.33349 1.32834
R-sq 4.66 % 5.63 %
Mallows’ Cp 6.24 4.08

Table 11 Stepwise Regression on Q15 vs. Q13 – 14, Q25: Temperature and Air Quality
Stepwise Regression
Step 1 Step 2 Step 3
Coef P Coef P Coef P
Constant 0.2295 0.1927 0.2969
Q13 0.3903 0.000 0.3685 0.000 0.2914 0.000
Q25 0.2974 0.000 0.2762 0.000
Q14 0.2088 0.000
S 1.43970 1.36287 1.32848
R-sq 16.84 % 25.66 % 29.54 %
Mallows’ Cp 72.83 24.24 4.00

Table 12 Stepwise Regression on Q29 vs. Q13 – 14, Q25: Temperature and Air Quality
Stepwise Regression
Step 1 Step 2 Step 3
Coef P Coef P Coef P
Constant 0.6049 0.5845 0.6284
Q13 0.2293 0.000 0.2172 0.000 0.1847 0.000
Q25 0.1652 0.000 0.1562 0.000
Q14 0.0880 0.022
S 1.17813 1.15040 1.14431
R-sq 9.45 % 13.88 % 15.00 %
Mallows’ Cp 26.35 7.32 4.00

Similar analysis is conducted on temperature and air quality satisfaction. Q26
(satisfaction on overall temperature) and Q15 (satisfaction on overall air quality) are
62

analyzed together as a group with Q13 (satisfaction on temperature), Q14 (satisfaction on
accessibility to thermostats), and Q25 (satisfaction on air movement) as it affects each other
regarding human perception. The result is shown in Table 10, Table 11, and Table 12 that
Q13 and Q25 consistently show significance on its relevance to temperature and air quality,
while Q14 does not show its relevance with overall temperature satisfaction.

Table 13 Stepwise Regression on Q28 vs. Q16 – 17: Acoustic Quality
Stepwise Regression
Step 1
Coef P
Constant 0.5564
Q17 -0.2094 0.000
S 1.33693
R-sq 5.67 %
Mallows’ Cp 1.37

Table 14 Stepwise Regression on Q29 vs. Q16 – 17: Acoustic Quality
Stepwise Regression
Step 1 Step 2
Coef P Coef P
Constant 0.6930 0.6525
Q16 0.2145 0.000 0.1453 0.002
Q17 0.1180 0.010
S 1.19252 1.18421
R-sq 6.86 % 8.38 %
Mallows’ Cp 7.75 3.00

Finally, stepwise regression analysis is conducted between questions of acoustic
quality. The result in Table 13 and Table 14 indicates the significance of Q17 (non-verbal
noise) on both Q28 (acoustic condition) and Q29 (overall environment) while Q16 (noise
from the conversation) only found in Q29.
63

Table 15 Stepwise Regression on Q29 vs. Q13, Q16, Q22, Q25
Stepwise Regression
Step 1 Step 2 Step 3 Step 4
Coef P Coef P Coef P Coef P
Constant 0.3653 0.3263 0.3334 0.3280
Q22 0.3936 0.000 0.3397 0.000 0.3135 0.000 0.2958 0.000
Q13 0.1324 0.000 0.1314 0.000 0.1209 0.000
Q25 0.1200 0.001 0.1019 0.004
Q16 0.0984 0.008
S 1.10995 1.09192 1.07727 1.06931
R-sq 19.75 % 22.53 % 24.78 % 26.07 %
Mallows’ Cp 33.37 20.26 10.02 5.00

Through the series of analysis using stepwise regression, questions that are most
relevant to Q29 (satisfaction on overall environment) could be marked in each IEQ criteria.
For instance, Q22 (quality of overall lighting) was the most significant factor on Q29
among the questions related to lighting quality. Similarly, Q13 (temperature) from
temperature, Q25 (air movement) from air quality, and Q16 (noise from the conversation)
are selected as representing question of each IEQ. With these selected questions, stepwise
regression is once again conducted to find out the most significant IEQ criteria on overall
environment satisfaction. The result is displayed in Table 15, which shows Q22 as a
question with the most relevance to Q29 and Q16 as the least relevant to the selected
questions. Once again, its significance of lighting satisfaction on overall environmental
satisfaction is seen from the analysis. Compared to the result of the very first stepwise
regression analysis in Table 6, Q22 is included in both groups of selected questions. It can
be interpreted that quality of lighting has significant relevance in occupants’ overall
environmental satisfaction.

64

5.1.2 Correlation between IEQ and Occupants’ Satisfaction
To find out its relationship between IEQ measurements and satisfaction survey
questions, correlation analysis is conducted. As mentioned in the previous chapter,
correlation analysis does not represent causation. However, it is a meaningful process to
point out the combinations that might have relevance in certain ways. In correlation
analysis, Pearson Correlation is the most common measure used. The value on Pearson
Correlation lies between -1 and 1. The closer the value gets to zero, the greater the variation,
meaning lower correlation. Another measure used in correlation analysis is P-value. The
P-value is the probability that would have found the current result if the correlation
coefficient were, in fact, zero (null hypothesis). If this probability is lower than 5% (P-
value < 0.05) the correlation coefficient is called statistically significant. Based on these
two measures, combinations with possible relation based on correlation analysis are
selected and recorded in Table 16.

Combinations in Table 16 are first selected based on its significance of P-value that
shows lower than 0.05 (or similar). Most of the selected combinations show low value in
its Pearson Correlation, which means low correlation. However, this correlation analysis is
made based on an entire dataset without considering human factors or spatial factors of the
occupants who took the survey and workstations where the data are collected. Therefore,
the necessity of analysis on distributed data based on human factors and spatial factors are
found and conducted in following sections.

65

Table 16 Correlation Analysis between IEQ Measurement and Satisfaction Survey
Correlation Analysis
Question
Category
IEQ
Measurement
Pearson
Correlation
P-value
Q6
Job Satisfaction
CO2 0.095 0.055
Ambient Temperature 0.154 0.055
Acoustic -0.112 0.024
Average Luminance -0.312 0.000
UGR -0.239 0.002
Q8
Privacy
Acoustic -0.094 0.058
Q11
Privacy
Acoustic -0.131 0.008
Average Luminance -0.179 0.023
Q12
Privacy
Acoustic -0.104 0.036
Q13
Temperature
Air Velocity 1.1m 0.134 0.022
Air Velocity 0.1m -0.142 0.015
Average Luminance -0.161 0.042
Q14
Temperature
Air Velocity1.1m 0.244 0.000
Air Velocity 0.1m 0.146 0.013
Q15
IAQ
Air Velocity 0.1m -0.151 0.010
Ambient Temperature -0.183 0.001
Q25
Air Movement
Air Velocity 1.1m 0.413 0.000
Air Velocity 0.1m 0.191 0.001
Ambient Temperature 0.112 0.047
Q27
Lighting
Screen Illuminance 0.114 0.048
Q29
Overall Satisfaction
CO2 0.099 0.046

66

5.1.3 Analyses by Different Occupant Group
As explained in Chapter 3, analysis of variance (ANOVA) is used to determine
whether there are any significant differences between the means of three or more
independent groups. Two sample t-test is usually used when there are only two groups to
compare the means of the samples, but ANOVA can also be used to bring out the same
result. In this section, the process of analysis on distributed data based on human factors
and spatial factors are explained and illustrated. As an example, a combination of measured
air velocity at 1.1m and Q25 (satisfaction on air movement) is used.

Figure 14 Interval Plot of Data by Different Point Scale

3 2 1 0 -1 -2 -3
0.4
0.3
0.2
0.1
0.0
Q25
AIR VELOCITY 1.1
Interval Plot of AIR VELOCITY 1.1
95% CI for the Mean
Individual standard deviations are used to calculate the intervals.
Positive Neutral Negative
0.4
0.3
0.2
0.1
0.0
Q25_1
AIR VELOCITY 1.1
Interval Plot of AIR VELOCITY 1.1
95% CI for the Mean
Individual standard deviations are used to calculate the intervals.
Positive Negative
0.4
0.3
0.2
0.1
0.0
Q25_2
AIR VELOCITY 1.1
Interval Plot of AIR VELOCITY 1.1
95% CI for the Mean
Individual standard deviations are used to calculate the intervals.
67

The very first step to identify the distribution of the data is simply running ANOVA
between two selected data. As the survey is taken using 7-point scale satisfaction level on
each question, distribution is made in 7 different groups. Then, the analysis is conducted
again using lower number-point scale (3-point scale and 2-point scale) to simplify and
clarify the relationship as displayed in interval plot graphs of Figure 14.

It is clearly seen that group of occupants responded positively on air movement
tend to seated in a workstation with higher air velocity compared to a group of occupants
responded negatively to air movement. However, this analysis is still made based on an
entire dataset that different features of each occupant are not considered. Since each of the
occupants have a different preference of environmental setting depending on one’s physical
or biological status, analysis on occupant groups of similar status is required. To take into
consideration of occupants’ status, human factors (gender and age) and spatial factors
(building type and workstation location) are to be considered in next step.

Human factors used for data distribution are gender and age of the occupants as
illustrated in Figure 15. Data can be simply divided into two groups by gender: In the graph,
F represents Female and M for Male. However, five different groups were set in original
survey for occupants’ age group: 18-29, 30-39, 40-49, 50-59, and 60-69. To simplify the
grouping and increase the credibility of the statistical analysis, data is re-distributed into
three different age groups: Junior, Mid-age, and Senior. In the graph, N stands for negative
responses while P represents positive responses. As seen in the graphs, the mean value of
68

air velocity varies by different groups of occupants by human factor. Especially, occupants
of the senior group show distinctive pattern compared to the rest of the occupant groups.

Figure 15 Interval Plot of Data by Human Factor

Similarly, data is divided into different groups by its spatial factors to see its impact
on occupants’ environmental satisfaction. In Figure 16, interval plot graphs are generated
by different building types and workstation location. It can be observed that overall pattern
of the graphs is the same, but the graph of occupant group in commercial buildings are
generated in a lower value of air velocity. In this case, the finding can be a starting point
69

to make a suggestion on setting of an air conditioner or a fan of workstations in a different
type of building.

Figure 16 Interval Plot of Data by Spatial Factor

Series of analysis using analysis of variance (ANOVA) is conducted on each
combination of IEQ measurement and its corresponding satisfaction survey questions.
Corresponding IEQ measurements for each satisfaction survey questions are selected based
on its correlation analysis and intuitive match from common sense standpoint.
Combinations analyzed are as follow in Table 17. Combinations can be categorized into
four different groups based on IEQ comfort: Lighting, Temperature / IAQ, Acoustic, and
Overall Environment.

70

Table 17 IEQ Measurement and Survey Combinations Analyzed
Category Survey Question IEQ Measurements
Lighting
Q18
Q19
Q20
Q21
Q22
Q27
Computer Work
Screen Reflection
Glare from Light Fixture
Glare from Daylight
Overall Lighting Quality
Lighting Condition
Work Surface Illuminance
Screen Illuminance
Reading Zone Illuminance
Average Luminance
UGR
Temperature

/ IAQ
Q13
Q14
Q15
Q25
Q26
Temperature
Thermostat Access
Air Quality
Air Movement
Ambient Temperature
CO2
Ambient Temperature
Air Velocity 1.1m
Air Velocity 0.1m
Acoustic
Q16
Q17
Q28
Verbal Noise
Non-verbal Noise
Acoustic Condition
Acoustic Decibel
Overall
Environment
Q29 Overall Environment
Work Surface Illuminance
Screen Illuminance
Reading Zone Illuminance
Average Luminance
UGR
CO2
Ambient Temperature
Air Velocity 1.1m
Air Velocity 0.1m
Acoustic Decibel

Among the series of analysis, combinations that show interesting result are selected
and carefully observed to make inferences from it. Two of noticeable findings from each
factor (gender, age, building type, and workstation location).

71

5.2 Impact of Factors on Occupants’ Satisfaction
5.2.1 Impact of Gender
The first set of findings are the analyses showing the impact of gender on occupants’
satisfaction. Graphs on the left side of Figure 17 represent the distribution of Q21
(occupants’ satisfaction on direct glare from daylight) by average luminance level. The
distribution of the data is observed to be similar in overall data, but the pattern changes
when the data are re-distributed by gender. According to graphs on the right side of Figure
17, female occupants answered positively on Q21 at average luminance level of around 60
cd/m
2
while male occupants answered negatively. Male occupants responded positive on
Q21 are found at average luminance level of around 100 cd/m
2
. Interestingly, the mean
value of average luminance which female occupants responded as satisfied is almost same
as the mean value of male occupants responded dissatisfied. Different level of control on
luminance at the workstation can be suggested for occupants with different gender based
on this finding.

Figure 17 Interval Plot of Average Luminance Level by Gender

Positive Negative
140
120
100
80
60
40
20
0
Q21_2
Avg Luminance cd/m2
Interval Plot of Avg Luminance cd/m2
95% CI for the Mean
Individual standard deviations are used to calculate the intervals.
72

Another example showing the impact of gender is an impact of air velocity at 1.1m
on Q15 (occupants’ satisfaction on air quality). In Figure 18, the graph shows that female
occupants are feeling satisfied with air quality to tend to answer at the condition with higher
air velocity at 1.1m height than satisfied male occupants. Especially, mean values of air
velocity where male and female responded negative and positive on air quality are almost
opposite to each other. Therefore, considering air velocity at the workstation depending on
the gender of the occupant can be recommended.

Figure 18 Interval Plot of Air Velocity at 1.1m by Gender

5.2.2 Impact of Age
The second series of finding is that aging has an impact on occupant’s satisfaction
in the indoor environment. As illustrated in the graph on the left side of Figure 19, people,
in general, tend to answer higher satisfaction on Q18 (lighting for computer work) as its
work surface illuminance level increase. The data generates slightly direct proportional
graph when it is drawn. However, when the data are distributed into different age groups,
the tendency appears differently. While junior and mid-aged group shows the similar
tendency to the overall result, senior group shows a graph with opposite direction.
Positive Negative
0.3
0.2
0.1
0.0
Q15_2
AIR VELOCITY 1.1
Interval Plot of AIR VELOCITY 1.1
95% CI for the Mean
Individual standard deviations are used to calculate the intervals.
73

Interestingly, mean work surface illuminance level of the senior group who answered
positively on Q18 is close to illuminance level of the junior and mid-aged respondents who
answered negatively. Moreover, mean illuminance level of the negative-answered senior
respondents lies at around 600 lux, which is higher than the illuminance level of positively
answered occupants from other groups. This difference may be caused by the aging process
of the human body that different age group has a different preference in work surface
illumination level. Based on this result, illuminance level higher than 400lux is suggested
for the junior and mid-aged group, while illuminance level lower than 400lux is
recommended for the senior group.

Figure 19 Interval Plot of Work Surface Illuminance by Age Group

Similarly, there were other analyses showing the impact of different age group on
occupants’ satisfaction. In Figure 20, occupants’ answered positively on Q25 (satisfaction
on air movement) lies in condition with 0.1 m/s or higher air velocity at 1.1m height.
However, data of senior group generates opposite directional graph to other age groups,
which means negative response at the condition with air velocity higher than 0.1m/s. Also,
the mean value of air velocity where the occupants of each group responded positively on
Positive Negative
1200
1000
800
600
400
200
0
Q18_2
WORK SURFACE LUX
Interval Plot of WORK SURFACE LUX
95% CI for the Mean
Individual standard deviations are used to calculate the intervals.
74

air movement gradually goes lower as the occupant group becomes older. The graph can
be interpreted that people in a senior group tend to prefer lower air velocity for air
movement satisfaction compared to junior and mid aged occupant group. Based on this
finding, lower air velocity can be suggested on the design of senior occupants’ workstation.

Figure 20 Interval Plot of Air Velocity at 1.1m by Age Group

5.2.3 Impact of Building Type
The third set of findings is analyses on the impact of building type on occupants’
environmental satisfaction. As described in the previous chapter, data are measured from
office environment within two different types of buildings: commercial buildings and
university buildings. Graphs in Figure 21 shows the distribution of answers on Q27
(satisfaction on lighting condition) by measured work surface illuminance level. In overall
data, higher work surface illuminance level is measured from workstations of occupants
who answered positively on Q27, compared to occupants answered negatively. The
tendency remains the same even after the graphs are generated by different building type.
However, the graph of university buildings is located in higher work surface illuminance
level than the graph of commercial buildings. Mean work surface illuminance level of
Positive Negative
0.3
0.2
0.1
0.0
Q25_2
AIR VELOCITY 1.1
Interval Plot of AIR VELOCITY 1.1
95% CI for the Mean
Individual standard deviations are used to calculate the intervals.
75

negatively responded occupants in university buildings is even higher than it of positively
responded occupants in commercial buildings.

Figure 21 Interval Plot of Work Surface Illuminance by Building Type

Another combination that showed the impact of different building type in its result
is air velocity at 1.1m and Q25 (satisfaction on air movement). In overall data, occupants
who answered positively on Q25 were seated in workstations with higher air velocity at
1.1m height. Mean air velocity of positively answered occupants lies at around 0.17 m/s
while negative answered occupants read 0.07 m/s. When the data are distributed by
building type, the similar shape of graphs are generated in both commercial and university
buildings, but the graph of university buildings is observed in higher air velocity.
Occupants of commercial buildings are observed to be satisfied with air movement in air
velocity around 0.1 m/s. However, occupants of university buildings answered negatively
in air velocity of 0.1 m/s and preferred air velocity around 0.2 m/s. Especially, the
combination of air velocity at 1.1m and Q25 is significant, since the combination is already
observed once in the previous section as an example showing the impact of different age.
Positive Negative
600
500
400
300
200
Q27_2
WORK SURFACE LUX
Interval Plot of WORK SURFACE LUX
95% CI for the Mean
Individual standard deviations are used to calculate the intervals.
76

It can be inferred that occupants’ comfort on air movement is sensitively affected by both
occupant’s age and type of building where the workstation is located.

Figure 22 Interval Plot of Air Velocity at 1.1m by Building Type

5.2.4 Impact of Workstation Location
The last set of findings is analyses on the impact of workstation location to
occupants’ environmental satisfaction. Graphs in Figure 23 illustrate Q27 (occupants’
satisfaction on lighting condition) based on illuminance level of the reading zone.
Distribution of occupants’ satisfaction on lighting condition seems to be irregular at first.
However, the difference in pattern occurs when the data are redistributed into two groups
based on their workstation location; center and perimeter. It is clearly seen that occupants
of a workstation located at the perimeter area of the building tend to feel satisfied in higher
illuminance level of reading zone compared to occupants at the center area of the building.
It is somewhat obvious that workstations located at the perimeter area of the building to
have a higher chance of exposure to the natural light; therefore, higher illuminance is
required at the reading zone of perimeter workstation for a clearer view. Thus, higher
reading zone illuminance of 400lux is recommended for the design of workstation at
Positive Negative
0.25
0.20
0.15
0.10
0.05
0.00
Q25_2
AIR VELOCITY 1.1
Interval Plot of AIR VELOCITY 1.1
95% CI for the Mean
Individual standard deviations are used to calculate the intervals.
77

perimeter area, while lower reading zone illuminance level of 270lux is recommended for
a workstation at the center area of the building.

Figure 23 Interval Plot of Reading Zone Illuminance by Workstation Location

Another example of the impact of workstation location was found in analyses
between air velocity at 0.1m height and Q25 (occupants’ satisfaction on air movement).
Figure 24 illustrates the tendency of overall occupants answering satisfied at higher air
velocity to the ones answered negatively. The preference of occupants remains the same
even the data are distributed by different workstation location. However, the distinguishing
difference is found in its actual measure of air velocity. The mean value of air velocity
from satisfied occupants at the center of the building was around 0.07m/s. However, data
shows that occupants at the perimeter of the building responded negatively in condition
with this air velocity. Instead, the mean value of 0.09m/s was marked by occupants’
responded positively at the perimeter of the building. Due to its location, higher air velocity
at 0.1m height was measured in workstations at the perimeter of the building to the center
of the building. However, still, occupants at the perimeter require higher air velocity for
the satisfaction. This may result from the higher temperature of the perimeter. Therefore,
Positive Negative
500
400
300
200
Q27_2
READING ZONE LUX
Interval Plot of READING ZONE LUX
95% CI for the Mean
Individual standard deviations are used to calculate the intervals.
78

consideration of the location of the workstation is suggested when air velocity is controlled
in the building.

Figure 24 Interval Plot of Air Velocity at 0.1m by Workstation Location

5.3 In-depth Analysis Using an Advanced Data Mining Tool
As an additional step to investigate the correlations and hierarchy within collected
data, data mining computer software is used. Data mining is the computational process of
discovering patterns in large data sets involving methods at the intersection of artificial
intelligence, machine learning, statistics, and database systems. The overall goal of the data
mining is to capture information from data and transform it into a simple format. Data
mining computer software Weka is developed by machine learning group at the University
of Waikato in New Zealand [53]. The software automatically analyses a large body of data
and decide what information is most relevant. Decision Tree (J48) is generated using
various combination of collected data, about statistical analyses conducted in the previous
section. The first round is conducted with an entire set of survey questions to identify the
questions with recognizable hierarchy. Then, the second round of analyses are conducted
Positive Negative
0.15
0.10
0.05
0.00
Q25_2
AIR VELOCITY 0.1
Interval Plot of AIR VELOCITY 0.1
95% CI for the Mean
Individual standard deviations are used to calculate the intervals.
79

with a combination of survey questions and IEQ measurement. In the end, human factors
and spatial factors are included to find its impact just as it was done with statistical analyses.

As illustrated in Figure 25, certain questions among the entire survey are found to
have higher relevance and hierarchy between them. Referring to the survey questions listed
in Table 2, Q22 (satisfaction on the quality of lighting) has shown the most significance.
The result means, to identify human comfort Q22 is to be asked first. Depending on the
answer to Q22, positively answered occupants are to be asked about their satisfaction on
Q15 (air quality), Q20 (direct glare from light fixture), and Q25 (air movement). On the
other hand, occupants with negative satisfaction on Q22 are asked with Q7 (size of the
workstation) to redistribute them into two groups. Positive answer on Q7 leads to Q28
(acoustic condition), Q18 (light for computer work), Q16 (noise from conversation), Q27
(lighting condition), Q8 (level of privacy), and Q24 (color / texture / finishes) in order,
while negative in Q7 leads to Q16 (noise from conversation) and Q25 (air movement). The
questions mentioned are found to have relatively higher significance based on data mining
process. Thus, occupants’ comfort and satisfaction on a workstation can be determined
through this process.

80

Figure 25 Decision Tree (J48) from Entire Survey Questions

However, using entire survey questions may cause the redundancy in identifying
IEQ criteria that affects occupants’ comfort. Therefore, 5 of the questions selected from
previous stepwise regression analysis are again used to represent each of the IEQ
components: Q13 (temperature), Q16 (noise from the conversation), Q22 (quality of
overall lighting), and Q25 (air movement). Based on these selected questions, occupants
are asked with Q22 first to determine whether the occupant is feeling satisfied or not. As
seen in Figure 26, Q16 is asked to occupants’ who answered negatively to find out their
satisfaction on noise. Finally, Q13 is asked to occupants with constant negative answers to
determine their thermal satisfaction. Just from this result, Lighting has shown the most
significance in determining human comfort, followed by acoustic and thermal comfort.

81

Figure 26 Decision Tree (J48) from 5 Selected Survey Questions

In addition to previously selected five survey questions, five of selected IEQ
measurements are added to the database to analyze the relation. Selected IEQ
measurements include Work Surface Illuminance, Air Velocity at 1.1m, CO2 Level,
Ambient Temperature and Acoustic Decibel. These measurements correspond to IEQ
comfort that is asked to occupants of survey questions. As shown in Figure 27, Q22 (quality
of overall lighting) and Q16 (noise from the conversation) from survey remain as the two
most significant factors on occupants’ satisfaction. Depending on the answer to Q16,
negatively answered group is subdivided by Q13 (temperature) and CO2 level while the
positive group is clustered by Q25 (air movement) and acoustic Decibel.

82

Figure 27 Decision Tree (J48) from Selected Survey Questions and IEQ Measurements

Similarly, human factors (gender, age) and spatial factors (building type,
workstation location) that are found to have an impact on occupants’ comfort are combined
with five selected survey questions to determine the hierarchy in the decision tree (J48). In
Figure 28, Q22 (quality of overall lighting) remain as the most significant factor in the
decision tree (J48). Then, workstation location and building type become the next
significant factors to divide groups. In university buildings, Q25 (air movement) and
gender are important factors, while Q16 (noise from the conversation) is essential in
commercial buildings.
83

Figure 28 Decision Tree (J48) from Selected Survey Questions and Impact Factors

Finally, selected factors from each data group are combined to generate a decision
tree (J48) in Figure 29. Among combined data of selected five survey questions, IEQ
measurements, human factors, and spatial factors, Q22 (quality of overall lighting) is found
to be the most significant factor in deciding satisfaction on workstation environment. As
the result is somewhat combined outcome of previous decision trees, Q13 (temperature)
and workstation location takes the second most significant position depending on the
answer to Q22. For the group of occupants that are positive on the quality of lighting,
factors such as Q13, acoustic decibel, and workstation location are predominant. Q25 (air
movement), Q16 (noise from the conversation), building type, temperature, and age
follows as supplementary factors. On the other hand, workstation location and building
type are dominant factors to a group of occupants who answered negatively on Q22. For
this group of occupants, Q25, temperature, work surface illuminance, and Q13 plays a role
as minor factors affecting occupants’ comfort at the workstation.
84

Figure 29 Decision Tree (J48) from Combined Data

As a result suggested decision-making tool has identified the hierarchy of the
factors by its relevance to satisfaction on overall environment of the workstation.
Comparing the result with statistical analyses conducted in the previous section, Q22
(satisfaction on the quality of overall lighting) has been identified as one of the significant
factors among all. The decision-making tool not only supports the previously conducted
analyses but also suggests a possible analysis for the future research. For instance,
conducting a new set of statistical analysis on subdivided groups based on decision-making
software may bring out the result with higher statistical relevance. However, the tool is
introduced in this research just to support the necessity of the research and widen up the
possibility of future studies.

85

Chapter 6: Conclusion
Chapter 6 summarizes the thesis with a discussion of the current problem in
building indoor environmental control through findings. Confronted difficulties and
limitations found throughout the progress of the research are mentioned to suggest a future
direction. Especially, adding more number of supplementary data into the current database
may change the result and increase the reliability of the result. In the end, the conclusion
is made to close the thesis.

6.1 Discussion
The notion is accepted that each human being has different preferences in
surrounding environment, and detailed adjustment is required to meet everyone’s
satisfying level. However, it is also true that it is almost impossible to fulfill everyone’s
request due to the contradictions among them. Therefore, post-occupancy evaluation (POE)
is commonly examined and seeks for problems that are affecting occupants’ comfort within
the building.

Modern buildings these days follow regulated standards such as ASHRAE and
IESNA to assure quality indoor environment for occupants. The buildings examined in this
research are also operated by following IEQ standards and guidelines. However, the results
show the existence of unsatisfied occupants even under uniformly controlled environment.
Based on the result, it can be inferred that more detailed and adaptive environmental control
in consideration of human factor and the spatial factor is required to meet individual
occupants’ environmental satisfaction.
86

The research has conducted a post-occupancy evaluation of indoor environmental
conditions (lighting, thermal, air quality and acoustic) at 411 workstations in 14 modern
buildings of Southern California. By integrating actual measurement and building
attributes with occupants’ satisfaction surveys, the common weakness of objectivity in
POE was strengthened. Also, statistical analyses on collected data have found interesting
correlations and impacts between various IEQ factors. Especially, correlations regarding
human factors and spatial factors are highlighted from the research. Occupants of different
gender and age group have shown different satisfaction in different IEQ condition. This
supports the existing notion of the impact of a human factor on satisfaction. The difference
in building type and location of the workstation was also identified as an impact factor for
human satisfaction.

6.2 Limitations and Future Directions
Besides these findings, there still are limitations and future improvements for the
research. As a common issue for research with statistical analysis methods, clearer and
stronger results can be made with a larger number of input data. Even though 411 research
samples are not a small number for statistical analyses, the number becomes smaller when
it is subdivided into groups based on age, gender, and location. Especially, numbers of
occupants were not evenly distributed in age groups due to characteristics of the buildings
where the research took place. The findings from the different age group might become
more distinctive and interesting with a larger number of data collection.

87

Another limitation and interesting improvement that can be made for this research
is control of the building size or layout of the office in the floor plan. Due to a limitation
in access, data was collected from some different buildings with various building footprints.
The analyses done in this research has its value regarding selected spatial factors of the
building, it might be even more accurate and make sense to control the factors to clarify
the impact.

For future direction, findings from this type of research can be utilized as a
modification of current IEQ standards and guidelines. Instead of uniform and fragmentary
standards for generalized occupants, a diverse and adaptive suggestion for specified target
occupant may be possible. As partially explained with the decision-making tool,
considering impact factors by their hierarchy may be a solution. An outcome of the research
can be made in the form of design guidelines based on POE, which may be able to replace
the existing standards and regulations.

6.3 Conclusion
This research has discussed the satisfaction of occupants with their environmental
quality in an office environment. Findings of the research support an idea that
environmental satisfactions are significantly affected by human factors and spatial factors
such as gender and age, building type, and location of the workstation. Identifiable
differences were found in some of the IEQ measurements based on different gender and
age groups. Female occupants are satisfied with lower luminance levels and higher air
velocity compared to male occupants. Similarly, lower work surface illuminance level and
88

air velocity are preferred by the senior age group compared to junior and middle aged group.
By spatial factor, occupants of university buildings tend to show a preference for higher
work surface illuminance level and air velocity. Occupants of perimeter area are more
satisfied with higher reading zone illuminance level, and air velocity than occupants at the
center of the building.

Based on these results, indoor environment conditions of office environments can
be suggested to be designed with adjustable lighting and air movement system based on
the occupant of the workstation. As an example, senior aged female occupants located in
the center area of the building may require a lower level of lighting compared to junior
aged male occupants located at the perimeter area of the building. Equivalently, junior aged
female occupants located at the perimeter area of the building may prefer higher air velocity
than senior aged male occupants located in the center area of the building.

The result of the research may be limited to the collected data. However, the
research has value as it suggests a methodological approach to making design guidelines
based on impact factors. Adding more numbers to the dataset will reinforce the solidity of
the research findings and significance of impact factors. Also, revealing the hierarchy of
the impact factors through data mining will reinforce the research findings. With all the
improvements made, research findings may be used for developing design guidelines that
can effectively carry out the project within limited resources such as time and budget.
However, as statistical analysis may always have accidental erroneous findings, expert
knowledge on the issue should be accompanied.
89

Appendix: Satisfaction Survey by Question
In this section, satisfaction survey data is displayed using different types of graphs.
Distribution of answers to each satisfaction survey question is illustrated through bar chart
graph. The most frequent answer among seven choices (-3: very dissatisfied, -2:
dissatisfied, -1: slightly dissatisfied, 0: neutral, +1: slightly satisfied, +2: satisfied, +3: very
satisfied) made by occupants can be found in this graph. Also, boxplot is used to display
data in another way with a mark of mean value.

Question 6 asks occupants’ satisfaction on job or institution. As shown in Figure
30, the most frequent answer made by occupants is +2 (satisfied). The mean value of
satisfaction level lies between +1 and +2.

Figure 30 Distribution of Answers to Q6 and Box Plot

Question 7 asks occupants’ satisfaction on the size of the workspace. As shown in
Figure 31, the most frequent answer made by occupants is +2 (satisfied). The mean value
of satisfaction level lies near +1.

3 2 1 0 -1 -2 -3
160
140
120
100
80
60
40
20
0
Q6: Job Satisfaction
Count
Chart of Q6
3
2
1
0
-1
-2
-3
Q6
Boxplot of Q6
90

Figure 31 Distribution of Answers to Q7 and Box Plot

Question 8 asks occupants’ satisfaction on the level of privacy in the workstation.
As shown in Figure 32, the most frequent answer made by occupants is 0 (neutral). The
mean value of satisfaction level lies between 0 and +1.

Figure 32 Distribution of Answers to Q8 and Box Plot

Question 9 asks occupants’ satisfaction on alterability of physical condition of the
workstation. As shown in Figure 33, the most frequent answer made by occupants is 0
(neutral). The mean value of satisfaction level lies near 0.
3 2 1 0 -1 -2 -3
120
100
80
60
40
20
0
Q7: Size of Workspace
Count
Chart of Q7
3
2
1
0
-1
-2
-3
Q7
Boxplot of Q7
3 2 1 0 -1 -2 -3
120
100
80
60
40
20
0
Q8: Level of Privacy
Count
Chart of Q8
3
2
1
0
-1
-2
-3
Q8
Boxplot of Q8
91

Figure 33 Distribution of Answers to Q9 and Box Plot

Question 10 asks occupants’ satisfaction on outside view from the workstation. As
shown in Figure 34, the most frequent answer made by occupants is +3 (very satisfied).
The mean value of satisfaction level lies between 0 and +1.

Figure 34 Distribution of Answers to Q10 and Box Plot

Question 11 asks occupants’ satisfaction on distance from other occupants’
workstation. As shown in Figure 35, the most frequent answer made by occupants is 0
(neutral). The mean value of satisfaction level lies near +1.
3 2 1 0 -1 -2 -3
120
100
80
60
40
20
0
Q9: Altering Physical Condition
Count
Chart of Q9
3
2
1
0
-1
-2
-3
Q9
Boxplot of Q9
3 2 1 0 -1 -2 -3
90
80
70
60
50
40
30
20
10
0
Q10: Outside View
Count
Chart of Q10
3
2
1
0
-1
-2
-3
Q10
Boxplot of Q10
92

Figure 35 Distribution of Answers to Q11 and Box Plot

Question 12 asks occupants’ satisfaction on the degree of enclosure of workstation.
As shown in Figure 36, the most frequent answer made by occupants is +1 (slightly
satisfied). The mean value of satisfaction level lies near +1.

Figure 36 Distribution of Answers to Q12 and Box Plot

Question 13 asks occupants’ satisfaction on Temperature at the workstation. As
shown in Figure 37, the most frequent answer made by occupants is +2 (satisfied). The
mean value of satisfaction level lies near +1.

3 2 1 0 -1 -2 -3
100
80
60
40
20
0
Q11 Distance from Others
Count
Chart of Q11
3
2
1
0
-1
-2
-3
Q11
Boxplot of Q11
3 2 1 0 -1 -2 -3
100
80
60
40
20
0
Q12 Degree of Enclosure
Count
Chart of Q12
3
2
1
0
-1
-2
-3
Q12
Boxplot of Q12
93

Figure 37 Distribution of Answers to Q13 and Box Plot

Question 14 asks occupants’ satisfaction on accessibility to the thermostat. As
shown in Figure 38, the most frequent answer made by occupants is 0 (neutral). The mean
value of satisfaction level lies near 0.

Figure 38 Distribution of Answers to Q14 and Box Plot

Question 15 asks occupants’ satisfaction on indoor air quality. As shown in Figure
39, the most frequent answer made by occupants is 0 (neutral). The mean value of
satisfaction level lies between 0 and +1.

3 2 1 0 -1 -2 -3
100
80
60
40
20
0
Q13 Temperature
Count
Chart of Q13
3
2
1
0
-1
-2
-3
Q13
Boxplot of Q13
3 2 1 0 -1 -2 -3
120
100
80
60
40
20
0
Q14 Access to Thermostats
Count
Chart of Q14
3
2
1
0
-1
-2
-3
Q14
Boxplot of Q14
94

Figure 39 Distribution of Answers to Q15 and Box Plot

Question 16 asks occupants’ satisfaction on the level of verbal noise. As shown in
Figure 40, the most frequent answer made by occupants is +1 (slightly satisfied). The mean
value of satisfaction level lies near 0.

Figure 40 Distribution of Answers to Q16 and Box Plot

Question 17 asks occupants’ satisfaction on the level of non-verbal noise. As shown
in Figure 41, the most frequent answer made by occupants is +2 (satisfied). The mean value
of satisfaction level lies between 0 and +1.

3 2 1 0 -1 -2 -3
120
100
80
60
40
20
0
Q15 Indoor Air Quality
Count
Chart of Q15
3
2
1
0
-1
-2
-3
Q15
Boxplot of Q15
3 2 1 0 -1 -2 -3
100
80
60
40
20
0
Q16 Verbal Noise
Count
Chart of Q16
3
2
1
0
-1
-2
-3
Q16
Boxplot of Q16
95

Figure 41 Distribution of Answers to Q17 and Box Plot

Question 18 asks occupants’ satisfaction on light for computer work. As shown in
Figure 42, the most frequent answer made by occupants is +1 (slightly satisfied). The mean
value of satisfaction level lies near +1.

Figure 42 Distribution of Answers to Q18 and Box Plot

Question 19 asks occupants’ satisfaction on reflected light on a computer screen.
As shown in Figure 43, the most frequent answer made by occupants is +2 (satisfied). The
mean value of satisfaction level lies near +1.

3 2 1 0 -1 -2 -3
100
80
60
40
20
0
Q17 Non-verbal Noise
Count
Chart of Q17
3
2
1
0
-1
-2
-3
Q17
Boxplot of Q17
3 2 1 0 -1 -2 -3
100
80
60
40
20
0
Q18 Light for Computer Work
Count
Chart of Q18
3
2
1
0
-1
-2
-3
Q18
Boxplot of Q18
96

Figure 43 Distribution of Answers to Q19 and Box Plot

Question 20 asks occupants’ satisfaction on direct glare from the light fixture. As
shown in Figure 44, the most frequent answer made by occupants is +2 (satisfied). The
mean value of satisfaction level lies near +1.

Figure 44 Distribution of Answers to Q20 and Box Plot

Question 21 asks occupants’ satisfaction on direct glare from daylight. As shown
in Figure17, the most frequent answer made by occupants is +2 (satisfied). The mean value
of satisfaction level lies near +1.

3 2 1 0 -1 -2 -3
120
100
80
60
40
20
0
Q19 Reflected Light in Computer Screen
Count
Chart of Q19
3
2
1
0
-1
-2
-3
Q19
Boxplot of Q19
3 2 1 0 -1 -2 -3
100
80
60
40
20
0
Q20 Direct Glare from Light Fixture
Count
Chart of Q20
3
2
1
0
-1
-2
-3
Q20
Boxplot of Q20
97

Figure 45 Distribution of Answers to Q21 and Box Plot

Question 22 asks occupants’ satisfaction on the quality of overall lighting. As
shown in Figure18, the most frequent answer made by occupants is +2 (satisfied). The
mean value of satisfaction level lies near +1.

Figure 46 Distribution of Answers to Q22 and Box Plot

Question 23 asks occupants’ satisfaction in the building, office, and workstation
layout. As shown in Figure 47, the most frequent answer made by occupants is +1 (slightly
satisfied). The mean value of satisfaction level lies between 0 and +1.

3 2 1 0 -1 -2 -3
120
100
80
60
40
20
0
Q21 Direct Glare from Daylight
Count
Chart of Q21
3
2
1
0
-1
-2
-3
Q21
Boxplot of Q21
3 2 1 0 -1 -2 -3
120
100
80
60
40
20
0
Q22 Quality of Overall Lighting
Count
Chart of Q22
3
2
1
0
-1
-2
-3
Q22
Boxplot of Q22
98

Figure 47 Distribution of Answers to Q23 and Box Plot

Question 24 asks occupants’ satisfaction on color, texture, and finishes. As shown
in Figure 48, the most frequent answer made by occupants is 0 (neutral). The mean value
of satisfaction level lies near +1.

Figure 48 Distribution of Answers to Q24 and Box Plot

Question 25 asks occupants’ satisfaction on air movement. As shown in Figure 49,
the most frequent answer made by occupants is 0 (neutral). The mean value of satisfaction
level lies near 0.

3 2 1 0 -1 -2 -3
90
80
70
60
50
40
30
20
10
0
Q23 Building / Office / Workstation Layout
Count
Chart of Q23
3
2
1
0
-1
-2
-3
Q23
Boxplot of Q23
3 2 1 0 -1 -2 -3
90
80
70
60
50
40
30
20
10
0
Q24 Color / Texture / Finishes
Count
Chart of Q24
3
2
1
0
-1
-2
-3
Q24
Boxplot of Q24
99

Figure 49 Distribution of Answers to Q25 and Box Plot

Question 26 asks occupants’ satisfaction on temperature. As shown in Figure 50,
the most frequent answer made by occupants is 0 (neutral). The mean value of satisfaction
level lies near 0.

Figure 50 Distribution of Answers to Q26 and Box Plot

Question 27 asks occupants’ satisfaction on the lighting condition. As shown in
Figure 51, the most frequent answer made by occupants is 0 (neutral). The mean value of
satisfaction level lies near 0.

3 2 1 0 -1 -2 -3
120
100
80
60
40
20
0
Q25 Air Movement
Count
Chart of Q25
3
2
1
0
-1
-2
-3
Q25
Boxplot of Q25
3 2 1 0 -1 -2 -3
160
140
120
100
80
60
40
20
0
Q26 Temperature
Count
Chart of Q26
3
2
1
0
-1
-2
-3
Q26
Boxplot of Q26
100

Figure 51 Distribution of Answers to Q27 and Box Plot

Question 28 asks occupants’ satisfaction on acoustic condition. As shown in Figure
52, the most frequent answer made by occupants is +1 (slightly satisfied). The mean value
of satisfaction level lies between 0 and +1.

Figure 52 Distribution of Answers to Q28 and Box Plot

Question 29 asks occupants’ satisfaction on the overall environment. As shown in
Figure 53, the most frequent answer made by occupants is +1 (slightly satisfied). The mean
value of satisfaction level lies near +1.

3 2 1 0 -1 -2 -3
120
100
80
60
40
20
0
Q27 Lighting Condition
Count
Chart of Q27
3
2
1
0
-1
-2
-3
Q27
Boxplot of Q27
3 2 1 0 -1 -2 -3
120
100
80
60
40
20
0
Q28 Acoustic Condition
Count
Chart of Q28
3
2
1
0
-1
-2
-3
Q28
Boxplot of Q28
101

Figure 53 Distribution of Answers to Q29 and Box Plot

3 2 1 0 -1 -2 -3
140
120
100
80
60
40
20
0
Q29 Overall Environment
Count
Chart of Q29
3
2
1
0
-1
-2
-3
Q29
Boxplot of Q29
102

Bibliography
[1] Fisk, William J. 2002. How IEQ affects health, productivity. ASHRAE Journal 44,
(5): 56

[2] Kosonen, Risto, Mervi Ahola, Kirsi Villberg, and Tarja Takki. 2011. Perceived IEQ
conditions: Why the actual percentage of dissatisfied persons is higher than
standards indicate? In , 75-88. Berlin, Heidelberg: Springer Berlin Heidelberg,

[3] Omer, Abdeen Mustafa. 2008. Energy, environment and sustainable development.
Renewable and Sustainable Energy Reviews 12, (9): 2265-2300

[4] Hern, Stephen C., Andy M. Tsang, John P. Robinson, Klepeis, Neil E.
Nelson,William C., Paul Switzer, William H. Engelmann, Wayne R. Ott, and Joseph
V. Behar. 2001. The national human activity pattern survey (NHAPS): A resource
for assessing exposure to environmental pollutants. Journal of exposure analysis and
environmental epidemiology 11, (3): 231-252

[5] "INDOOR ENVIRONMENTAL QUALITY." Centers for Disease Control and
Prevention. August 19, 2015. Accessed December 16, 2015.
http://www.cdc.gov/niosh/topics/indoorenv/.

[6] Candido, Christhina, Richard de Dear, Leena Thomas, Jungsoo Kim, and Thomas
Parkinson. "Introducing BOSSA."

[7] Preiser, Wolfgang FE. "Post-occupancy evaluation: how to make buildings work
better." Facilities 13, no. 11 (1995): 19-28.

[8] Meir, Isaac A., Yaakov Garb, Dixin Jiao, and Alex Cicelsky. "Post-occupancy
evaluation: an inevitable step toward sustainability." Advances in building energy
research 3, no. 1 (2009): 189-219.

[9] Beutel, Lauren, and Andre Dalton. "POST-OCCUPANCY EVALUATION: A case
study of a public space." Australian Planner 38, no. 3-4 (2001): 178-181.

[10] "Office Building." Office Building. Accessed December 24, 2015.
https://www.wbdg.org/design/office.php.

[11] Abbaszadeh, S., Leah Zagreus, D. Lehrer, and C. Huizenga. "Occupant satisfaction
with indoor environmental quality in green buildings." Center for the built
environment (2006).

[12] "Indoor Air Quality (IAQ)." EPA. Accessed November 2, 2015.
http://www.epa.gov/iaq.

103

[13] "BRE Group: Post-Occupancy Evaluation (POE)." BRE Group: Post-Occupancy
Evaluation (POE). Accessed November 2, 2015.
https://www.bre.co.uk/page.jsp?id=1793.

[14] "Lighting and Daylighting Design | Sustainability Workshop." Lighting and
Daylighting Design | Sustainability Workshop. Accessed December 18, 2015.
http://sustainabilityworkshop.autodesk.com/buildings/lighting-and-daylighting-
design#sthash.HpTJvdqc.dpuf.

[15] "Why Join ASHRAE." Standards & Guidelines. Accessed November 2, 2015.
http://www.ashrae.org/technology/page/548.

[16] "Thermal Comfort." Occupant Comfort. Accessed December 13, 2015.
http://sustainabilityworkshop.autodesk.com/buildings/occupant-comfort.

[17] Watson, Chris. 2003. Review of building quality using post occupancy evaluation.
PEB Exchange 2003 (48): 15.

[18] Hadjri, Karim, and Carl Crozier. "Post-occupancy evaluation: purpose, benefits and
barriers." Facilities 27, no. 1/2 (2009): 21-33.

[19] Liang, HH, CP Chen, RL Hwang, WM Shih, SC Lo, and HY Liao. 2014.
Satisfaction of occupants toward indoor environment quality of certified green office
buildings in taiwan. Building and Environment 72 : 232-42.

[20] Morhayim, Lusi, and Isaac Meir. "Survey of an office and laboratory university
building: An unhealthy building case study." In Proceedings of Indoor Air 2008
Conference, Copenhagen, Denmark, pp. 17-22. 2008.

[21] Meir, I. A., W. Motzafi-Haller, E. Kruger, L. Morhayim, S. Fundaminsky, and L.
Oshry-Frenkel. "Towards a comprehensive methodology for Post Occupancy
Evaluation (POE): A hot dry climate case study." In keynote presentation) in M.
Santamouris and P. Wouters (eds), Building Low Energy Cooling and Advanced
Ventilation in the 21st Century, Proceedings of the 2nd PALENC and 28th AIVC
Conference, Crete, II, pp. 27-29. 2007.

[22] Kansara, Tia. "Post Occupancy Monitoring: An introduction." Environmental
Management and Sustainable Development 2, no. 2 (2013): 86-100.

[23] Choi, Joon-Ho, Vivian Loftness, and Azizan Aziz. 2012; 2011. Post-occupancy
evaluation of 20 office buildings as basis for future IEQ standards and guidelines.
Energy & Buildings 46, : 167-175

[24] Zagreus, Leah, Charlie Huizenga, Edward Arens, and David Lehrer. 2004. Listening
to the occupants: A Web ‐ based indoor environmental quality survey. Indoor Air 14
(s8): 65-74.
104

[25] Frontczak, Monika, Rune Vinther Andersen, and Pawel Wargocki. "Questionnaire
survey on factors influencing comfort with indoor environmental quality in Danish
housing." Building and Environment 50 (2012): 56-64.

[26] Li, Baizhan, and David Lim. "Occupant behavior and building performance." In
Design and Management of Sustainable Built Environments, pp. 279-304. Springer
London, 2013.

[27] Abbaszadeh, S., L. Zagreus, D. Lehrer, and C. Huizenga. 2006. Occupant
satisfaction with indoor environmental quality in green buildings. In . Vol. 3, 365-
370,

[28] Pfafferott, Jens Ü ., Sebastian Herkel, Doreen E. Kalz, and Andreas Zeuschner. 2007.
Comparison of low-energy office buildings in summer using different thermal
comfort criteria. Energy & Buildings 39 (7): 750-7.

[29] Langston, Craig, Yu Song, and Brian Purdey. 2008. Perceived conditions of workers
in different organizational settings. Facilities 26 (1/2): 54-67.

[30] Wagner, A., E. Gossauer, C. Moosmann, Th Gropp, and R. Leonhart. 2007. Thermal
comfort and workplace occupant satisfaction—Results of field studies in german low
energy office buildings. Energy & Buildings 39 (7): 758-69.

[31] "Measuring Light Levels | Sustainability Workshop." Measuring Light Levels |
Sustainability Workshop. Accessed December 18, 2015.
http://sustainabilityworkshop.autodesk.com/buildings/measuring-light-levels.

[32] del Mar Castilla, Maŕia, José Domingo Álvarez, Francisco Rodríguez, and Manuel
Berenguel. Comfort control in buildings. Springer-Verlag, 2014.

[33] Sharma, Rajesh, and D. G. M. Prouhit. "Energy analysis for building designers
through computer software." International Journal of Innovative Science,
Engineering & Technology 1, no. 6 (2014).

[34] Goldman, R.F. and Kampmann, B., 2007. Handbook of Clothing. Biomedical
Effects of Military Clothing and Equipment Systems, Report NATO Research Study
Group, 7.

[35] "Human Thermal Comfort | Sustainability Workshop." Human Thermal Comfort |
Sustainability Workshop. Accessed April 11, 2016.
http://sustainabilityworkshop.autodesk.com/buildings/human-thermal-
comfort#sthash.IDqG5izE.dpuf.

[36] "Indoor Air Quality | Sustainability Workshop." Indoor Air Quality | Sustainability
Workshop. Accessed December 18, 2015.
http://sustainabilityworkshop.autodesk.com/buildings/indoor-air-quality.
105

[37] Fincher, Warren, and Michael Boduch. "Standards of human comfort: relative and
absolute." 2009.

[38] "UNITED STATES DEPARTMENT OF LABOR." Safety and Health Topics.
Accessed April 01, 2016. https://www.osha.gov/SLTC/noisehearingconservation/.

[39] GSA Public Building Service. 2011 Sound Matters: How to achieve acoustic
comfort in the contemporary office.

[40] ASHRAE, ANSI. "ASHRAE 129-1997:(1997)“Measuring air change effectiveness”,
American Society of Heating." Refrigeration, and Air Conditioning Engineers, Inc.,
Atlanta, GA.

[41] Newsham, Guy R. Making the open-plan office a better place to work. Institute for
Research in Construction, National Research Council of Canada, 2003.

[42] "What Is a Likert Scale and How Do You Pronounce Likert?" What Is a Likert Scale
and How Do You Pronounce Likert? Accessed August 24, 2015.
http://core.ecu.edu/psyc/wuenschk/StatHelp/Likert.htm.

[43] Illuminating Engineering Society of North America. Lighting Handbook: Reference
& application. Illuminating Engineering, 2000.

[44] ASHRAE, A., 2010. Standard 55-2010:“Thermal Environmental Conditions for
Human Occupancy”; ASHRAE. Atlanta USA.

[45] ASHRAE, ASHRAE. "ASHRAE 62.1-2010 Ventilation for Acceptable Indoor Air
Quality." Atlanta, GA (2002).

[46] "Stepwise Regression and All-possible-regressions." Stepwise Regression and All-
possible-regressions. Accessed December 24, 2015.
http://people.duke.edu/~rnau/regstep.htm.

[47] "Regression Analysis: How Do I Interpret R-squared and Assess the Goodness-of-
Fit?" Regression Analysis: How Do I Interpret R-squared and Assess the Goodness-
of-Fit? Accessed December 18, 2015.
http://blog.minitab.com/blog/adventures-in-statistics/regression-analysis-how-do-i-
interpret-r-squared-and-assess-the-goodness-of-fit.

[48] "Introduction to Correlation and Regression Analysis." Introduction to Correlation
and Regression Analysis. Accessed December 24, 2015.
http://sphweb.bumc.bu.edu/otlt/MPH-
Modules/BS/BS704_Multivariable/BS704_Multivariable5.html.

106

[49] "Basics of Correlation and Covariance." - Minitab. Accessed December 24, 2015.
http://support.minitab.com/en-us/minitab/17/topic-library/modeling-
statistics/regression-and-correlation/correlation-and-covariance/basics-of-
correlation-and-covariance/.

[50] "Why Should I Use a 2-sample T Test?" - Minitab. Accessed December 24, 2015.
http://support.minitab.com/en-us/minitab/17/topic-library/basic-statistics-and-
graphs/hypothesis-tests/tests-of-means/why-use-2-sample-t/.

[51] "What Is ANOVA?" - Minitab. Accessed December 24, 2015.
http://support.minitab.com/en-us/minitab/17/topic-library/modeling-
statistics/anova/basics/what-is-anova/.

[52] "One-way ANOVA." - An Introduction to When You Should Run This Test and the
Test Hypothesis. Accessed December 24, 2015.
https://statistics.laerd.com/statistical-guides/one-way-anova-statistical-guide.php.

[53] "Machine Learning Group." Machine Learning Project at the University of Waikato
in New Zealand. Accessed February 20, 2016.
http://www.cs.waikato.ac.nz/ml/index.html.

[54] "What Is Mallows' Cp?" - Minitab. Accessed December 18, 2015.
http://support.minitab.com/en-us/minitab/17/topic-library/modeling-
statistics/regression-and-correlation/goodness-of-fit-statistics/what-is-mallows-cp/.

Abstract (if available)

Abstract Rapidly developing technologies and growing demands of people for a better quality of life have enhanced the collaboration between design research and practice. In this context, Post Occupancy Evaluation (POE) and its research are one of the most well-developed examples of the research-practice feedback system. However, the weakness of the POE is often pointed out its excessive reliance on the survey and general-targeted solution. This weakness may lead to irrelevant modification and occupants’ dissatisfaction with the indoor environmental quality (IEQ). As a solution, comprehensive POE is suggested. Quantifiable data such as IEQ measurements and occupants’ satisfaction is collected from over 400 workstations in modern buildings of Southern California through on-site measurement and satisfaction surveys on the occupants. Statistical analyses are conducted on collected data within specified categories of building types, building attributes, and human factors. The analysis not only correlates the influence on one another but also allows sorting the factors by the intensity of the impact. The results verified the necessity of the modified POE regarding practice-based approach and specific-targeted solution. Based on findings, specific recommendations and strategic design guidelines are suggested to help enhance the environmental conditions in working environment.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Impacts of building performance on occupants' work productivity: a post occupancy evaluation study

PDF

Impacts of indoor environmental quality on occupants environmental comfort: a post occupancy evaluation study

PDF

Enhanced post occupancy evaluation (POE) for office building: improvement of current methodology to identify impact of ambient environment

PDF

Nursing comfort study: are hospitals performing for nurses? IEQ analysis of working conditions inside hospitals

PDF

Indoor environmental quality and comfort: IEQ adaptation and human physiological responses in commercial buildings

PDF

IEQ, sleep quality, and IoT: meta-analysis on improving IEQ and sleep quality using IoT

PDF

Human-building integration based on biometric signal analysis: investigation of the relationships between human comfort and IEQ in a multi-occupancy condition

PDF

Impact of bamboo materials on indoor air quality

PDF

Human-environmental interaction: potential use of pupil size for office lighting controls

PDF

Human-building integration: Investigation of human eye pupil sizes as a measure of visual sensation in the workstation environment

PDF

Human–building integration: machine learning–based and occupant eye pupil size–driven lighting control as an applicable visual comfort tool in the office environment

PDF

Multi-occupancy environmental control for smart connected communities

PDF

Indoor air quality for human health in residential buildings

PDF

Post occupancy energy evaluation of Ronald Tutor Hall using eQUEST; computer based simulation of existing building and comparison of data

PDF

Daylighting study of a LEED platinum laboratory building: a post-occupancy evaluation comparing performance in use to design intent

PDF

Impact of occupants in building performance: extracting information from building data

PDF

Exploration for the prediction of thermal comfort & sensation with application of building HVAC automation

PDF

Energy use intensity estimation method based on building façade features by using regression models

PDF

Developing environmental controls using a data-driven approach for enhancing environmental comfort and energy performance

PDF

Development of data-driven user-centered building façade design guideline models: machine learning-based approaches to predict user preferences

Asset Metadata

Creator Moon, Jehyun (author)

Core Title Considering occupants: comprehensive POE research on office environment of Southern California

School School of Architecture

Degree Master of Building Science

Degree Program Building Science

Publication Date 06/22/2018

Defense Date 04/29/2016

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag human comfort,indoor environmental quality,OAI-PMH Harvest,office workstation,post occupancy evaluation,user satisfaction

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Choi, Joon-Ho (committee chair), Mutlow, John V. (committee member), Schiler, Marc (committee member)

Creator Email jaymoonla@gmail.com,jehyunmo@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-257440

Unique identifier UC11279441

Identifier etd-MoonJehyun-4469.pdf (filename),usctheses-c40-257440 (legacy record id)

Legacy Identifier etd-MoonJehyun-4469.pdf

Dmrecord 257440

Document Type Thesis

Format application/pdf (imt)

Rights Moon, Jehyun

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA