University of Southern California Dissertations and Theses

Enabling human-building communication to promote pro-environmental behavior in office buildings

(USC Thesis Other)

Enabling human-building communication to promote pro-environmental behavior in office buildings

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
Enabling Human-Building Communication to Promote
Pro-Environmental Behavior in Office Buildings

Saba Khashe
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
Submitted in Partial Fulfillment of the Requirements for the Degree
DOCTOR OF PHILOSOPHY
(CIVIL ENGINEERING)

Guidance Committee Members:
Professor Burcin Becerik-Gerber (Chair), Professor Lucio Soibelman,
Professor Jonathan Gratch, and Dr. Gale Lucas
May 2018
`
1

Contents
Acknowledgement .......................................................................................................................... 8
Executive Summary ...................................................................................................................... 10
Chapter 1: Introduction ................................................................................................................. 14
1.1. Problem Statement and Motivation .................................................................................... 14
1.2. Scope .................................................................................................................................. 17
Chapter 2. Literature Review ........................................................................................................ 18
2.1. Green Branding Intervention Strategy ............................................................................... 20
2.2. Communication and Compliance Gaining ......................................................................... 23
2.2.1. Intended Behaviors: Areas of Wasteful Behaviors in Buildings ................................. 23
2.2.2. Messaging Strategies ................................................................................................... 26
2.2.3. Communication Modality or Delivery Style ............................................................... 29
2.2.4. Communicator Persona ................................................................................................ 32
2.2.5. Social Dialog ............................................................................................................... 34
2.2.6. Users’ Characteristics .................................................................................................. 36
2.3. Frequency of Interactions ................................................................................................... 38
Chapter 3. Research Objectives and Questions ............................................................................ 40
Chapter 4. Research Tools ............................................................................................................ 44
4.1. Virtual Environments ......................................................................................................... 44
4.2. Modeling (Creation of Virtual Environment) .................................................................... 45
4.3. Benefits of the Virtual Environments................................................................................. 47
4.4. Virtual Environments vs. Physical Environments.............................................................. 48
4.5. Virtual Environment Platforms .......................................................................................... 49
4.5.1. PC laptop vs. HMD ..................................................................................................... 50
4.5.2. Research Methodology ................................................................................................ 51
4.5.4. Results ......................................................................................................................... 59
4.5.5. Discussions and Conclusions....................................................................................... 67
Chapter 5. Influence of LEED Branding on Building Occupants’ Pro-environmental Behavior 70
5.1. Methodology ...................................................................................................................... 72
5.2. Experiment Setup ............................................................................................................... 73
5.2.1. Pre-Experiment Session ............................................................................................... 75
`
2

5.2.2. Experiment Session ..................................................................................................... 76
5.2.3. Post-Experiment Session ............................................................................................. 79
5.3. Results and Data Analysis .................................................................................................. 81
5.4. Discussions ......................................................................................................................... 88
5.5. Conclusions ........................................................................................................................ 89
Chapter 6. Exploring the Effects of Social Influence Methods .................................................... 91
6.1. Methodology ...................................................................................................................... 92
6.2. Experiment Setup ............................................................................................................... 93
6.2.1. Pre-Experiment Session ............................................................................................... 94
6.2.2. Experiment Session ..................................................................................................... 95
6.2.3. Carry over effect .......................................................................................................... 98
6.2.4. Post-Surveys ................................................................................................................ 99
6.3. Results, and Discussion .................................................................................................... 100
6.4. Conclusions ...................................................................................................................... 108
Chapter 7. Exploring the Effects of Communication Delivery Style and Communicator's Persona
..................................................................................................................................................... 110
7.1. Research Methodology ..................................................................................................... 111
7.2. Experiment Setup ............................................................................................................. 112
7.2.1. Pre-Experiment Session ............................................................................................. 115
7.2.2. Experiment Session ................................................................................................... 116
7.2.3. Post-Experiment Session ........................................................................................... 119
7.3. Results .............................................................................................................................. 120
7.3.1. Influence of manipulated variables on compliance ................................................... 121
7.3.2. Influence of individual difference variables on compliance ..................................... 126
7.4. Discussion ........................................................................................................................ 131
7.5. Conclusion ........................................................................................................................ 133
Chapter 8. Exploring the Effects of Social Dialog and Communicator's Persona ...................... 135
8.1. Research Methodology ..................................................................................................... 136
8.2. Study 1a ............................................................................................................................ 137
8.2.1. Pre-Experiment Session ............................................................................................. 138
8.2.2. Experiment Session ................................................................................................... 139
`
3

8.2.3. Post-Experiment Session ........................................................................................... 142
8.2.4. Results and Discussion .............................................................................................. 143
8.3. Study 1b............................................................................................................................ 149
8.2.1. Pre-Experiment Session ............................................................................................. 150
8.2.2. Experiment Session ................................................................................................... 150
8.2.3. Post Experiment Session ........................................................................................... 152
8.2.4. Results ....................................................................................................................... 152
8.4. Conclusions ...................................................................................................................... 153
Chapter 9. Limitations and Potential Future Work ..................................................................... 155
Chapter 10. Conclusions ............................................................................................................. 160
References ................................................................................................................................... 163
Appendixes ................................................................................................................................. 194
Appendix 1: Green buildings and LEED certification ............................................................ 194
Appendix 2: Allo-inclusive identity scale ............................................................................... 196

`
4

Table of Figures
Figure 1 – Different views of the model: (a) Top view; (b) perspective view ............................. 45
Figure 2 – Modeled environment in 3ds Max
©
............................................................................. 46
Figure 3 – Modelled office – illustrating what participants see on the PC laptop ........................ 46
Figure 4 – Modeled office - illustrating what participants see through Oculus ............................ 47
Figure 5 – Different VE platforms: (a) laptop display; (b) HMD................................................. 52
Figure 6 – Pro-environmental requests delivered to the participant in the VE: (a) before the
request was delivered to the participant; (b) while the request was being delivered to the
participant ..................................................................................................................................... 53
Figure 7 – Different lighting levels in the room: (a) All the lights on while the blind was closed;
(b) dimming the light one level while the blind was open; (c) dimming the light two levels while
the blind was open; (d) turning off all the lights while the blind .................................................. 56
Figure 8 – Different temperature settings in the room: (a) 73 degree Fahrenheit; (b) 74 degree
Fahrenheit; (c) 75 degree Fahrenheit; (d) 76 degree Fahrenheit ................................................... 57
Figure 9 – Training process: a participant is trained how to navigate and interact within the IVE.
The left image shows the participant navigating through the room, and the right image shows the
participant opening and closing the oven door ............................................................................. 76
Figure 10 – Different lighting levels in the room: (a) dark room; (b) bright room with all light
bulbs on and the blind closed; (c) bright room with the blind open and all of the light bulbs off;
(d) bright room with the blind open and all of the light bulbs on ................................................. 77
Figure 11 – (a) turn on/off the lights; and (b) open/close the blind .............................................. 78
Figure 12 – (a) experimental components: passage on computer monitor; (b) scrap papers on the
table ............................................................................................................................................... 79
Figure 13 – Recycling bin and regular trashcan in the experimental setting ................................ 79
Figure 14 – A participant during training ..................................................................................... 95
Figure 15 – Different lighting levels in the room: (a) all light bulbs on and all blinds closed; (b)
all light bulbs on and the left blind open; (c) all light bulbs on and the middle blind open; (d) all
light bulbs on and the right blind open; (e) all light bulbs ............................................................ 96
Figure 16 – How request was delivered to the participants in the immersive virtual environment:
(a) a participant watching a video in the immersive virtual environment; (b) Request delivered to
the participant while watching the video ...................................................................................... 97
`
5

Figure 17 – How request was delivered to the participants in the physical office environment: (a)
a participant completing surveys in the physical office; (b) request delivered to the participant
while completing the survey ......................................................................................................... 99
Figure 18 – Participants’ intentions to use a similar suggestion system if it was employed in their
building in the future ................................................................................................................... 106
Figure 19 – Participants intentions to respond in a similar fashion to the buildings’ request if it
was employed in their own office ............................................................................................... 107
Figure 20 – Training process: a participant is trained how to navigate and interact within the
IVE. Figure also illustrates what participants see through Oculus DK2 Head-Mounted Display
..................................................................................................................................................... 116
Figure 21 – Requests: (a) female text; (b) male text; (c) female voice; (d) male voice; (e) female
avatar; (f) male avatar ................................................................................................................. 118
Figure 22 – Participants' compliance distribution based on communicator's gender ................. 123
Figure 23 – Participants rating for communicator positivity ...................................................... 125
Figure 24 – Virtual office space embedded in the on-line survey .............................................. 138
Figure 25 – Pro-environmental requests delivered to the participant in the VE: (a) before the
request was delivered to the participant; (b) while the request was being delivered to the
participant (right) ........................................................................................................................ 140
Figure 26 – Compliance (count) for independent variables, persona and social dialog ............. 145
Figure 27 – Pro-environmental requests delivered to the participant in the physical environment:
(a) before the request was delivered to the participant; (b) while the request was being delivered
to the participant ......................................................................................................................... 151

`
6

Table of Tables
Table 1- Main effects of VE platforms on presence ..................................................................... 61
Table 2 – Simulator sickness effects before and after the experiment ......................................... 62
Table 3 – Correlation between presence sub-scales and SSQ subscales ...................................... 63
Table 4 – Difference in participants’ immersive tendency subscales ........................................... 64
Table 5 – Correlation between ITQ subscales and presence subscales ........................................ 65
Table 6 – Gender main effects and interaction effects with VE platforms on compliance,
presence, and performance............................................................................................................ 66
Table 7 – Sample questions about environmental responsibility.................................................. 80
Table 8 - IVE interaction questions .............................................................................................. 82
Table 9 - Branding group vs. control group .................................................................................. 83
Table 10 – Branding group vs. control group ( 
2
analysis - natural vs. artificial light) ................ 84
Table 11 – Branding group vs. control group ............................................................................... 85
Table 12 – Requests delivered to participants in different groups................................................ 92
Table 13 – Direct request vs. reciprocal requests vs. FITD in virtual environment ................... 102
Table 14 – Persona – delivery style - request combinations that the participants will be assigned
to ................................................................................................................................................. 114
Table 15 - Main effects of delivery style and persona on compliance ....................................... 121
Table 16 - Main and interaction effects of delivery style, persona, and order of communicator
gender on compliance ................................................................................................................. 122
Table 17 – Effect of communicator gender on compliance ........................................................ 123
Table 18 – Main and interaction effects of communicator gender on compliance ..................... 124
Table 19 – Main effects of demographic variables on compliance ............................................ 126
Table 20 – Interaction effects of demographic variables on compliance ................................... 127
Table 21 – Main and interaction effects of Technology Readiness and Allo-inclusive Identity on
compliance .................................................................................................................................. 127
Table 22 – Main and interaction effects of personality variables on compliance ...................... 129
Table 23 – Main effects of delivery style, persona, and communicator gender on how
affectionate, friendly, and likable the communicator seemed .................................................... 130
Table 24 - Effects of demographic variables on effectiveness of persona and delivery style .... 146
Table 25 – Effects of personality traits on effectiveness of persona and delivery style ............. 147
`
7

Table 26 – Effects of technology readiness on effectiveness of persona and delivery style ...... 149

`
8

Acknowledgement
Dedicated to my wonderful parents, sister, and husband, for supporting me through this journey.
First and foremost, I want to thank my advisor Dr. Burcin Becerik. It has been an honor to work
with her. I truly appreciate all the time, advice, and support she gave me throughout my Ph.D. The
joy and enthusiasm she has for her research was contagious and motivational for me. I am also
thankful for the excellent example she has provided as a successful woman engineer and professor.
I gratefully acknowledge the members of my Ph.D. committee, Dr. Lucio Soibelman, Dr. Jonathan
Gratch, and Dr. Gale Lucas. I appreciated your guidance, support and willingness to take time to
discuss my research. Again, thank you for your timely advice throughout my graduate studies.
I am very thankful to Dr. Najmedin Meshkati who encouraged me to start this journey and
supported me during my studies. I would also like to thank Dr. Wendy Wood, Dr. David Gerber
and Dr. Tim Hayes for their guidance during the first semesters of my Ph.D.
Thanks to the entire Civil and Environmental Engineering Department at University of Southern
California, especially the iLab team. All of the students, faculty and staff that I worked with, took
classes from, and met throughout my graduate studies were wonderful people and a real pleasure
to know.
I gratefully acknowledge the support by the National Science Foundation under the Grant numbers
1231001 & 1548517. Any opinions, findings, and conclusions or recommendations expressed in
this document are my views and do not necessarily reflect the views of the National Science
Foundation.
`
9

Lastly, I would like to thank my family for all their love and encouragement. Mom and Dad, I
cannot find a way to thank you enough and express my feelings for you. I am who I am today
because of you. You are my inspirations, role models, best friends, and so much more. Yalda, my
lovely sister, I could not ask for a better sister and friend, thanks for always being there for me.
And I owe thanks to a very special person, my husband Siavash, if I wrote down everything I ever
wanted in a husband and best friend I would not have believed I could meet someone better!
Thanks for your continued support during this journey that made the completion of this dissertation
possible. I deeply appreciate your belief in me. I would also like to thank my uncle, Alireza, who
supported me during these years. I am so lucky to have you all in my life.

`
10

Executive Summary
Americans spend 90 percent of their time inside buildings, working, living, shopping, studying,
and performing other activities, which cause interactions with building systems. These human-
building interactions influence building performance, which impacts the occupants’ life quality.
The quality of life could be enhanced by promoting occupants’ productivity, learning, comfort,
satisfaction, and health, which can be affected by design, energy efficiency, safety and security in
buildings. There are technological approaches that try to improve the performance of building
systems (e.g., the use of wireless sensor networks to enhance energy efficiency of heating,
ventilating, and air-conditioning (HVAC) systems). However, studies showed that occupants and
their behaviors have significant impacts on a building’s performance and can even override these
approaches. Therefore, these technological improvements are only successful when implemented
and conformed correctly by building occupants.
In this dissertation, we investigated behavior change intervention strategies that influence the way
occupants behave in buildings. This behavior change can support different goals (e.g., improving
energy efficiency, health, and comfort in different types of built environments). However, we
specifically focus on pro-environmental behavior in commercial buildings
1
. Commercial buildings
consume half of the total building energy use and considerably contribute to the issue of climate
change. In addition, the occupants’ lack of direct financial incentives for energy conservation
forms interesting research challenges. Among several commercial building types, we specifically

1
Pro-environmental behaviors refer to occupants’ activities such as reducing energy consumption and
recycling which reduce negative environmental impacts.
`
11

focus on office buildings, which have the largest floor space and energy consumption share of
commercial buildings and are responsible for the largest amount of CO2 emissions.
We designed indirect and direct
2
behavior change strategies to promote building occupants’ pro-
environmental behaviors. First, we explored the effects of an indirect behavior change strategy
(i.e., Leadership in Energy and Environmental Design (LEED)
3
branding) on occupants’ pro-
environmental behaviors in buildings. LEED branding, in the context of this dissertation, is defined
as introducing the building as being LEED certified and emphasizing the aspects of the LEED
building’s green features to its users. This strategy took advantage of functional attributes and
emotional benefits of LEED certified buildings to influence occupants’ behaviors. We also
investigated the effects of a direct behavior change strategy, which is effective for both green and
conventional buildings. We explored approaches to transform buildings into interactive living
spaces that communicate with their occupants and create conditions for behavior change.
Communication intervention strategies, which provide people with relevant information and
motivation through well-defined strategies, appropriate methods, tools and channels, have been
used in different domains to encourage behavior change. We investigated the effects of
incorporating social and relational features into the design of different components of a
communication system.

2
Indirect strategies influence behavior by changing its underlying determinants (e.g., user’s attitude) and
direct strategies directly ask users to change their behaviors.

3
Leadership in Energy and Environmental Design (LEED) is a green building rating system used
increasingly in the United States. LEED was developed by the United States Green Building Council
(USGBC) to provide nationally accepted third-party certification that a building is designed and built using
an approach that reduces or eliminates negative environmental impacts [224].
`
12

First, we examined the effects of adapting behavior change tactics more commonly seen in face-
to-face communication (e.g., direct requests and compliance-gaining techniques) into human-
building communication context. Compliance gaining refers to interactions, in which one
individual (the agent) attempts to induce another person (the target) to perform a desired behavior
that the target person otherwise might not have performed. Second, we investigated the impacts of
incorporating social surface cues (such as face and voice) to the delivery of messages on behavior
change. In addition, we explored the persona (the character of the communicator) that fitted the
purpose of the human-building communication. A communicator’s persona should be based on its
relationship with the user, its capabilities and functions, and the role of the communicator in the
system. Third, we explored the effects of small talk as a common example of social dialog on
facilitating behavior change. Considering that people apply social rules to social and relational
agents and react differently to them based on their own characteristics, we also investigated the
effects of user characteristics on the intervention outcomes. Finally, considering that building-
occupant communication requires continuous interactions between buildings and their occupants,
we also investigated the effects of the requests over more interactions to examine if repetition
impacts the results.
For our investigations, we used a single occupancy office, which was either a physical or a virtual
environment (VE). We tested the differences between participants’ behaviors in the physical and
virtual environments to ensure virtual environments were good venues for our investigations. In
the LEED branding study, participants were given informational text about green buildings and
the goals and requirements of the LEED certification process at the beginning of the study. In the
studies investigating the effects of direct communication, participants were exposed to different
pro-environmental requests while they were performing some office related activities such as
`
13

reading a passage. Then, we observed participants’ compliance with the requests and sought the
reason behind their behaviors.
The results of our studies showed that LEED branding motivated occupants’ pro-environmental
behaviors in buildings. In addition, the incorporation of social and relational features into the
design of intervention strategies aiming to promote pro-environmental behaviors in buildings
increased the positive effects of the interventions. Reciprocal requests, inclusion of face followed
by voice, human-like persona, and social dialog engagement increased compliance with the
requests. However, the effects of LEED branding and social and relational features were not
equally effective across all types of people and their effects varied for people with different
characteristics. For example, participants with higher environmental values and views
4
were more
likely to adopt pro-environmental behaviors under the influence of LEED branding. In addition,
extroverts complied more with the pro-environmental requests when engaged in a social dialog by
a human-like persona agent. Our results also showed that more interaction with the users can
improve the outcomes of the interventions.
Our findings emphasize the importance of occupants’ perception of buildings. These results can
be used to create an interactive environment, which influences occupants’ relationships with the
building. Our findings can also provide useful design choices for persuasive technologies aiming
to promote pro-environmental behaviors.

4
Environmental values and views refer to people’s pro-environmental concerns, awareness and knowledge
about environment related issues.
`
14

Chapter 1: Introduction
1.1. Problem Statement and Motivation
Americans spend 90 percent of their time inside buildings [1], working, living, shopping,
studying, and performing other activities, which cause interaction with building systems. Buildings
interact with occupants by providing space and services to them and occupants interact with
building systems by overriding the systems settings. These human–building interactions influence
building performance as well as the occupants’ quality of life. In fact, the mutual goal of a building
and its occupants is to enhance the quality of the built environment in order to improve occupants’
living experiences. The quality of life could be enhanced by promoting occupants’ productivity,
comfort, and health, which can be impacted by design, energy efficiency, safety and security in
buildings.
Currently, there are technological approaches that try to improve the performance of building
systems (e.g., the use of wireless sensor networks to enhance energy efficiency of heating,
ventilating, and air-conditioning (HVAC) systems [2]). However, these technologies are only
successful when implemented correctly by occupants [3]. In fact, occupant behavior is the major
contributing factor to the uncertainty of building performance [4]. Studies show that buildings with
sustainable technologies (e.g., green buildings) could consume more energy in comparison to
similar conventional buildings due to their occupants’ behaviors [5,6]. This underperformance is
attributed to the differences in occupancy patterns and occupants' characteristics, which are defined
as the presence of people in a building and their actions [7-11]. How occupants interact with
building systems and respond to environmental discomfort directly affect the operation of
buildings. Therefore, to improve the quality of the built environment, behavioral approaches as
well as technological approaches are required to modify occupant behavior.
`
15

Behavioral approaches focus on modifying occupant behavior through intervention strategies. For
example, providing feedback to occupants on their energy usage encourages them to further reduce
their energy consumption [12]. However, existing intervention strategies adopted in buildings were
not successful to fully influence occupant behaviors. There is a need to investigate the potentials
of improving the effectiveness and persuasiveness of these strategies. Information sharing and
increasing awareness may modify occupant behaviors by increasing people's awareness of their
behavioral impacts and knowledge about possibilities to mediate these impacts. Communication
is also an important factor in persuasion and lifestyle changes [13,14].
Communication intervention strategies, which provide people with relevant information and
motivation through well-defined strategies; appropriate methods, tools and channels have been
used in different domains to encourage behavior change. Communication strategies can support a
variety of functions in buildings (e.g., learning, health safety, and security). However, we focus on
human–building interactions that contribute to environmental impacts, which is one of the major
concerns of our time.
Climate change issue is among the greatest threats to the world and residential and commercial
buildings significantly contribute to the issue through greenhouse gas emissions [15]. Energy
consumption is the main cause of greenhouse gas emissions [16]. In the developed countries, the
building sector accounts for more energy consumption than the transportation and industry sectors,
including manufacturing, agriculture, and mining [17]. In the U.S., residential and commercial
buildings steadily increased their share reaching 39% of the total energy consumption in 2016 [18].
Growth in the population, increasing energy demands of building services, life style changes and
improvements together with the rise in time spent in buildings assure the upward trend in energy
demand will continue in the future. Based on the current trends, it is predicted that between the
`
16

years of 2013 and 2040, electricity consumption in commercial buildings will increase by a higher
rate (19%) than other building types, including residential buildings (13%) [19]. Given these
increasing trends in the energy consumption, we focused on commercial buildings.
Within the commercial building sector, we focus on office buildings which have the largest floor
space and energy consumption share of commercial buildings and are responsible for the largest
amount of CO2 emissions [20]. In addition, the amount of three key energy end uses in office
buildings including artificial lighting, air-conditioning, and IT equipment use, have steadily
increased over the past years [21] and measured electricity demands can be approximately 60–
70% higher than predicted [22]. These buildings present great opportunities in energy reduction
with behavioral approaches due to occupants’ lack of direct financial incentives for energy
conservation (occupants are not paying energy bills and consequently, they are not motivated to
reduce energy costs [23]), and the average time that a person spends in an office building (7.9
hours per day [24]). Therefore, we focus on promoting pro-environmental behaviors among
occupants of office buildings.
Simulations of occupant behavior demonstrated that occupant behavior has significant impact on
energy use of private offices [25]. These simulations and findings of other studies prove that
occupant behavior is one of the most significant contributors to building energy consumption and
motivating occupants toward adaption of pro-environmental behaviors, such as adaption of
daylighting instead of artificial lighting and adaption of natural ventilation instead of mechanical
ventilation, could result in considerable energy savings (5 to 30%) [26-28]. In addition, improving
occupants’ pro-environmental behaviors provides higher savings with much lower costs than
investing in building systems and it can be applied to both new and existing buildings [27].
`
17

We investigated behavior change strategies to promote sustainable behaviors and obtain significant
reductions of energy usage. We focused on two approaches: 1) LEED branding, which is an
indirect behavior change strategy aiming to change occupants’ perception of the building by
developing a green image of a building by introducing the building as being LEED certified; 2)
direct communication, which simply focuses on persuasion through communication with
individuals, which can occur in any contexts (e.g., energy efficiency, security, and comfort) and
settings (e.g., commercial buildings, residential buildings, LEED certified buildings, and
conventional buildings).
1.2. Scope
We focused on the factors that are influenced the most by occupant behavior and have significant
impacts on buildings’ environmental impacts, such as lighting operations [29-31], temperature set-
point adjustments [29,32], and recycling [33]. In addition, we focused on private offices with
single occupancy, assuming that the occupant have freedom to interact and change his/her indoor
environment. Offices with multiple occupants involve the complexity of group behavior, which is
not covered in this dissertation.
This dissertation is structured as follows: Chapter 1 describes the problem, the motivation and the
scope of the work. Chapter 2 proposes the literature review and concludes the gaps in existing
research. Chapter 3 discusses the objectives and research questions. Chapter 4 introduces the
research tools and experimental environments that were used for data collection. Chapter 5, 6, 7,
8 present research methodology, results, and conclusions that address the research questions.
Chapter 9 lists the limitations of the research and provides future research steps. Chapter 10
concludes the dissertation.
`
18

Chapter 2. Literature Review
Behavioral intervention strategies are designed to influence the actions that individuals take with
regard to their health and environment. Examples are programs to encourage pro-environmental
behaviors, such as encouraging energy conservation [34,35] and recycling [36,37]. Various
intervention strategies aiming to induce behavior change have been used in different domains
especially in health, marketing, and psychology, targeting wide range of different behaviors and
populations. Although intervention strategies are important to behavior change, these interventions
were not successful in taking the full range of significant influences on human behavior into
account. We conducted a comprehensive review of the previous behavior change strategies to
identify the success or failure factors of these interventions and transfer these findings into the
design of a successful communication strategy.
Generally, there are two types of behavior change intervention strategies: indirect and direct
strategies. Indirect strategies such as feedback change behavior by influencing its underlying
determinants (e.g., increasing self-efficacy
5
) or evoking existing feelings and values (e.g., inducing
a feeling of guilt or making an individual’s pro-environmental values salient). On the other hand,
direct strategies, such as compliance gaining techniques, directly ask users to change their
behaviors.
Behavior change intervention strategies in the energy domain are mostly indirect strategies. These
strategies can be categorized into structural and informational strategies. Structural strategies try
to influence behavior by changing contextual factors, such as price policies (e.g. energy tax).
Informational strategies often try to influence behavior indirectly by changing people’s attitudes,

5
Confidence in ability to perform a behavior
`
19

preferences, and other social judgments. We reviewed the previous work in the energy domain
over a 50-year period and identified the following indirect strategies: commitment (a pledge to
behave in a specific way or achieve a certain goal) [38,39]; goal setting (setting a reference point
of 5% or 10% energy savings for occupants) [38,39]; information (general knowledge about
energy-related problems or various energy-saving measures occupants can adapt) [40]; feedback
(providing occupants with information about their energy consumption or energy savings) [41,42]
gamification (the process of using games to engage users and solve problems) [43,44]; message
framing (highlighting either the benefits of engaging in a particular behavior (a gain-frame) or the
consequences of failing to engage in a particular behavior (a loss frame) [45]; modeling (presenting
an appropriate model that performs a desired action) [46]; and providing an incentive or
disincentive (strategies that motivate a range of pro-environmental behaviors) [47,48]. Among
these behavior change strategies, feedback has received the greatest attention and many studies
showed its superior impacts [49-51]. However, collecting, processing and, representing users’
energy information require advanced sensing technology and energy infrastructure. It requires the
replacement of traditional electricity meters with digital meters, which allow for wireless
communication of information back to the users. These technologies are too expensive and the
required infrastructure is not available in many parts of the U.S. There are also other challenges
associated with feedback strategies. For example, comparative feedback (comparison between
individuals or groups about their energy consumption) was found to be one the most effective
feedback strategies [34,52]. However, for this strategy to be successful, there is a need to have
peers in the network with pro-environmental behavior [53,54]. In fact, interventions to promote
pro-environmental behaviors are needed before promoting this kind of behavior in social networks.
Therefore, a practical strategy should be compatible with the existing infrastructures (e.g.,
`
20

metering infrastructure and social networks). Green branding is an indirect behavior change
strategy that try to promote behavior change only by highlighting the positive environmental
aspects of a product without any extra effort and cost to implement the interventions. It simply
takes advantage of functional attributes and emotional benefits of a green product to promote
sustainable consumption. There is a growing interest among researchers regarding green branding
strategy and its influence on user behavior. Several studies found significant and positive
relationships between green product features (e.g., eco-labels) and purchasing behavior of
consumers [55,56]. Studies in the marketing literature also confirm that green branding can
introduce greener patterns of consumption into contemporary lifestyles, such as customers'
purchasing patterns for green products [57]. Green branding could also affect intentions, for
example, deciding to stay at a green hotel, to recommend it, and to pay more for it [58-61].
Although research suggests that methods that use green branding have the potential to evoke
greater conformity (acting in compliance to the green standard) [55,56], such a possibility has yet
to be explored for influencing pro-environmental behaviors in buildings. In this chapter, first, we
provide a review of studies investigated whether knowing being in a green building positively or
negatively affects the building’s occupants’ environmental behavior. Then we explore the
characteristics of direct behavior change strategies.
2.1. Green Branding Intervention Strategy
A number of studies investigated the level of practice in energy saving behaviors between green
and conventional buildings. Some argue that occupants in green buildings practice energy saving
behaviors better than the occupants in conventional buildings [3,62]. For example, Steinberg et al.
[3] conducted a survey investigating occupants’ willingness to adopt energy saving behaviors in a
new green building and in a conventional commercial building. The results showed that the
`
21

occupants anticipating to move into a new green building showed more willingness to adopt energy
saving behavior as they were encouraged by the green building certification and wanted to ensure
the building performance success. In addition, Azizi et al. [62] conducted another survey to
investigate the level of energy saving behavior between green and conventional office buildings
in Malaysia. The results of their study showed that occupants in green buildings performed better
energy saving behaviors than occupants in conventional buildings because of the intervention
strategies implemented in the green buildings such as putting up posters on energy efficiency
features of the building. The results of another survey conducted by Azizi et al. [63] showed that
participants in the LEED certified office buildings were more willing to sacrifice their comfort and
adopt more energy saving behavior in comparison with the occupants in conventional buildings.
For example, in response to being cold, the occupants in green buildings adopted more personal
adjustments (e.g., wearing warm clothes) and less environmental adjustment (e.g., changing
temperature set point) compared to occupants in conventional buildings. The authors suggested
that the perception of working in a green building could be the reason to influence occupants’
behaviors towards more green behaviors. These studies showed that green buildings promote more
pro-environmental behavior among the occupants (conformity).
However, there are also studies that show green buildings do not motivate occupants’ pro-
environmental behaviors. It is possible to have a “green” building with “gray” occupants, due to
their lack of knowledge about how to use the systems in the building in the most efficient and
effective way [64]. Gray occupants are also more likely to be found in green buildings due to the
“rebound effect” or "take back", which is the term used to describe the effects caused by
sustainable solutions [65,66]. For example, a study showed that when the recycling option was
available, consumers increased usage of products that were free (e.g., office paper, bathroom paper
`
22

towels, etc.), creating adverse effects [67]. A reason for this kind of behavior might be the fact that
a green building might be considered as remedy to relieve the negative emotions, such as guilt
associated with wasteful behavior and disposing of a product [68]. Another reason might be the
"licensing" effect, which suggests that a prior environmentally responsible choice affirms
individuals’ values of social responsibility and ethical consciousness; therefore, it can license less
environmentally responsible and morally questionable behavior in a subsequent choice [69].
“Greenness” of a building might also seem as a remedy, which undermine consumers' risk
perceptions and increase their intentions to consume the product [70]. Applying the concept to
buildings, occupying a green building might be considered as a remedy for the occupants and
license them for less environmentally behavior as they feel that the building's sustainable
technologies will compensate for their wasteful behavior (compensation). Therefore, it is possible
that occupants in green buildings behave in a more wasteful manner. In other words, being an
occupant of a green building may have the unintended consequence of increasing consumption
[71].
As summarized above, there are conflicting conclusions about whether knowing being in a green
building positively or negatively affects the building’s occupants’ environmental behavior.
Sustained progress in green buildings results from understanding of sustainability values of the
building and occupants' engagement with them and without proper engagement, occupants might
alter the way, in which the building is designed to operate. Therefore, there is a need to investigate
how green branding, developing a green image of a building, as an indirect behavior change
technique can be used to promote occupants’ pro-environmental behaviors in buildings.
`
23

2.2. Communication and Compliance Gaining
Direct behavior change intervention strategies focus on persuasion through communication with
individuals. In general, five components have been identified for the development of a successful
communication. These components include intended audience, intended behavior, message
design, delivery style, and communicator [72,73]. Therefore, operational decisions in developing
successful communication efforts should include selecting effective message strategies,
determining optimal modalities for the delivery of persuasive messages, and choosing persuasive
communicators [74].
In this dissertation, first, we explored the areas of waste in buildings in order to identify the
behaviors that have the potential to provide higher savings if changed (intended behaviors);
second, we assessed different social influence methods to identify persuasive messaging strategies;
third, we investigated the possible modalities to deliver these messages; forth, we examined the
different assumptions for communicator's persona; fifth, we assessed the effects of different
communication modes (monologue vs. dialogue); sixth, we tested the effects of user characteristics
on the outcomes of these interventions; and finally, we examined the effects of intervention over
longer periods of time
2.2.1. Intended Behaviors: Areas of Wasteful Behaviors in Buildings
Waste in buildings refers to the unnecessary amount of energy and materials that are currently
being used. Office occupants are the main contributors to building energy consumption and
generate a lot of waste in buildings. In an office space, paper is the greatest component of the waste
stream [75]. An average office worker uses 10,000 papers annually and 45% of them end up in the
trash by the end of day [76].
`
24

Energy use in buildings is defined by the interactions between end users and building systems and
features. Therefore, accurately assessing behavioral components of occupant-building interactions
is critical for identifying areas of waste in buildings. Direct or indirect occupant interactions with
building systems and features impact buildings’ operations and energy usage. Examples of a direct
human-building interaction are opening/closing windows, turning on/off or dimming lights,
turning on/off office equipment, turning on/off heating, ventilation and air-conditioning (HVAC)
systems, setting indoor thermal and comfort criteria [25]. An example of an indirect human-
building interaction is adjustment of clothing to feel more comfortable when the room temperature
is high or low [77].
Studies showed that occupants’ conscious and unintentional actions in response to a specific
stimulus, have significant impacts on buildings’ energy demands [78,79]. Occupants might change
their energy usage characteristics by adapting more energy efficient practices or on the contrary,
by adapting poor consumption habits, which negatively impact energy consumption by increasing
energy use. As an example of these poor habits, we can refer to the "rebound effect", which is
explained by the tendency of occupants to use more electric lighting, following the installation of
energy saving bulbs, assuming that their actions will have less impact on the environment [80].
Numerous studies analyzed the way building occupants influence total energy consumption levels,
mainly through their behaviors and interactions with building systems and appliances. These
studies identified the most influential factors on total energy demand as: lighting operations [29-
31], temperature set-point adjustments [29,32], blind operations [81], window operations [29,82-
84], appliance usage [85,86], and occupant schedules [87-90]. For example, one study indicated
`
25

that adaption of day lighting rather than adaption of artificial lighting could reduce overall primary
energy expenditure by more than 40 percent [91].
Many studies focused on energy usage in commercial buildings during working hours and
neglected the influence of occupant behavior on building energy usage during non-working hours.
However, considerable amount of energy consumption happens during non-working hours. This
consumption is associated with employee’s behavior of leaving lights and office equipment on at
the end of the day [28,92-95]. In 70 case studies done by Rurke in office buildings, it was
demonstrated that unnecessary use of artificial lighting had the greatest opportunities for
behavioral waste reduction, followed by office equipment, and cooling. The results showed that it
was possible to reduce lighting energy usage in office buildings by 36%, office equipment energy
usage by 23%, and cooling energy usage by 12%. Considering that lighting usage constitute 30%,
office equipment 21%, cooling 15% of total energy usage in office buildings in 2012, the total
amount of energy usage that could be reduced is considerable [96].
Other factors, such as occupancy sensors for lighting, might also influence occupants' lighting
related behaviors and energy consumption. For example, Eliers [97] investigated lighting usage in
private offices in a university office building, where there were occupancy sensors in each room.
The results showed that existence of occupancy sensors in these rooms not only increased the
energy costs including the cost of sensors and their installation, but also resulted in doubling the
lighting energy usage due to the occupants’ tendency to rely on the sensors to control the lights.
The presence of occupancy sensors in these offices changed the propensity to turn off the lights
when leaving the room, and decreased the likelihood that an occupant would choose a light level
setting other than full illumination.
`
26

In general, occupant’s presence and behavior in buildings have significant influence on buildings’
energy consumption through the use of lighting, space heating, cooling, ventilation systems, and
appliances. Occupant’s wasteful behavior can result in more than 30% increase in a building's
predicted energy consumption [98] and 45% increase in paper consumption [76]. Therefore,
lighting, air-conditioning, office equipment, and paper consumption are the key areas of waste due
to occupant behaviors in office buildings. Many studies anticipated energy and material savings if
occupant behavior was modified. These modified behaviors include adaption of day lighting
instead of artificial lighting, natural ventilation instead of mechanical ventilation, shading to
control solar heat gains and glare, efficient heating by closing the windows when heater is working,
choosing efficient temperature setpoints, turning off office lights and HVAC during non-working
hours [28,92,93], turning off office equipment during non-working hours [99,100], and recycling
[101].
2.2.2. Messaging Strategies
Compliance gaining is a popular approach in persuasion. Persuasion is defined as the process of
influencing a person’s values, beliefs, attitudes, or behaviors [102]. Compliance gaining refers to
interactions, in which one individual (the agent) attempts to induce another person (the target) to
perform a desired behavior that the target person otherwise might not have performed [103].
According to Scott, compliance gaining techniques focus on modifying behavior directly rather
than influencing the presumed cognitive determinants of behavior [104].
Different strategies are used to obtain compliance from users. Among the most common
techniques used in the literature are foot-in-the-door and reciprocity techniques. Foot in the door
(FITD) technique is one of the classic compliance-gaining strategies receiving great attention in
`
27

the marketing literature. In using the FITD technique, the requester first makes a request so small
that nearly anyone would comply. After the compliance with the first request occurs, a second,
larger request is made - the one desired from the outset. People who comply with the initial, small
request are more likely to then comply with the second request, compared with a control situation
without a first request [105,106]. This technique may be especially effective in social contexts,
where initial attitudes may not be positive and gaining compliance to the larger request is difficult.
In addition, the theoretical rationale proposed for this technique suggests that it may result in more
persistent and more widespread behavioral influence because it modifies an individual’s
perception of himself. That is, performing the initial behavior may lead the individual to perceive
himself as one who takes action on his beliefs, and thus increase his sense of self efficacy [104].
Many studies investigated the effects of FITD on gaining compliance. For example, Freedman and
Fraser [105] conducted two experiments investigating the influence of FITD technique on
convincing housewives to allow a survey team of five or six men to come into their homes for two
hours to classify their household products or to display a large ugly sign concerning safe driving
in their front yard. These interactions involved phone contact, and the results showed that carrying
out a small request increased the likelihood that the subjects would agree to a similar larger request.
In another study, Gue´guen [107] carried out an experiment on fifty computer science students by
sending an e-mail with a forty-question survey on their food habits, which required 15–20 minutes
of their time. This questionnaire came from a hypothetical student at participants’ own university.
Half of the students had earlier responded to a small solicitation made by this same solicitor.
Results showed that more compliance to the second request was obtained than in a control situation
that did not involve the initial small solicitation. The results of similar studies showed that FITD
technique increased compliance to the final request [108-110]. Studies also showed that FITD
`
28

technique, which is a human–human communication technique, can be used in computer mediated
communication settings [107].
The norm of reciprocation, often referred to as the "golden rule" [111], obliges us to repay others,
for what we have received from them. Reciprocation is a strong, pervasive social force in all human
cultures and can be a powerful persuasive strategy [112]. People reciprocate because they do not
want to be judged as ungrateful for favors they receive [113]. Many studies have investigated the
effects of reciprocity on gaining compliance. For example, Regan [114] conducted a laboratory
experiment to examine the effect of a favor from a confederate on a compliance with a request.
The results of his study showed that participants were more willing to purchase raffle tickets from
a confederate if that confederate had earlier given them a soft drink as an unexpected favor. In
another study, Whatley et al. [115] examined the influence of the social (public) and internal
(private) consequences of reciprocation on behavior. In their study, participants were randomly
assigned to conditions, in which they were or were not given a small favor, and then were asked
to comply with a request. In addition, participants expected that the person who asked for the favor
would either know or not know whether they complied with the request. The results showed that,
although, public compliance was greater than private compliance, the presence of a favor appeared
to increase compliance in both public and private conditions.
Research suggests that the rule of reciprocity in human-computer interactions can also influence
users’ behaviors [116-118]. For example, in creating a system to promote behavior change, Fogg
and Nass [119] showed that it is possible to take advantage of the rule of reciprocity to motivate
participants. In a web search task, participants were more likely to agree with, comply with, or
help out a computer that had previously helped them. The influence of obligations to reciprocate
`
29

has been shown in many other studies [120,121], highlighting the importance of reciprocation for
increasing the compliance.
Considering the positive effects of compliance gaining techniques (FITD and reciprocity) on
increasing adherence, we believe that adaption of these techniques in message design can
potentially enhance the effectiveness of the intervention strategies designed to promote sustainable
behaviors in buildings.
2.2.3. Communication Modality or Delivery Style
Previous studies suggest that communication modality (delivery style) is an important factor
influencing the effectiveness of behavioral interventions. In general, studies showed that people
interact differently to media based on its relationalism, which refers to the social aspects of media.
For example when interacting with another person, text has less relationalism than talking on the
phone and is considered as leaner media. On the other hand, face to face communication has more
relationalism than talking on the phone and is considered as richer media [198]. The required extent
of the relationlism of an effective communication media depends on the context of the
communication (e.g., health consultation in marketing). Context investigated has varied widely
among studies and different results have been obtained based on the context. For example, in some
tasks such as problem solving and information transmission, in which the social relationships
between the users are relatively unimportant, research has shown that nonverbal cues are of no
great importance in determining outcome of these tasks [122]. For example, Davies [123]
compared two modalities (audio only and face to face) for solving problems with objective
solutions and found little difference between these two modalities in the accuracy of the solution
achieved. In another study, Champness and Davies [124] compared two modalities (audio only
`
30

and face to face) for solving an open-ended human relations task and found no difference between
these two modalities in the final solutions agreed upon.
In addition, some tasks such as negotiations and social dilemmas
6
, in which social relationships
between the users are relatively important, are sensitive to media differences [125]. For example,
Bos Olson et al. [126] demonstrated that richer media (e.g., audio-mediated communication)
increases trust based corporation in a social dilemma game compared to less rich media (e.g., text
chat). Another study in social dilemma research paradigm investigated the effects of four forms of
communication, including no communication, text-chat, text-to speech, and voice, on the
development of trust and cooperation. The results showed significant differences between these
forms of communication, with the voice condition resulting in the highest levels of cooperation
[127]. In another study, Short [125] conducted a study, in which pairs of participants were asked
to participate in a simple negotiation task and discuss the areas that there is a need for cut in the
expenditures of hypothetical government corporations. The results showed that communication
modality had significant effects on the negotiation outcome and when the participants were arguing
the case that was constant with their views did relatively well, in terms of the negotiation outcome,
under face-to-face or audio-video communication. In another study, Chaiken and Eagly [128]
investigated the effects of using written, audiotaped, and videotaped modalities to deliver
persuasive messages and the results showed that more opinion change was obtained in the
videotaped and audiotaped, as compared to written condition. Overall, these results are parallel to
the findings that people are more cooperative in such social dilemmas with computers that look
like humans compared to those without human-like features. Indeed, more human-like delivery

6
Social dilemma is a situation in which there is a conflict between an individual’s self-interest and his or
her shared interests with other individuals.
`
31

styles (e.g., voice, face) evoke more cooperation, persuasion, and opinion change than those
without a human likeness (e.g., text, computer without a face) [129].
However, there are some contexts such as healthcare, in which people prefer leaner media [130].
For example, Greaney et al. [131] investigated the influence of reminder modality (voice reminders
vs. text reminders) on people’s engagement with behavior change (i.e., cancer prevention) and
association between peoples’ characteristics and preferred reminder modality. The results showed
that more than two-third of the participants chose text reminders over voice reminders. Another
study conducted by Crawford et al. [132] investigated the difference in the success of health
messages delivered through an SMS or a voice messaging. The results showed that SMS was the
preferred modality and intended or actual behavior change was significantly higher among
participants that received the message through an SMS compared to the participants that received
the voice message.
Although many examples of persuasive technologies have been developed in the energy domain,
these communication strategies lack the inclusion of social features if were incorporated could
enhance the effectiveness of these strategies. Communication delivery styles (e.g., text and voice)
are one of the important components of every communication strategy that tends to influence
users’ behavior and their satisfaction with the communication system [133]. In general, studies
showed that people interact differently to media based on its relationalism. Considering that more
human-like delivery styles (e.g., voice, face) evoke more cooperation and persuasion in social
context than those without a human likeness (e.g., text, computer without a face), one situational
variable, which has not yet been investigated in energy domain, but potentially may be very
important given the findings in literature, is relationalism of the communication modality.
`
32

2.2.4. Communicator Persona
A series of empirical studies conducted by Reeves and Nass [134] suggests that people interact
with computers in social way. Many studies showed that it is possible to communicate and build
a relationship between people and their cars and other inanimate objects using an agent [135-
137]. An agent is a program that appears to have the characteristics of an animate being. These
agents are called relational agents. Relational agents are intelligent agents designed to develop
and maintain long time socio‑emotional relations with users [138].
Relationships can play a significant role in persuasion [139]. The theory of relational agents
includes a complex model of human-computer relationships. Over the last decade, a series of these
agents have been developed for a variety of counseling, education and behavioral-change
interventions. For example, an agent can play the role of a personal trainer, designed to motivate
users to exercise more [139] or users can interact with agents, who play the role of different human
professions like a therapist [140] or a medical assistant [139]. Studies showed that computer agents
that function in helping roles and attempt to induce a change in behavior could be much more
effective if they first build empathetic relationships with their users [141]. An agent’s “persona”
or the character the agent portrays have significant roles in shaping relationships. An agent’s
persona should be based on its relation to the user, and its capabilities and functions and should fit
the purpose of the system and the role of the agent in that system [138]. Some believe that artificial
agents should not be presented as a human as it makes false promises and results in overblown
expectations about what an agent can or should do. However, research suggests that people treat
computers as human, even when the computer interface is not explicitly anthropomorphic [134].
On the other hand, some believe that the agents should represent more believable persona to know
that characters really care about what happens in the world, that they truly have desires [142].
`
33

Studies showed that persona, who is controlling the agent, impacts the user’s behaviors. Some
showed that agents representing human roles are more persuasive in social contexts [143] and the
perception of interacting with human increases the sense of social presence and facilitates social
responses more than the perception of interacting with a computer or an inanimate object [143-
145]. For example, Lim and Reeves [146] examined how a character’s persona (perception of
playing a game with a human vs. playing a game with a computer) impacts the game outcomes
(valence, presence, and likeability). Their results showed that persona can significantly influence
game play outcomes (i.e., physiological arousal, greater presence and likability), suggesting that
interacting with an agent representing a human increases the realism and consequence of media
interactions by increasing the social presence experienced by the participants. Okita et al., [147]
investigated whether the perception of having a social interaction with a human vs. a computer
influences participants’ learning and understanding. In their study, participants asked scripted
questions (about the mechanism that causes fever) to a virtual representation, which was believed
to be either a human or a computer and the virtual representation provided scripted responses. The
results showed that the virtual representation led to better learning outcomes when participants
believed that they were interacting with a human.
On the other hand, agents representing inanimate objects especially machines and computers are
more effective in some other contexts, especially in the contexts that social influence has negative
impacts. For example, de Visser et al., [148] investigated individuals’ compliance and performance
while receiving recommendation and feedback from a human or a computer while trying to
discover a pattern in a sequence of numbers. The results of their study showed that performance
and compliance was higher with computers than humans in reliable trials. Hoyt et al., [149]
investigated if individuals’ performance on categorization and pattern recognition tasks is affected
`
34

under the perception of being in front of a human or machine. The results showed that being in
front of a human negatively impacted their behavior, while being in front of a machine did not
influence their performance.
In general, virtual representations can be used as a method of influence and social interactions with
agents like human interactions can modify users’ behaviors [143]. Given the effects of social
representations in persuasive contexts, it is important to identify whether there is a difference in
how the communicator’s persona is perceived by the building occupants and how different
personas impact the effects of different components of the communication (e.g., social dialog).
2.2.5. Social Dialog
A relational agent can be designed to use various verbal and nonverbal strategies to move the
relationship in a desired direction towards the achievement of both task goals and relational goals.
Social dialog is an example of a verbal strategy. Social dialog involves an interaction between two
speakers and can be used by an embodied interface agent to build a relationship with the user to
satisfy a task goal such as obtaining compliance with a request. Several studies have explored the
variables that affect the persuasiveness of a virtual representation. However, potential capabilities
of interactive approaches such as dialog are not being fully explored in most current behavior
change research [150]. According to Garrod and Pickering [151], humans are designed for dialog
rather than monolog. In general, a successful communication requires an active process of both
listening and talking and this bidirectional communication is a precondition for any positive
change in behavior [152]. Social dialog is an essential mechanism for enhancing relationships
[153]. The level of interactivity in a communication can affect communication processes and
outcomes. Degree of participation is one of the main factors that affect the level of interactivity.
`
35

Dialogic communication involves more participation than monologic communication. Dialog
increases the mutuality and interactivity between the two sides of the communication and provides
better opportunity for influence [154].
Social dialog has been used in different contexts to encourage the desired behavior. For example,
Cassell et al. [155] designed an agent representing a real-estate agent using social dialog
interviewing clients about their housing needs. Agents adopting social dialog are also being used
increasingly in healthcare. For example, Bickmore et al. [139] designed an agent who played the
role of an exercise advisor and engaged the patients in a social dialog to persuade them to exercise
more. Research has also demonstrated the effectiveness of social dialog on promoting compliance
with a request. For example Howard [156] investigated how initiating a conversation with
individuals by asking them about how they feel and acknowledging their response elicit
compliance with a charitable request. In another study, Dolinski et al. [157] examined the effects
of monolog vs. dialog on compliance with charitable requests and the results showed that
participants in the dialog condition were more likely to comply with the request for a donation
than the participants in the monolog condition. Dolinski and Grzyb [158] investigated the
effectiveness of social dialog as a social influence technique on a paltry contribution and the results
showed that dialog, specifically social dialog, resulted in more success than monolog.
The effects of social dialog might be mediated by other characteristics of the communicator. For
example, the persona of the communicator might impact the individuals’ willingness to engage in
a social dialog. However, in review of previous studies, we found no empirical study exploring the
effects of social dialog on the effectiveness of interventions aiming to change human behavior
while the persona of the communicator varied. In most of the studies that social dialog was
implemented, the agent was representing a human-like persona and it is not clear if the same results
`
36

were obtained if the virtual representation personified an inanimate object. Therefore, there is a
need to investigate if social dialog has the same effects on people in human-building interactions
as in human-human and human-computer interactions and if the effects of social dialog are
associated with the persona that communicator represents.
2.2.6. Users’ Characteristics
Occupant related determinants, which are related to the attributes of occupants such as occupants’
knowledge, psychological factors, and physiological factors can dictate how an occupant behaves
and interacts with a building and building systems [159]. In addition, studies showed that people
apply social rules to social agents and react differently to them based on their own characteristics
(e.g., demographics and personality traits) [117]. In this section, we reviewed the user
characteristics that impact their energy consumption behaviors as well as their interactions with
social and relational agents.
2.2.6.1. Determinants of energy related behaviors
Occupant related determinants such as knowledge, which is defined as awareness of high-energy
consumption consequences, understanding of energy use, and energy saving options, have
essential impacts on the occupants’ energy consumption behaviors [160-162]. Occupants’ attitudes
and values can also determine their behaviors. Specifically, the "energy attitude" of an occupant is
important. The link between occupant’s environmental attitudes and patterns of energy
consumption is relatively strong. Environmentally concerned occupants tend to be more
conservative on energy use [163]. For example, studies suggest that there is a relationship between
green consumption and an individual's environmental attitudes (values and views). Thogerson and
Olander [164] found that sustainable consumption patterns are influenced by individual’s value
priorities. Karp [165] investigated the influence of values on environmental behavior and found
`
37

that altruistic values have positive influence on environmental behaviors. Azizi et al. [166]
suggested that energy awareness and knowledge around the existing solutions to the environmental
issues could be the motivating factor for occupants to practice energy saving behaviors. In
addition, occupants need to have knowledge and skills to engage in these practices. In summary,
the attitudes and behaviors of occupants can significantly affect a building’s performance during
the building’s life [167].
Several studies also showed that physiological factors, such as occupant’s gender and age
(especially gender) influence occupants’ energy consumption behavior [168-170]. For example,
Karjalainen showed that females prefer higher room temperatures than males [171]. In addition,
studies focused on age as a determinant of energy related behavior, showed that in general, older
occupants tend to consume more energy than younger occupants, especially for space heating
[172]. Occupants’ personality traits also influence their energy related behaviors. For example,
Paivio [173] showed that preference for dimmer light in different situations was associated with
occupant’s level of nervousness (neuroticism).
2.2.6.1. User characteristics and intervention strategies
Studies showed that user characteristics impact the effects of behavior change interventions. For
example, Nass and Moon [117] showed that individuals apply social rules to computers. Their
results showed that users perceived the agents with their own ethnicity to be more attractive,
persuasive, trustworthy, and intelligent compared to those in the different ethnicity condition.
Furthermore, in another study Nass and Lee [174] showed that there is a link between computer
voice personality and user’s personality for voice attractiveness, such that introverts preferred the
introvert voice and extroverts preferred the extrovert voice. Studies in the health domain also
showed that effectiveness of intervention strategies can be mediated by the recipients of the
`
38

interventions. For example, Albarracín et al. [175] investigated the effectiveness of some behavior
change intervention strategies (e.g., educational information and attitudinal arguments) on
promoting HIV prevention behavior. These reviewers identified that the impacts of interventions
were moderated by the recipients’ characteristics, such as age, gender, and ethnicity. For example,
normative argument strategies were more effective than threat-inducing arguments for groups
under 21 years old, while, behavioral skills arguments, HIV counseling and testing, and self-
management skills training, were more effective for groups over 21 years old [175]. Studies showed
potentials of improvements by designing interventions based on user characteristics. For example,
Shen et al., [176] showed that by incorporating personality traits into an energy-saving behavior
analysis, they could generate a 15.5–20% greater reduction in building energy consumption.
Review of previous studies showed that different user characteristics can impact user behavior and
the effects of behavior change interventions. Therefore, it is important to investigate how users’
characteristics (e.g., demographic and personality traits) can affect the ways intervention strategies
influence energy consumption behaviors.
2.3. Frequency of Interactions
Although some of the behavior change interventions reported promising results over one time
interaction, it is not clear whether their effects were maintained over more interactions and a few
studies examined the subject did not find positive results. Studies reported short-term compliance
rather than long-term adherence as results of behavior change interventions. For example,
Bickmore and Picard [141] conducted an intervention using a humanoid agent to promote physical
activity and they observed short-term increase in physical activity during the intervention but
decrease in physical activity after the follow-up. Building-occupant communication requires
`
39

continuous interactions between buildings and their occupants and differs from one time
communications that are observed in other contexts such as marketing. Therefore, it is important
to investigate the effects of interventions over more interactions.
`
40

Chapter 3. Research Objectives and Questions
In this dissertation, we investigated the effects of behavior change strategies designed to influence
the way that occupants behave in buildings. Specifically, we focused on strategies aiming to
change building occupants’ behaviors associated with energy consumption and recycling. These
strategies were designed to promote occupants’ pro-environmental behaviors in buildings. We
explored the effects of indirect and direct behavior change strategies including LEED branding
and direct communication.
First, we examined the influence of LEED branding on building occupants' pro-environmental
behaviors in office buildings. LEED branding, in this dissertation, is defined as developing a green
image of a building by introducing the building as being LEED certified. In addition, we
investigated how occupants’ environmental values/views affect the influence of LEED branding.
Objective 1: To investigate the influence of LEED branding on occupants' pro-environmental
behaviors in buildings.
 Q 1-1a: How does LEED branding influence building occupant’s lighting preferences
(natural light vs. artificial light) and recycling behaviors in buildings?
 Q 1-1b: How do environmental values and views affect the influence of LEED branding
on lighting preferences and recycling behaviors?

We also explored the effects of direct behavior change techniques. We focused on enabling a
human-building communication and investigated the effects of integrating social and relational
features into the design of different components of the human-building communication (e.g.,
messaging strategy, delivery style, persona, and social dialog). In addition, we investigated the
effects of user characteristics and continuous interactions on the effects of these communications.
`
41

Objective 2: To investigate the effects of incorporating social features (i.e., social influence
methods and social surface cues) in design of communication intervention strategies aiming to
promote occupants’ pro-environmental behaviors in buildings.
Q 2-1: Social Influence Methods
 Q 2-1a: What are the effects of social influence methods (i.e., foot in the door and
reciprocity) on building occupants’ compliance with pro-environmental requests (e.g.,
adaption of day lighting instead of artificial lighting)?
 Q 2-1b: Do occupants’ characteristics (e.g., demographics, personality traits, and
technology readiness) impact the effects of social influence methods on building
occupants’ compliance with pro-environmental requests?
Q 2-2: Social Surface Cues
 Q 2-2a: What are the effects of incorporating social surface cues (i.e., voice and face) to
the delivery of pro-environmental requests on building occupants’ compliance with the
requests?
 Q 2-2b: Do occupants’ characteristics (e.g., demographics, personality traits, and
technology readiness) impact the effects of different delivery styles on building occupants’
compliance with pro-environmental requests?
Objective 3: To investigate the effects of incorporating relational features (persona and social
dialog) in design of communication intervention strategies aiming to promote occupants’ pro-
environmental behaviors in buildings.
Q 3-1: Communicator’s Persona
`
42

 Q 3-1a: What are the effects of different personas (i.e., building facility manager and
building itself) on building occupants’ compliance with pro-environmental requests?
 Q 3-1b: Do occupants’ characteristics (e.g., demographics, personality traits, and
technology readiness) impact the effects of communicator’s persona on building
occupants’ compliance with pro-environmental requests?
Q 3-2: Social Dialog
 Q 3-2a: What are the effects of social dialog on building occupants’ compliance with pro-
environmental requests?
 Q 3-2b: Do occupants’ characteristics (e.g., demographics, personality traits, and
technology readiness) impact the effects of social dialog on building occupants’
compliance with pro-environmental requests?
Considering that the persona of the communicator might impact the individuals’ willingness to
engage in a social dialog. We also examined the influence of the social dialog on the persuasive
effects of pro-environmental requests under the perception of interacting with agents with different
personas. We investigated the following questions:
Q 3-3: Interaction of Persona and Social Dialog
 Q 3-3a: Does persona of the communicator influence the effects of social dialog on
promoting compliance with pro-environmental requests?
 Q 3-3b: Do occupants’ characteristics (e.g., demographics, personality traits, and
technology readiness) impact the interaction effects of persona and social dialog on
building occupants’ compliance with pro-environmental requests?
`
43

Objective 4: Considering that building-occupant communication requires continuous interactions
between buildings and their occupants, we also investigated the effects of the requests over more
interactions to examine if the repetition of the requests impacts the results. Therefore, we
investigated the following question:
 Q 4-1: Does occupants’ compliance with the pro-environmental requests change over one
more interaction with the building?

`
44

Chapter 4. Research Tools
To study occupant behavior in general and occupants’ interactions with buildings, many studies
used field data collection techniques, such as field experiments along with survey methods to
collect occupant-related data. Several studies also found virtual environments (VEs) to be valuable
tools to study human behavior [149,177,178]. In this dissertation, we used different data collection
techniques including field experiments, surveys, and virtual environments (desktop/laptop based
and immersive virtual environments). Despite the differences between virtual interactions and
physical interactions, researches successfully use VEs to answer questions that cannot be easily
addressed just using traditional data collection techniques such as field experiments and surveys
[179], which are the tools that are used most frequently in the research for gathering data on
behavior. Although the application of virtual environments is still rare in the domain of pro-
environmental behavior, there are studies that illustrate the potential of virtual environments
[180,181]. In this chapter, first we provided an overview of the virtual environments over the
traditional data collection techniques. Then, we described some studies that we conducted to
ensure that virtual environments were good venue for testing real-world behavior change
interventions.
4.1. Virtual Environments
Virtual environments are defined as “synthetic sensory information that leads to perceptions of
environments and their contents as if they were not synthetic”. Immersive virtual environments
(IVEs) are ones that "surround" a person and create the feeling of presence in that environment
[145]. IVEs are products of the integration of different hardware and software systems and they
include a user interface to display the virtual environment to users, a tracking system to track the
`
45

users’ movements and a computer that choose proper portions of the virtual environment to be
shown within the interface [182,183].
4.2. Modeling (Creation of Virtual Environment)
For our investigations, we simulated an office environment. Virtual model of a single occupancy
office was created (Figure1). The main features of the modeled office included both natural light
and artificial light sources, and a thermostat. The participants were able to interact with the
environment by dimming or turning on/off the lights, opening/closing the blinds, and changing the
temperature setpoint.

(a) (b)
Figure 1 – Different views of the model: (a) Top view; (b) perspective view
The architectural geometry (e.g., walls, generic windows, ceiling, and light fixtures) of a 3D model
of a single occupancy office space was first created in Revit
©
2015 and then taken to 3ds Max
©
to
optimize the geometry and add materials, furniture, texture, lighting, reflection, and shadows and
in order to make it look more photo realistic (Figure 2).
`
46

Figure 2 – Modeled environment in 3ds Max
©
The model was then exported as an FBX file and imported to the Unity
©
game engine, where
interactive options (e.g., opening/closing blind animation, dimming and turning on/off lights
animation, changing the temperature set point) were programmed into the model in a way that
would allow the participants to have more realistic interactions with the virtual interactive model.
For example, participants were able to dim or turn on/off the lights by standing in front of the light
switch or open/close the blind by standing in front of a window. Through Unity
©
, it is possible to
connect the 3D models to different VE platforms. We used two different platforms to make users
to interact with the VE, PC laptop (Figure 3) and Head-Mounted Display (Figure 4). We used
Oculus DK2 Head-Mounted Display (HMD) to provide a fully immersive environment that
includes an Xbox-360 controller, and a positional tracker that would track the participants’ head
and neck movements (Figure 4).

Figure 3 – Modelled office – illustrating what participants see on the PC laptop
`
47

Figure 4 – Modeled office - illustrating what participants see through Oculus

4.3. Benefits of the Virtual Environments
Many studies in different domains have utilized VEs to simulate real-life settings [145,184-186].
VE tools, specifically IVEs offer several key methodological benefits to the researchers, including
the ability to manipulate complex, abstract objects, and concepts while maintaining high
experimental control and realism. VE provides the same environment for all participants (e.g. same
level of lighting). For example, uncontrollable factors, such as cloudy/sunny weather in different
days or different window types, which influence the level of natural light in a room, might affect
the results by causing experimental noise. VEs allow the experimenter to control for most of the
potentially confounding variables and isolate the variables of interest (e.g., influence of pro-
environmental requests). One significant advantage of VEs, specifically IVEs, is that users cannot
see the experimenter so they do not feel like the experimenter is watching and/or judging them and
it results in a behavior that is less reserved or more realistic [187]. In addition, there could be
various variables in design of interventions, which result in different scenarios to take place in
different environments (e.g., different indoor layouts, materials, all of which might impact
occupants’ behaviors). VE's flexibility let the researchers to easily and quickly change an
environment’s settings (e.g., lighting settings, architectural and interior design settings, etc.)
`
48

without any extra cost to investigate their influence on individuals’ behavior and cognition. This
is a unique opportunity brought about by the use of VEs as these kinds of modifications could be
costly, time consuming and in some cases not feasible in real environments.
4.4. Virtual Environments vs. Physical Environments
Research has shown that users act in similar ways within VEs as they do in physical environments
and have similar feelings of presence within such environments [177,188-190]. For example,
Heydarian et al. [191] investigated whether IVEs could be adequate representations of physical
environments. They measured user performance on a set of everyday office-related activities (e.g.,
reading text and identifying objects) and benchmarked the participants' performances in a similar
real-world environment. The results showed that the participants performed similarly in an IVE
setting as they did in the benchmarked environment for all of the measured tasks. We also
investigated if participants would behave similarly in the physical environment as they do in the
virtual environments. We used two different VE platforms including head mounted display (HMD)
and PC laptop to make users to interact with the VE. Our results showed that there were no
significant differences in participants' behaviors in the virtual environments (both HMD and laptop
based VEs) vs. physical environment. Detailed explanation of experiment design and data analysis
for these studies are provided in Chapter 6 (HMD based VE vs. physical environment) and Chapter
8 (laptop based VE vs. physical environment). We also investigated if application of different
platforms of virtual environments (i.e., HMD vs. PC laptop) impacted participants’ behaviors. The
results of this study is presented.

`
49

4.5. Virtual Environment Platforms
VE systems that have been used across different studies vary in their dimensions specifically in
their level of immersion, the degree to which a system can “deliver an extensive, inclusive,
surrounding, and vivid illusion of virtual environment to a user [192].” Choice of an appropriate
system for a specific application impacts the effectiveness of VEs [193]. VEs are mostly
characterized based on their platforms (e.g., desktop/ PC laptop, head mounted display (HMD),
etc.). However, there is no guideline leading us to the selection of an appropriate platform. One of
the main considerations in the selection of appropriate VE platform is the level of immersion,
which is often linked to the sense of presence and simulator sickness (motion sickness experienced
during an interaction with a VE) experienced by users of those environments. “Presence is defined
as the subjective experience of being in one place or environment, even when one is physically
situated in another [194].” According to Slater et al. [195], immersion refers to the quantifiable
characteristics of a technology while presence refers to the user’s state of consciousness, the
feeling of being in the virtual environment, and corresponding modes of behavior.
Selection of the VE platform could impact the behavioral responses of participants engaging with
the VEs [196]. However, it is relatively unclear whether the level of immersion, sense of presence,
and simulator sickness experienced through different VE platforms influence user performance
and behavior. Often, the effects of VE platforms on performance and behavior appear to be
overshadowed by other factors, such as the context of the task and user characteristics (e.g., age,
gender, and previous experience with VEs) [197,198].
Studies have used different VE platforms and each platform has different characteristics and user
interfaces. For example, desktop based VEs use the standard monitors as display devices, while,
HMDs are composed of two LCD screens embedded in a glasses-like device, which represents the
`
50

virtual environment by detecting the user’s head orientation and position using a tracking system.
Desktop based VEs use abstract interfaces like mouse and keyboard to move and change direction
in the environment, whereas, in more immersive VEs, users can physically move around to change
direction as well as using abstract interfaces.
Selection of an appropriate VE platform for a specific experiment may be influenced by the context
of the behavioral responses that are examined. Several studies have been completed to investigate
this issue and compared various platforms with each other in different contexts. We conducted a
study exploring the influence of VE platforms on measuring the compliance with pro-
environmental requests. We compared a non-immersive VE (i.e., PC laptop) with an immersive
VE platform (i.e., HMD).
4.5.1. PC laptop vs. HMD
We examined the compliance with persuasive pro-environmental requests and performance, using
an office related task across two representative VE platforms (PC laptop vs. HMD). Participants
were randomly assigned to one of the two VE platform conditions. Half of the participants
completed the study on a PC laptop, and the other half completed the study using an HMD. We
investigated the following dependent variables: (a) compliance with the pro-environmental
request, (b) task performance, (c) presence, and (d) simulator sickness. Considering that research
shows contradictory findings regarding the influence of the level of immersion as a characteristic
of VE platforms on the sense of presence, behavior change, and performance, and these effects
have not been investigated before in social contexts that aim to influence behavior, we could not
hypothesize whether different VE platforms would generate different levels of presence or
influence the compliance with the persuasive pro-environmental requests or task performance in
the context of our study. Therefore, we investigated the following questions: (1) Do different VE
`
51

platforms (laptop vs. HMD) influence the compliance with pro-environmental requests? (2) Do
different VE platforms (laptop vs. HMD) influence the presence experienced by the participants
during the experiment? (3) Do different VE platforms (laptop vs. HMD) influence the performance
on the assigned task (reading a passage and answering some questions about it)? Considering that
user characteristics can be another factor that influence the sense of presence, we also investigated
the following questions: (4) Do different VE platforms influence the simulator sickness
participants experienced while interacting with the environment? And 5) Do participants’
characteristics (immersive tendency, previous VE experience, and gender) influence the presence
that they experienced during the experiment? We also investigated the interaction between the
sense of presence and effects of VE platform type on compliance with pro-environmental
behaviors, as well as the task performance. The same for the interaction between simulator sickness
and effects of VE platform type on compliance with the pro-environmental behavior and task
performance.
4.5.2. Research Methodology
Prior to the beginning of recruitment, the experimental protocol was approved by the University
of Southern California’s Institutional Review Board. Participants were undergraduate students at
the University of Southern California and recruited through the USC psychology subject pool and
received course credits for their participation. Out of 112 participants, who consented to
participate, three withdraw during the experiment due to simulator sickness symptoms, and nine
were excluded due to self-reported simulator sickness. Therefore, the final sample included 100
participants (66 females and 34 males). Participant self-reported major included: engineering
(31%), psychology (40%), medicine (19%), and other majors (10%) including architecture,
business, science, policy, communication, and cinematic arts.
`
52

We focused on two VE environments that varied in their level of immersion but used the same
technology to interact with the environment (an Xbox controller). As the VE platform with lower
level of immersion, we used a standard laptop (17 in. monitor) (Figure 5a). As the VE platform
with higher level of immersion, we used Oculus DK2 Head-Mounted Display (HMD) to provide
an immersive environment, in which a positional tracker would track the participants’ head and
neck movements (Figure 5b). The VE in the PC laptop condition was identical to what presented
in the HMD system, but the technical components including stereoscopic view and head tracking
were excluded.

(a) (b)
Figure 5 – Different VE platforms: (a) laptop display; (b) HMD
Pro-environmental requests included contents requesting participants to change the light level and
temperature setpoint using the principles of reciprocation, which is a social influence method. A
female avatar, representing the building manager, delivered the request. The adopted social
influence method and delivery style were found to be among the most effective ones [199].
Participants were randomly assigned to one of the two experimental groups: laptop group, in which
participants interacted with VE using a standard laptop without HMD, and the HMD group, in
`
53

which participants were immersed in a VE using HMD. The only difference between these two
groups was the platforms used for interacting with the VE.
The participants were asked to perform two office related simulated tasks and in each task, they
were asked to read a passage. The reading levels were kept consistent for the two passages to
ensure there would not be any biases. While completing each task, a pro-environmental request
was delivered to the participants (Figure 6). Participants had to decide whether or not to comply
with the pro-environmental requests. If the participants chose to comply with the requests, they
had to use the controller to walk to the thermostat or light switch in the virtual environment and
change the artificial light levels or temperature setpoints.
(a) (b)
Figure 6 – Pro-environmental requests delivered to the participant in the VE: (a) before the request was
delivered to the participant; (b) while the request was being delivered to the participant

4.5.3. Experiment Setup
4.5.3.1. Pre-experiment Session
Upon arrival at the research lab, participants received a consent form providing information about
the study and informed and signed consents were obtained from the participants. As the first step
`
54

of the experiment, participants were asked to complete a questionnaire covering general questions,
including age, gender, major, immersive tendency and previous experience with VE, and simulator
sickness symptoms.
The Immersive Tendency Questionnaire (ITQ) was developed by Witmer and Singer [194] to
measure the tendency or capabilities of individuals to be involved or immersed in a virtual
environment. ITQ is composed of 16 items divided to three subscales including involvement, focus
and game. Involvement items investigate the user’s tendency to get involved in common activities,
such as watching television or movie and reading books. Focus items investigate user’s mental
awareness and their ability to ignore distractions and concentrate on activities. Game items
investigate how frequently users play video games and how involved they feel when they play the
game. Simulator Sickness Questionnaire (SSQ) was developed by Kennedy et al [200] to measure
the level of simulator sickness symptoms in a VE system. It included a 16-item symptom checklist
(e.g., fatigue, headache, and dizziness) using a scale from 1 (none) to 4 (severe). These items are
divided to three subscales including nausea (general discomfort, increased salivation, stomach
awareness, difficulty concentrating, and burping), oculomotor problems (general discomfort,
eyestrain, difficulty focusing and concentrating, blurred vision, headache, and fatigue), and
disorientation (difficulty focusing, fullness of head, dizziness with eyes open and closed, and
vertigo). Prior to the main VE task, participants underwent training to become familiar with the
virtual environments. During the training, we showed the participants how to work with the laptop
setting or put the HMD on and work with the controller. The participants were asked to perform
different tasks, such as moving around the room, dimming or turning off the lights, and changing
the temperature set point.
`
55

4.5.3.2. Experiment Session
After the participants completed the training session, we explained to them the experiment
procedure. One group of the participants interacted with the virtual office through the laptop
monitor and the other group was immersed in the virtual office through the HMD. The procedure
was as following: participants had to sit on a chair in an office and read two passages very carefully
as they had to answer comprehension questions about the passages. During the experiment, a
female avatar representing the building manager would communicate with them. Participants were
told to pay attention to the requests carefully, as they would be asked what the requests were at the
end of the experiment. However, it was noted that it was completely up to the participants whether
to comply or not with the requests, and that they had to act as they would in their own office. While
participants were reading the passages, a request was delivered to them (30 seconds after starting
the task). Two different pro-environmental requests were delivered to each participant: (1) “if I
open the blinds for you to have natural light, would you please dim or turn off the artificial lights?”;
and (2) “if I open the window for you to have a breeze and fresh air, would you please increase
the temperature setting on the thermostat?” The default lighting setting in the office room was the
lights on while the blind was closed (Figure 7a) and the default temperature set point was 73˚ F
(Figure 8a). If participants chose to comply with the pro-environmental request, they had to go to
the light switch and lower the level of lights or turn them off while the blind was open. Likewise,
if participants chose to comply with the other pro-environmental request, they had to go to
thermostat and increase the temperature setpoint. We observed the participants’ compliance with
the requests in the assigned conditions (laptop vs. HMD). In addition, we assessed the participants’
performance by measuring the time that they spent on reading each passage as well the reading
comprehension (number of correctly answered questions).
`
56

(a)
(c)
(b)
(d)
Figure 7 – Different lighting levels in the room: (a) All the lights on while the blind was closed; (b)
dimming the light one level while the blind was open; (c) dimming the light two levels while the blind
was open; (d) turning off all the lights while the blind

(a) (b)
`
57

(c) (d)
Figure 8 – Different temperature settings in the room: (a) 73 degree Fahrenheit; (b) 74 degree Fahrenheit;
(c) 75 degree Fahrenheit; (d) 76 degree Fahrenheit
4.5.3.3. Pest-Experiment Session
After completing the virtual part of the experiment, participants completed the post-experiment
questionnaires. The first questionnaire included three items assessing how participant’s values and
attitudes resembled the environmentalists’ values and attitudes using a five-point scale ranging
from 1 (strongly disagree) to 5 (strongly agree). The second questionnaire assessed participants’
environmental setpoint preferences with 3 items: preferences for lighting sources (daylighting vs.
artificial lighting), lighting levels (amount of lighting), and temperature setpoint. As the last set of
questionnaires, participants completed two questionnaires investigating their sense of presence in
the VEs: the Witmer and Singer Presence Questionnaire (PQ) [194] and another presence
questionnaire developed by Slater, Usoh and Steed [201].
Presence Questionnaire (PQ) was designed by Witmer and Singer (1998) to measure the degree of
presence and engagement that individuals experience in a virtual environment. The PQ
questionnaire has 24 items divided to six subscales designed to assess immersion (e.g., objective
factors, such as visual and auditory aspects of the virtual environment) and individual’s
involvement (e.g., self-motion and degree of interactivity possible) using a seven-point scale
format (1 = Not at all, 7 = Completely). The presence questionnaire six subscales include
`
58

involved/control, natural, resolution, interface quality, auditory, and haptic. We did not consider
haptic in our study, as our VEs did not provide any haptic simulations. Items designed to
investigate involved/control subscale, examined the degree of control an individual has in
performing the task and interacting with the VE, as well as VE’s responsiveness to individuals’
actions. The other five subscales investigated the level of involvement that individuals
experienced. Natural items investigated how natural and consistent with reality the interactions
were. Resolution items examined scene resolution, which is related to the realism of the
environment. The interface quality items investigated if the VE platforms distracted users from the
task they performed and the degree to which users could concentrate on the task [194]. Auditory
items referred to sound and audio aspects of the VE.
In general, the questionnaire developed by Witmer and Singer [194] assesses mostly the immersion
and interaction factors. Although immersion can be a factor that predicts presence, immersion and
presence are different concepts [202,203]. Therefore, we also used Slater, Usoh and Steed
questionnaire [201], which focuses mainly on psychological and behavioral response to immersion
and involvement. This questionnaire has 6 items (each on a 1 to 7 scale) designed to assess the
extent to which individuals compare the virtual environments with a place in the real world and
the extent to which the individuals experience the virtual environment as a place visited rather than
seen computer generated images. In Slater-Usoh- Steed presence questionnaire, the higher score
represents higher level of presence and in analyzing the data, presence was measured by counting
the number of high responses (scores of 6 or 7) in the answers to the six questions [201,204].
Finally, participants completed the Simulator Sickness Questionnaire (SSQ). Participants had also
completed SSQ in the pre-experiment session so that their pre- and post-test scores could be
`
59

compared to identify simulator sickness caused by interacting with VEs. The entire experiment
took about approximately 40 minutes.
4.5.4. Results
In order to analyze the data, we compared the compliance and other dependent measures including
presence, task performance, and simulator sickness between the two types of VE platforms. We
also investigated the effects of other factors, such as participants’ characteristics (i.e., gender and
immersive tendency), could have on the dependent measures.
To examine the effects of the VE platforms on compliance with pro-environmental requests, we
conducted an independent sample T-test (PC laptop vs. HMD) on the number of times participants
chose to comply with the pro-environmental requests. There was no significant differences in the
number of times that participants complied with the request in the PC laptop (M = 1.30, SD = .81)
and HMD (M = 1.20, SD = .83) conditions: t(98) = .61 , p = .54. These results suggest that the
level of immersion does not influence the participants’ decision regarding to compliance with pro-
environmental request. We also checked the possible factors that might have impacted the
compliance including participants’ preferred source of lighting, preferred lighting level and
temperature setpoint, and identification with environmentalists. Using linear regression analysis,
the results showed no main effects of these factors on compliance: preferred source of lighting ( 
= -.05, p = .61), preferred lighting level (  = -.04, p = .67), preferred temperature setpoint (  =
.01, p = .89), and identification with environmentalists (  = -.11, p = .26).
We also investigated the influence of VE platforms on task performance. As a measure of task
performance, we computed the average time to read the passages in two different conditions (PC
laptop vs. HMD). Using an independent sample T-test, the results showed that there was no
`
60

significant differences in the average time that the participants spend reading the passages in the
PC laptop (M = 57.97, SD = 15.63) and HMD (M = 60.04, SD = 17.86) conditions: t(98) = .53 , p
= .25. In addition, we measured participants’ reading comprehension by computing the number of
correct answers to the questions that were asked about the passage. The results showed no
significant difference in participants’ reading comprehension in the PC laptop (M = 7.76, SD =
1.33) and HMD (M = 7.58, SD = 1.37) conditions: t(98) = .81, p = .51. Suggesting that participants
performed in the same way in both conditions.
We examined the sense of presence using the two presence questionnaires. Investigating the effects
of VE platforms on the presence using the questionnaire, developed by Witmer and Singer,
conducting independent sample T-test, the results showed no significant difference in the sense of
presence experienced in the PC laptop (M = 89.10, SD = 15.25) and HMD (M = 91.90, SD =
18.22) conditions: t(98) = -.83 , p = .41. We also examined the influence of two VE platforms on
the participants’ scores for these five subscales of presence, the results showed no significant
effects for these factors (Table 1). In analyzing the data from Slater-Usoh- Steed presence
questionnaire (the number of 6 or 7 answers to the six presence questions), using linear regression
analysis, the results showed no significant difference in sense of presence experienced by different
VE platforms:  = .14, p = .15.
Next, we investigated the interaction between the VE platforms and sense of presence (using
Witmer and Singer’s questionnaire) using linear regression analysis. The results showed no
interaction between the PC laptop (0) vs. HMD (1) dummy-coded variable and sense of present
measuring compliance (  = -.40, p = .56) and performance, both average time completion (  = .21,
p = .75) and reading comprehension (  = .18, p = .79). The same results were observed for Slater-
Usoh- Steed presence questionnaire: no interaction between VE platform and sense of presence
`
61

measuring compliance (  = -.51, p = .15) and performance, both average time completion (  = .24,
p = .51) and reading comprehension (  = -.29, p = .41).
Table 1- Main effects of VE platforms on presence
Dependent
Measure
Measure
PC laptop HMD
T p
Mean sd Mean sd

Presence Subscales
Involved/Control 42.24 7.32 44.02 7.37 -1.13 .26
Natural 10.96 4.32 12.04 4.23 -1.26 .21
Resolution 8.86 2.31 9.04 2.43 -.38 .71
Interface Quality 13.54 3.12 12.40 3.79 1.64 .10
Auditory 13.50 2.79 14.40 3.23 -1.49 .14
Note. †p < .1; *p < .05; **p < .01; ***p < .001.
Results of a mixed ANOVA on participants’ simulator sickness symptoms before and after
exposure to VE as within subjects effect and platform type as between subjects effects revealed
significant interaction effects for simulator sickness (pre vs. post participation) and VE platform
type (F(1, 98) = 8.24, p = .005). Accordingly, we conducted independent samples T-tests to
examine the effect of VE platform type on participants’ level of simulator sickness both before
and after the experiment. The results showed no significant differences in the participants’ level
of simulator sickness before the experiment in both PC laptop (M = 154.32, SD =149.63) and
HMD (M = 166.81, SD = 187.44) conditions: t(98) = -.37 , p = .71. However, there was a
significant difference when comparing the participants’ levels of simulator sickness after the
experiment, such that participants experienced greater motion sickness in the HMD condition (M
= 1073.89, SD = 276.56) than the PC laptop condition (M = 974.76, SD =198.59, t(98) = -2.06 , p
`
62

= .04. Further analyses reveal that this difference motion sickness after the experiment was because
of disorientation. Specifically, we investigated the effects of different subscales of the simulator
sickness and the results showed significant effects on disorientation. However, no significant
effects on nausea and oculomotor problems were found (Table 2).
Table 2 – Simulator sickness effects before and after the experiment
Dependent
Measure
Questionnaires Measure
PC laptop HMD
T p
Mean sd Mean sd

Simulator
Sickness

Before VE

Nausea 8.01 14.47 10.49 15.45 -.83 .41
Oculomotor
Problems
20.16 20.97 20.47 22.15 -.07 .94
Disorientation 13.08 20.55 13.64 25.85 -.12 .91
After VE
Nausea 77.85 14.97 80.90 17.48 -.94 .35
Oculomotor
Problems
74.28 21.05 79.29 25.40 -1.07 .29
Disorientation 111.36 21.60 126.95 37.69 -2.54 .01
Note. †p < .1; *p < .05; **p < .01; ***p < .001.
We also investigated if simulator sickness is correlated with presence. A Pearson correlation was
calculated between the total score for the SSQ questionnaire and presence questionnaires’ scores
and the results showed significant negative correlation between post-participation levels on SSQ
and presence using both Witmer and Singe questionnaire (r = -.23, n = 100, p = .02) and Slater-
Usoh- Steed presence questionnaire (r = -.25, n = 100, p = .01). Pearson correlations conducted
between the questionnaires’ subscales showed significant or marginal significant correlations
`
63

between all subscales of these questionnaires (r < -.16, n = 100, p =< .1) except the Audio subscale
of the presence questionnaire and all subscales of SSQ questionnaire (r > -.06, n = 100, p > .58)
(Table 3). This finding is in line with the results of [205], reporting that sound has no effect on the
severity of simulator sickness. In general, these negative correlations suggest that the more
sickness symptoms participants experience, the more distracted they become from the VE and the
less presence they feel [206]. The analysis also showed that the interaction of VE platform type
and simulator sickness did not have any significant effects on compliance with pro-environmental
requests (  = -.12, p = .81) and performance, both average reading time (  = -.84, p = .1) and
reading comprehension (  = -.29, p = .56).
Table 3 – Correlation between presence sub-scales and SSQ subscales

Note. †p < .1; *p < .05; **p < .01; ***p < .001.

SSQ Subscales Presence Subscales r p

Nausea
Involved/Control -.29 .00
Natural -.19 .05
Resolution -.20 .04
Interface Quality -.31 .00
Auditory -.06 .58

Oculomotor Disturbance
Involved/Control -.29 .00
Natural -.26 .01
Resolution -.20 .05
Interface Quality -.41 .00
Auditory -.01 .88

Disorientation
Involved/Control -.28 .01
Natural -.18 .07
Resolution -.16 .11
Interface Quality -.42 .00
Auditory .02 .84
`
64

We used immersive tendency questionnaire to investigate if participants’ characteristics influenced
the sense of presence they experienced in VEs. We also tested the two experimental groups to
ensure that there was not failure of random assignment (i.e., there were no significant differences
between conditions in participants’ levels of immersive tendency). Using independent sample T-
tests, the results showed no significant differences between participants immersive tendency in the
PC laptop (M = 67.62, SD = 10.54) and HMD (M = 69.76, SD = 11.40) conditions: t(98) = -.97 ,
p = .33. The results also showed no significant differences between participants’ immersive
tendency three subscales (Table 4).
Table 4 – Difference in participants’ immersive tendency subscales
Dependent Measure Measure
PC laptop HMD
T p
Mean sd Mean sd
Immersive Tendency
Subscales
Involvement 42.24 7.32 44.02 7.37 -1.13 .26
Focus 10.96 4.32 12.04 4.23 -1.26 .21
Game 13.50 2.79 14.40 3.23 -1.49 .14
Note. †p < .1; *p < .05; **p < .01; ***p < .001.
Next, we examined if participants’ immersive tendency influenced the presence they experienced
in the VE, as well as their performance and compliance with pro-environmental requests using
linear regression and correlation analyses. We used the presence data collected by Witmer and
Singer’s presence questionnaire as this presence questionnaire and immersive tendency were both
developed by Witmer and Singer based on the same standards. The results showed significant
effects of immersive tendency on the sense of presence, meaning that participants with higher level
of immersive tendency experienced higher sense of presence (  = .26, p = .01). A person, who is
more likely to become immersed in a VE would experience a greater sense of presence while
interacting with a VE. We also calculated a Pearson correlation between the questionnaires’
`
65

subscales and the results showed that ITQ’s Involvement and Focus subscale significantly or
marginally significantly correlated with presence questionnaires’ Involvement and Audio
subscales (r >.26, n = 100, p < .04). ITQ’s Involvement was also significantly correlated with
presence’s Interface Quality subscale (r = -.21, n = 100, p = .04) and ITQ’s focus was significantly
correlated with presence’s resolution subscales (r = .21, n = 100, p = .03). However, no significant
correlation was found for ITQ’ game subscales and any of the presence’s subscales (Table 5).
These results suggested that participants, who had the ability to get deeply involved in an activity
or a stimulus, such as books or movies and showed a tendency to maintain focus on current
activities, were more likely to experience higher presence in the virtual environment and
participants’ tendency to play video games was not an effective factor causing sense of presence.
Table 5 – Correlation between ITQ subscales and presence subscales
ITQ Subscales Presence Subscales r p

Involvement
Involved/Control .26 .01
Natural .11 .26
Resolution .15 .15
Interface Quality -.21 .04
Auditory .21 .03

Focus
Involved/Control .36 .00
Natural .15 .12
Resolution .21 .03
Interface Quality .16 .12
Auditory .34 .00
Involved/Control .05 .62
`
66

Note. †p < .1; *p < .05; **p < .01; ***p < .001.
The effects of immersive tendency on sense of presence were independent of the platform type
and no significant interaction effects were found for platform type and immersive tendency (  =
.5, p = .38). No significant effects were found for immersive tendency on compliance with pro-
environmental request (  = .09, p = .66) and performance: average time (  = .08, p = .45) and
reading comprehension (  = -.06, p = .52). In addition, no significant effects were found for the
interaction of platform type and immersive tendency on compliance with pro-environmental
request (  = -.30, p = .66) and performance, both average time ( = -.53, p = .43) and reading
comprehension (  = -.94, p = .16).
In addition, we used regression analyses to examine whether participants’ gender had any effects
on compliance, presence, and performance or moderated impacts of VE platforms. There were no
main effects or interaction effects of gender and VE platforms on compliance, presence, and
performance (Table 6).
Table 6 – Gender main effects and interaction effects with VE platforms on compliance, presence, and
performance

Game
Natural .01 .88
Resolution -.03 .73
Interface Quality -.04 .65
Auditory -.07 .48
Dependent Measure Between Subject
Measure
Beta t

Involved/Control

-.12 .25
Natural -.04 .70
`
67

Note. †p < .1; *p < .05; **p < .01; ***p < .001.
4.5.5. Discussions and Conclusions
Direct comparison of effectiveness of VE systems (immersive and non-immersive) is complicated
since the differences between such systems may be not only due to the level of immersion, but
also to other factors, such as application context and user characteristics. In this study, we
investigated the effectiveness of two representative VE platforms (i.e., PC laptop vs. HMD) on
examining the compliance to persuasive pro-environmental requests and task performance. We
also explored the causes and consequences of presence, experienced in these environments.
Overall, the results of this study showed no significant differences in the participants’ sense of
presence and performance while interacting with these two VE platforms. The results are
consistent with the findings of other studies, which did not report any differences in the sense of
presence and performance between different VR platforms [207-211]. In addition, there was no
significant differences in participants’ compliance rates in the PC laptop and HMD conditions.
Previous studies did not explore the potentials of using VEs for investigating the effectiveness of
behavior change intervention strategies using social interactions. Our study can guide similar
Presence Subscales Resolution Gender
(Main Effects)

-.00 .99
Interface Quality -.05 .64
Auditory -.08 .42
Compliance
Gender
(Interaction with
platform type)
.20 .32
Presence (Witmer and Singer ) -.20 .33
Presence (Slater-Usoh- Steed) -.06 .76
Performance
(Average Reading Time)

-.01 .97
Performance
(Reading Comprehension)

.20 .31
`
68

studies by comparing the effectiveness of different VE platforms, which vary in the level of
immersion and by investigating the potential factors that impact the effectiveness of these
platforms.
We also investigated the relationship between the occurrence of simulator sickness symptoms and
VE platforms, compliance, and task performance. Results of our study showed HMD-based VE
induced a higher level of simulator sickness. Our results for simulator sickness in the present study
are consistent with those found in previous investigations [196,212]. Among the simulator sickness
subscales, disorientation showed significant effects, which can be justified by the fact that
disorientation symptoms are more dominant of the other symptoms including nausea and
oculomotor problems [213] and remain for some time following immersion [214,215]. Our results
also showed negative correlation between the post-participation levels of simulator sickness and
sense of presence, suggesting that an individual, who experience symptoms of motion sickness in
a VE, becomes distracted from the key aspects of the environment and experience less feeling of
presence [205,206,216]. These results suggest that studies designed with the aim of minimizing
simulator sickness symptoms may take advantage of using a PC laptop over an HMD [196].
We also explored factors, such as individuals’ gender and immersive tendency that could have
influenced the effects of VE platforms on participants' sense of presence, behavior, and
performance indirectly. Investigating how individuals’ characteristics (e.g., gender and immersive
tendency) are related to the effects of VEs, we found no relationship between participants’ gender
and their sense of presence, behavior and performance. These results are also consistent with other
studies reported no correlation between participants’ gender and sense of presence and
performance [217,218]. In addition, we found no relationship between participants’ immersive
tendency and their behavior and performance. On the other hand, the results showed that there was
`
69

a strong relationship between participants’ immersive tendency and the presence they experienced.
The participants with higher immersive tendency experienced higher sense of presence, however,
these effects were independent of the type of VE platforms. These findings are consistent with the
findings of other studies suggesting that an individual who is more likely to become immersed in
a VE will experience a greater sense of presence while interacting with the VE [219-221]. We also
found that individuals’ tendency to become involved in situations and maintain focus on current
activities can predict the presence that they experience in the VE and their previous gaming
experience is not an influential factor.
The results of this study suggested that PC laptop and HMD can be used interchangeably for
investigating the effects of pro-environmental requests on occupants’ pro-environmental
behaviors. However, for studies that experimenters’ presence in the room was necessary, IVE had
the advantage that participants could not see the experimenter and it helped to avoid the assumption
that the experimenter would judge them for their behavior. On the other hand, for the studies that
larger sample sizes with more diversity were required, laptop based VEs had the advantage that
could be embedded in the surveys and sent to a large number of people in different parts of the
U.S. In the following chapters of this dissertation, we demonstrated the results of studies, in which
we investigated the effectiveness of VEs (both PC laptop and HMD) vs. physical environments
for collecting data while exploring the effects of incorporating social and relational features in
design of different components of a behavior change intervention strategy designed to promote
occupants’ pro-environmental behaviors in buildings.
`
70

Chapter 5. Influence of LEED Branding on Building Occupants’ Pro-
environmental Behavior
In this chapter, we presented a study conducted to investigate the effects of an indirect behavior
change strategy on building occupants’ pro-environmental behaviors. Advances in building
technology and science have made it possible to reduce buildings’ negative impacts on the
environment by employing sustainable building practices. However, in spite of the fact that
technological improvements can have an impressive influence on how the built environment
impacts the natural environment, studies have shown that green office buildings consume more
energy in comparison to conventional office buildings of the same function and size [5,6].
In recent years, rating systems and certification programs have been developed in an effort to
provide the architecture, engineering and construction (AEC) industry with the tools necessary to
design and construct high performance green buildings [222]. Among these certification programs,
Leadership in Energy and Environmental Design (LEED) green building rating system is used
increasingly in the United States. LEED was developed by the United States Green Building
Council (USGBC) to provide nationally accepted third-party certification that a building is
designed and built using an approach that reduces or eliminates negative environmental impacts.
LEED provides four levels of certification: certified, silver, gold, and platinum, which are based
on a building’s performance, measured using a set of prerequisites and credits (standards). “Each
higher level of certification represents an incremental step toward integrating different components
of sustainable design, construction, and operation to achieve optimal building performance [223].”
The LEED standards create a specific kind of built environment that consumes minimal resources
and produces minimal waste [224].
`
71

Energy efficient design strategies have been increasingly adopted in office buildings, which make
up a large portion (in terms of gross area) of LEED-certified buildings [225]. Company policies
that geared toward environmental stewardship [226], stringent building codes and government
regulations [227], cost effectiveness, utility incentives, and occupant productivity [228] and
satisfaction [229] are among the reasons that motivates an organization to occupy LEED certified
office buildings. However, paired comparisons of the energy use of LEED certified buildings with
similar conventional buildings showed that although energy performance for LEED-certified
buildings was better on average, many individual LEED buildings performed worse than their
conventional counterparts [230]. Research has shown that the major cause of the
underperformance is the behavior of occupants [231,232]. These buildings might have sustainable
designs but there is no provision for sustainable usage by the occupants. One suggested approach
is for building occupants to change their energy consumption behavior by adopting more energy
efficient practices, such as adoption of natural light instead of artificial light. In previous studies,
researchers have used intervention strategies in green buildings to influence occupants’ behaviors.
For example, occupants’ behaviors were influenced by increasing building occupants’ knowledge
(e.g., knowledge about a building system), which can make the occupants change their behavior
[3]. However, it has not been investigated to date if being in a LEED building influences occupant
behavior in a positive or negative manner.
In this study, we specifically examined the influence of LEED branding, as an indirect behavior
change strategy, on building occupants' lighting preferences and recycling behavior in office
buildings. In addition, we investigated how occupants’ environmental values/views affect the
influence of LEED branding. . LEED branding, in the context of this study, is defined as
`
72

introducing the building as being LEED certified and emphasizing the aspects of the LEED
building’s green features to its users.
5.1. Methodology
To the best of our knowledge, there is no empirical research that has investigated the influence of
LEED (green) branding on energy consumption behavior of building occupants. Therefore, the
objective of this study was to investigate the influence of LEED branding on occupants' lighting
preferences and recycling behaviors in a single occupancy office space. We specifically
investigated if LEED branding influences occupant behavior in a positive way, which means that
occupants conform to the green building’s standards (conformity); or occupants would feel that
the building is handling environmental concerns and therefore it could compensate for their
wasteful behavior and this would result in negative environmental behavior (compensation).
In order to create a successful LEED branding strategy, we focused on the functional attributes as
well as the emotional benefits of the LEED certified buildings [233]. To include the functional
attributes, we delivered information on how LEED certified buildings are designed to reduce the
environmental impacts. In addition to enhancing the influence of LEED branding, we tried to
evoke positive emotions in the occupants by offering information on how LEED certified buildings
contribute to human health and preservation of the natural environment. LEED branding was
implemented by labeling the building as being LEED certified while explaining the LEED
certification procedure, its environmental goals and contributions to the natural environment and
human health, as well as LEED certified building’s standards, sustainable aspects of their design,
construction, and operation.
`
73

We used immersive virtual environments to study LEED branding and its effects on occupants’
behaviors when a building was presented as a LEED certified building since there were some
challenges conducting the experiment in the physical environment. For example, we had to identify
and control a LEED building’s design features and we also needed to collect actual occupant-
related data, for which we needed access to actual LEED certified buildings with specific LEED
features, which was challenging. Furthermore, aforementioned uncontrollable factors, such as
cloudy/sunny weather in different days or different window types, which influence the level of
natural light in a room, might affect the results by causing experimental noise, which is hard to
control in the physical environment.
5.2. Experiment Setup
A virtual model of a single occupancy office was created in an IVE. The modeled office included
the following features: both natural light and artificial light sources, a controllable blind and
controllable lighting fixtures, two trash storage bins, a recycling bin and a regular trashcan, and a
LEED certification signage on the door. The participants were able to interact with the
environment by turning on/off the lights, opening/closing the blinds, and grab and move the
objects.
Participants were randomly assigned to one of the two experimental groups: Group 1 (branding
group): information was provided to the participants about “green buildings” and LEED
certification (its functional attributes and emotional benefits) and the room was introduced as being
part of a LEED certified building; and Group 2 (control group): no information was provided to
the participants about the room being LEED certified or having any green features. The only
`
74

difference between the two groups’ models was the inclusion of LEED certification signage on the
door (LEED group).
We tested the following hypotheses by comparing the participants’ behaviors between these two
groups:
 H1: When placed in a dark room, participants in the LEED branding group would choose
natural light to increase the lighting levels significantly more than the participants in the
control group (conformity).
 H2: When asked to place the scrap paper in a trash can, participants in the LEED branding
group would choose recycling bin significantly more than the participants in the control
group (conformity).
 H3: LEED branding would be more influential on lighting preferences and recycling
behaviors among participants with higher environmental values and views compared to the
participants with lower environmental values and views.
To test these hypotheses, experiments were conducted in IVEs for a period of three months (March
2015 through May 2015). First, a pilot study was conducted with a limited number of participants
to ensure that the model, procedure, and the questionnaires were designed adequately for the
experiment. Based on the findings of the pilot study, we modified the immersive virtual
environment and the questionnaires that were administered after the experiment. We had 50
participants for each condition (n = 100 participants, with 50 in each group). Participants were
undergraduate and graduate students at the University of Southern California. Participants either
voluntarily participated in the experiment or received course credit or incentives (e.g., pizza) for
`
75

their participation. Prior to participating in the experiments, participants were asked to read and
sign a consent form.
5.2.1. Pre-Experiment Session
Prior to the experiment, participants underwent training to become more familiar with IVEs. The
training was done in a different environment than the experimental one in order to avoid any
learning effects and biases that could possibly affect the participants' decision making during the
experiment, as well as to reduce any noise that would influence the experimental results other than
the experiment manipulations (Figure 9). The training included showing the participants how to:
work with the equipment, put the head mounted display on and adjust it to their most comfortable
state; move around in the immersive virtual environment; turn on/off the lights; open/close the
blinds; read text; and grab and move the objects in the space. In addition, we showed them how
the lighting level would be if they open the blinds or turn on the lights, so that participants know
that there would be enough light to read the passage in both settings (natural light or artificial
light). In addition, all participants, both in Group 1 and 2, were told that the study was about how
to link Building Information Models (BIM) to immersive environments to be incorporated in
architecture and engineering courses and the participants input was going to be used to understand
how realistic these virtual environments were compared to the actual and/or BIM environments.

`
76

Figure 9 – Training process: a participant is trained how to navigate and interact within the IVE. The left
image shows the participant navigating through the room, and the right image shows the participant
opening and closing the oven door
Then, the participants were told that they would be placed in the middle of an office room (starting
equal-distant to both the light switch and the blind) and they would have to walk around the desk
to read a passage, which was located on a monitor placed on the desk. Participants were told that
they could ask the experimenters to stop the experiment whenever they experienced motion
sickness. The participants were randomly assigned to one of the two experimental groups: the
branding group and control group. The participants in the branding group were asked to read an
informational text (see Appendix 1) about green buildings and the goals and requirements of the
LEED certification process. The participants in the branding group were also told that they were
going to be immersed in an environment that is a part of a LEED certified building. No further
information about green buildings, the LEED certification process or the virtual office being part
of a LEED certified building was provided to the control group participants.
5.2.2. Experiment Session
After the pre-experiment session, the participants were asked to put back on the head mounted
display so that they could be immersed to the experimental virtual environment. In the virtual
office, the default lighting setting was minimum amount of available lighting, where the
participants were not able to read the passage but could safely go to any location in the room
`
77

(Figure 10a). Through real-time rendering, three other lighting settings were created: (1) a bright
room with all light bulbs on while the blind was closed (Figure 10b); (2) a bright room with the
blind open with none of the light bulbs on (Figure 10c); and (3) a bright room with the blind open
and all of the light bulbs on (Figure 10d). The model was designed to ensure that lighting levels
met the code regulated lighting levels for occupants to read documents in the office building [234].

(a) (b)

(c) (d)
Figure 10 – Different lighting levels in the room: (a) dark room; (b) bright room with all light bulbs on
and the blind closed; (c) bright room with the blind open and all of the light bulbs off; (d) bright room
with the blind open and all of the light bulbs on
Participants were told that they had the following options to increase the lighting levels: turn the
lights on/off (Figure 11a) and open/close the blind (Figure 11b). The order, in which we described
these options randomly varied between the participants to eliminate any bias due to any order
effect. However, to be able to read the passage, the participants had to increase the lighting levels
`
78

in the room choosing from one of the three options. In order to eliminate any effect due to the
variations in lighting levels in different times of day, the experiments were conducted on sunny
and non-cloudy days between 10 AM and 4 PM Los Angeles time to make sure the participants
understand there would be adequate natural light to read the passage if they opened the blinds. The
model was also designed to reflect natural light levels for a sunny Los Angeles day at 12:00 PM
so that all of the participants experienced the same level of natural light during the experiment.
(a) (b)
Figure 11 – (a) turn on/off the lights; and (b) open/close the blind

The duration of the experiment was 20 minutes and it included two tasks. First, the participants
were asked to read a passage (Figure 12a) and they were told that they would be asked to answer
a few questions about the passage at the end of the experiment. Participants' behaviors were
observed in terms of whether they chose to turn on the artificial lights or they just opened the blind
(for natural light) in order to have enough light to read the passage. The second task was to throw
the scrap paper, which was placed on the desk, in in one of the trash storage bins (Figure 12b).
`
79

(a) (b)
Figure 12 – (a) experimental components: passage on computer monitor; (b) scrap papers on the table

We asked the participants to perform the second task after they were done with the first task. The
goal was to observe whether participants would throw the scrap paper in the regular trashcan or in
the recycling bin (the recycling bin was clearly labeled) (Figure 13). When participants completed
this task, they were asked to remove the head mounted display and thanked for their participation.

Figure 13 – Recycling bin and regular trashcan in the experimental setting

5.2.3. Post-Experiment Session
In the post-experiment session, participants were asked to fill out post-surveys. The post-surveys
were composed of six parts: (1) a questionnaire about the passage they read. These questions were
to divert participants' attention from the fact that we were observing their pro-environmental
behaviors. We did not use this data for the analysis; (2) demographics questionnaire (participants'
`
80

age, gender, degree objective, and major, etc.) to investigate the influence of these factors on
participants’ lighting preference and recycling behaviors and assess the underlying reasons behind
the participants' behaviors indirectly; (3) an IVE interaction questionnaire asking participants what
they think about the virtual environment (realism and presence); (4) an environmental value
questionnaire consisted of 15 questions with a 5-point scale, used to assess participants’ pro-
environmental concerns and any potential effect of such concerns on their behavior. Scores on the
scale items varied from a low of 1 (strongly disagree) to a high of 5 (strongly agree), with disagree,
neutral, and agree as interval points, and (5) an environmental view questionnaire consisted of 11
questions with Yes and No as two possible responses. These questions measured participants’
general knowledge about environmental issues, awareness of eco-friendly products and packaging,
and awareness of recycling options. These surveys were designed based on questionnaires used by
Ishaswini et al. [235] to measure general consumers’ environmental concern, knowledge of
environmental issues, awareness of eco-friendly products, trust in performance of eco-friendly
products and green buying behavior. The survey questions were modified to assess factors more
related to building occupants’ behaviors rather than consumers’ buying behaviors. A sample of
these questions are shown in Table 7. The sixth questionnaire assessed if the participants were
familiar with environmental related terminologies. After filling out the post surveys, the
participants were thanked and dismissed.
Table 7 – Sample questions about environmental responsibility
Environmental Values Mean SD
I am concerned about the current environmental state the world is / in
(1 - Strongly disagree, 5-Strongly agree)
3.94 0.81
`
81

I always turn the lights off when I leave a room in my house/the /
place I work
(1 - Strongly disagree, 5-Strongly agree)
4.32 0.69
Environmental Views Frequency
(Yes)
Frequency
(No)
Have you ever sent old or broken household electronic products to be
recycled?
(1- Yes, 2 - No)
18% 82%
Do you purchase correct from the environmental point of view?
(1- Yes, 2 - No)
27% 73%
5.3. Results and Data Analysis
We tested the research hypotheses by performing statistical analysis on the data collected from the
participants in the branding and control groups. Chi-squared goodness of fit test ( 
2
), t-test (t),
ordinary least square (OLS) linear regression, and multinomial logistic regression were used to
analyze the data. The results of chi-squared goodness of fit test was reported with degrees of
freedom and sample size in parentheses, the Pearson chi-square value ( 
2
) and the significance
level (p), ( 
2
(d.f, N) , p). The results of t- test are reported by t statistic and the significance level
(p), (t (d.f), p). Ordinary least square (OLS) linear regression output was reported with un-
standardized slope ( ), with the t-test (t) and the corresponding significance level (p), ( , t(N- k -
1), p , k equals the number of predictor variables) [236] The analyses are based on alpha (  equals
to 0.05. In order to help interpret the results, p-values (p) are provided below and the null
hypotheses (H0) are rejected when p ≤ alpha.
`
82

The results presented here are based on the 100 participants’ (n =100) data (50 participants in each
group with equal distributions between genders: 25 females and 25 males). Participants were
undergraduate students (66%), master students (28%), and PhD students (6%) at the University of
Southern California enrolled in different majors including: engineering (28%); psychology (28%);
architecture (17%); and other majors (27%) including business, cinematic arts, communication,
preventive medicine, etc. The majority of the participants were 18 to 29 years old (71%) and the
rest were around 30 to 39 years old (29%).
In order to ensure that the virtual environment was an adequate representation of the physical
environment, at the end of the experiments, we asked the participants how realistic the virtual
environment was and how easy it was to perform the tasks in the virtual environment. The results
showed that the participants rated the realism of the environment to be realistic on average (M =
3.89, SD = 0.42) based on a five-point Likert scale. 85% of the participants rated the virtual
environment to be realistic or very realistic and only 15% of the participants rated it as neutral. In
addition, when participants were asked how easy it was to interact with the virtual environment,
they rated it to be somewhat easy on average for both the blind (M = 6.01, SD = 0.77) and lights
(M = 6.09, SD = 0.67) based on a seven-point Likert scale (Table 8). 93% of the participants who
opened the blind and 98% of the participants who turned on the artificial lights believed that it was
slightly, somewhat or extremely easy to open the blind or turn on the lights in the virtual
environment and the remaining participants rated it to be neutral.
Table 8 - IVE interaction questions
Questions Mean SD
How realistic was the virtual environment?
(1 Not realistic, 5 Very realistic)
3.89 0.42
`
83

How easy or difficult was it to open the blind in the virtual office?
(1 Extremely difficult, 7 Extremely easy)
6.01 0.77
How easy or difficult was it to turn on the lights in the virtual office?
(1 Extremely difficult, 7 Extremely easy)
6.09 0.67

For investigating the influence of LEED branding on participants’ lighting preferences and
recycling behaviors, we observed the participants’ choice of natural light vs. artificial light in order
to increase the lighting level in the room. It is important to note that in our analysis, we assumed
the behavior is not environmentally conscience if the participant chose to use a combination of
natural and artificial light at the same time. Therefore, herein after, when we analyze the data for
artificial light use, we include the choice of artificial light even if the blinds were open at the same
time (Table 9).
Table 9 - Branding group vs. control group
(number of the participants in terms of choice of natural light vs. artificial light)
Group Artificial light Artificial + natural light Natural light
Branding 10 15 25
Control 20 21 9

The results showed that, in the branding group, the percentage of participants who chose artificial
lights and chose both natural and artificial lights (50%) were the same as the percentage of the
participants who solely chose the natural light (50%). However, in the control group, the
percentage of participants who turned on the artificial lights or both opened the blinds and turned
on the artificial lights (82%) were significantly more than the participants that just opened the
blinds to increase the lighting levels in the room (18%). Statistical analysis, using a chi-squared
`
84

goodness of fit test, showed that the percentage of participants who chose just natural light in the
branding group was significantly more than the percentage of participants who made this choice
in the control group ( 
2
(1, N = 100) = 11.408, p = .001 < 0.05). Therefore, the result was
significant at p < 0.05, accepting Hypothesis 1, which suggested that the participants in branding
group would choose natural light to increase the lighting levels significantly more than the
participants in the control group (conformity). Table 10 shows more details of the analysis.
Table 10 – Branding group vs. control group ( 
2
analysis - natural vs. artificial light)

In addition to the lighting preference, we investigated the impact of LEED branding on
participants’ recycling behavior. In this study, recycling was used as an example of pro-
environmental behaviors. Participants’ choices of use of recycling bin vs. regular trashcan for
placing the scrap paper were observed. The results showed that the percentage of the participants
who used recycling bin in the branding group (92%) was more than the percentage of the
participants who used recycling bin in the control group (76%). The results of the statistical
analysis, using a chi-squared goodness of fit test, showed that the percentage of the participants
chose recycling bin in the branding group was significantly more than the percentage of the
participants who did the same in the control group ( 
2
(1, N = 100) = 4.762, p = 0.029 < 0.05).
Therefore, the result was significant at p < 0.05, accepting Hypothesis 2, which suggested that the
participants in the branding group would choose recycling bin to place scrap paper significantly
Experimental Group Natural light Artificial light Total
Branding group 25 (17) 
2
= [3.76] 25 (33) 
2
= [1.94] 50
Control group 9 (17) 
2
= [3.76] 41 (33) 
2
= [1.94] 50
Total 34 66 100
`
85

more than the participants in the control group (conformity). Table 11 shows more details of the
analysis.
Table 11 – Branding group vs. control group
( 
2
analysis (Recyclable trashcan vs. regular trashcan)

We also examined the effects of the participants' environmental values and views, and demography
characteristics on their lighting preference and recycling behaviors. We first checked the collected
data to make sure that participants were distributed evenly in the control and branding groups in
terms of gender, age, degree objectives (e.g., undergraduate, master, PhD), major (e.g., engineering
psychology, architecture, etc.), and environmental values and views. The statistical analysis
confirmed that there was no significant difference between the two experimental groups in terms
of aforementioned factors. Based on data collected from the personality test, and demographic
characteristic surveys using linear regression, the results showed that there were no significant
effects between participants’ demographic characteristics and their lighting preference and
recycling behavior in both the branding and control groups.
Linking the data collected from the environmental values and views surveys to the experiment
results by using simple linear and logistic regression analyses with environmental values
predicting each dependent variable (DV), we found that among participants in the total sample,
environmental values influenced their lighting preferences (= 0.015, t(98) = 2.160, p = 0.033 <
Experimental Group Recycling Bin Regular trashcan Total
Branding group 46 (42) 
2
= [0.38] 4 (8) 
2
= [2] 50
Control group 38 (42) 
2
= [0.38] 12 (8) 
2
= [2] 50
Total 84 16 10
`
86

.05) and recycling behaviors (  = 0.012, t(98) = 2.175, p = 0.032). In addition, significant effects
were observed between participants’ environmental views and their lighting preferences (  =
0.059, t(98) = 2.272, p = 0.025 < .05) and recycling behaviors (  = 0.061, t(98) = 3.075, p =
0.003 < .05). Participants with more favorable environmental views and values were more likely
to open the blinds to increase the lighting levels in the room or place the scrap paper in the recycling
bin. Knowing that participants' values and views influenced their lighting preferences and
recycling behaviors, we investigated if these values and views would influence the effects of
branding on participants' lighting preferences and recycling behaviors. To do this, we ran separate
regression analyses using environmental values and views as predictors to predict participants'
lighting preferences and recycling behaviors in the branding and control groups, respectively. The
results showed that there were significant effects between participants' environmental values and
their lighting preferences (  = 0.26, t(98) = 2.481, p = 0.017) in the branding group. Predicting
our key DVs from participants’ environmental views (rather than environmental values), the
results showed significant effects between participants' environmental views and their lighting
preferences (  = 0.095, t(98) = 2.676, p= 0.010 < .05) and the recycling behaviors (  = 0.055, t(98)
= 2.898, p= 0.006 < .05) in the branding group. On the other hand, no significant effects were
found between participants environmental values and their lighting preferences (  = 0.002, t(98)
= 0.258, p= 0.798 > .05) in the control group. The same results were found when examining the
influence of the participants' environmental views on their lighting preferences (  = - 0.008, t(98)
= - 0.229, p = 0.820 > .05) and recycling behaviors (  = 0.058, t(98) = 1.618, p = 0.113 > .05) in
the control groups. These results showed that in the branding group, participants with higher
environmental values and views were more likely to open the blind and participants with higher
environmental views were more likely to recycle the paper but there were not any significant
`
87

effects between environmental values and views and lighting preference and recycling behavior in
the control group confirming that participants' environmental values and views influence the
effects of branding (Hypothesis 3). These findings might also suggest that LEED branding made
participants' environmental values and views salient to them, leading them to act more pro-
environmentally (conformity), while we cannot see these effects of environmentalism on behaviors
of the participants in the control group.
At the end of the experiment, we asked the participants in the branding group if they thought being
in a virtual LEED certified building affected their choices. 43 participants from the branding group
answered this question. The results showed that 18% of the participants in the branding group
believed that being in a LEED certified building affected their behaviors in terms of opening blinds
or turning on/off the lights. In addition, 44% of the participants in the branding group said that
being in a LEED certified building affected their behaviors in terms of recycling the scrap paper.
A chi-squared goodness of fit test showed that there was no significant difference between
participants' choices in terms of opening the blinds or turning on the artificial lights to increase the
lighting levels of the room ( 
2
(1, N = 43) = 1.010, p = 0.315 > 0.05) and participants’ acts of
recycling ( 
2
(1, N = 43) = 3.491, P = 0.062 > 0.05) between those who believed being in a LEED
certified building affected their behavior and those who did not.
We also investigated if participants' familiarity with environmental related terminologies, such as
EPA (Environmental Protection Agency), energy star, and WSE (Waste Stream Evaluations) had
any influence on participants' lighting preference and recycling behavior. Linear regression
analysis results revealed that there were significant effects between how familiar participants were
with different environmental related terminologies and their choices for natural vs. artificial light
to increase the lighting levels of them room (  = 0.089, t(98) = 2.176, p = 0.032 < .05) and
`
88

recycling bin vs. regular trashcan to place the scrap paper (  = 0.085, t(98) = 2.726, p = 0.008 <
.05). These results showed that the participants who had more knowledge about environment
related terminologies were more prone to use natural light and recycle the paper. These results
may suggest that the level of building occupants’ knowledge surrounding environmental issues
and their solutions have positive correlations with their electricity and waste reduction activities.
5.4. Discussions
In this study, we investigated the influence of LEED branding on building occupants' lighting
preferences and recycling behaviors. The results showed that LEED branding promotes occupants’
pro-environmental behaviors. The participants chose natural light significantly more than the
artificial light in the branding group compared to the control group, suggesting that LEED branding
influences occupants’ lighting preferences positively, potentially motivating electricity
consumption reduction. The results also revealed that participants in the branding group chose
recycling bin significantly more than regular trashcan in comparison with the participants in the
control group, suggesting that LEED branding influences building occupants’ recycling behaviors
positively, potentially motivating behavior that is towards waste reduction. Therefore, we can
conclude that branding a building as a LEED certified building could result in conformity and not
compensation. These results suggest that it is possible to motivate LEED certified building
occupants toward adoption of more pro-environmental behaviors by letting occupants know that
the building is LEED certified and provide them with knowledge about LEED certification, its
functional attributes and emotional benefits. For example, placing the LEED certification signage
at a place that all of the occupants could see the signage and provide some information about the
LEED certification next to the signage. There might be different ways, such as posters, pamphlet,
or virtual displays.
`
89

We also explored factors, such as individuals’ demographic characteristics, and environmental
values and views that could have influenced participants' behaviors indirectly. Investigating how
individuals’ demographic characteristics, such as age, gender, education, major is related to their
behaviors, we found no relationship between participants’ demographic characteristics and their
energy consumption and environmental behaviors. On the other hand, the results showed that there
was strong relationship between participants’ environmental values and views and the influence
of LEED branding on their choice of natural vs. artificial light and recycling bin vs. regular
trashcan. The participants with higher environmental values and views were more likely to adopt
pro-environmental behaviors under the influence of LEED branding. Our results also suggested
that increasing building occupants' knowledge about environmental issues might improve their
pro-environmental behaviors.
5.5. Conclusions
Although in green buildings, technological improvements and sustainable design have been
adopted, they are only successful when implemented correctly by the occupants of a LEED
certified building; therefore, there is a need to motivate occupants toward adoption of more
sustainable behavior. However, the review of the previous studies shows that there is no consensus
among scholars whether working in green buildings motivates occupants’ pro-environmental
behavior (conformity) or wasteful behavior (compensation). In this study, we investigated the
influence of LEED branding on building occupants' lighting preferences (natural light vs. artificial
light choices) and recycling behaviors (their use of a recycling bin vs. a regular trashcan to place
the scrap paper). The results showed that LEED branding motivated the participants to use natural
light more (potentially reducing electricity consumption) and recycling scrap paper more
(potentially reducing waste generation). Therefore, the results showed that branding a building as
`
90

a LEED certified building resulted in conformity, not compensation. In addition, the results
indicated that there were significant effects between participants’ environmental values and views
and the influence of LEED branding. LEED branding was more influential on lighting preferences
and recycling behaviors of the participants with higher environmental values and views compared
to the participants with lower environmental values and views.
This chapter addresses Research Question 1-1a: How does LEED branding influence building
occupant’s lighting preferences (natural light vs. artificial light) and recycling behaviors in
buildings? And research Question 1-1b: How do environmental values and views affect the
influence of LEED branding on lighting preferences and recycling behaviors?

`
91

Chapter 6. Exploring the Effects of Social Influence Methods
Results of the previous chapter showed that LEED branding promoted occupants’ pro-
environmental behaviors in buildings. In fact, LEED branding indirectly influenced underlying
determinants of occupants’ behaviors such as environmental attitudes. LEED branding made
occupants’ environmental values and views salient to them, leading them to act more pro-
environmentally. In the following chapters of the dissertation, we investigated the effects of direct
behavior change strategies, in which building directly communicates with occupants and tries to
influence their behaviors. In the current chapter, we investigated the effects of incorporating social
features into the human-building communication by adapting behavior-change tactics more
commonly seen in face-to-face communication. Social features represent the ways humans interact
with each other. Research on communication strategies in the energy domain has yet to explore a
number of social features that might promote pro-environmental behavior in buildings. Social
features investigated in this chapter include social influence strategies, such as making a direct
request and other common compliance gaining techniques including reciprocity and foot in the
door (FITD) [139,237,238]. In the present research, we adapted these social strategies to
communications designed to promote energy conservation behavior. We investigated whether
building occupants would adapt pro-environmental behaviors in response to pro-environmental
requests including several classic social influence techniques. In this study, we focused on adaption
of natural lighting than artificial lighting as a pro-environmental action as lighting is the building
system influenced the most by user behavior in office buildings and have significant impacts on
electricity consumption [30,239].
`
92

6.1. Methodology
The objective of this study is to investigate the effects of classic compliance gaining techniques
involving foot in the door and reciprocity on promoting pro-environmental behavior through
influence appeals. The aim of these requests is to encourage building occupants to perform pro-
environmental behaviors, in particular to use natural light instead of artificial light.
We examined the effectiveness of two classic compliance-gaining strategies: (a) foot-in-the-door
and (b) reciprocity [240,241] and then compared them to (c) a direct request, in which participants
were directly asked to engage in the desired behavior. The study included three experimental
groups, which are shown in Table 12. The phrasing of these requests differed between groups, but
they all included the same energy saving tip leading to energy conservation, i.e., opening the blinds
for natural light and turning off artificial lights.
Table 12 – Requests delivered to participants in different groups
Social Influence Method Request
Group 1: direct request
“Could you please do I a favor and open the blinds and turn off the
artificial lights?”
Group 2: reciprocity
"Could you please do I a favor and turn off the artificial lights if I
open the blinds for you?”
Group 3: foot-in-the-door
“Could you please do me a favor and open at least one of the blinds?”
and then “Could you please do me a favor and open the other blinds
and turn off the artificial lights?”

Participants were randomly assigned to one of the three groups. Based on the group that they were
assigned, the participants were given the request through text and given an opportunity to comply
with the request. We tested the following hypotheses by comparing the participants’ compliance
relating to their energy-use behaviors between these three groups:
`
93

H1a: Reciprocity technique engenders more compliance with the pro-environmental requests
compared with the direct request.
H1b: FITD technique engenders more compliance with the pro-environmental requests compared
with the direct request.
H2: Participants’ personality traits will engender different patterns of compliance with the appeals.
We also examined possible carryover effects between the behavioral changes in the virtual part of
the experiment and the real environment by testing the following hypothesis:
H3: Participants who complied or did not comply with the request in the virtual environment will
act in the same way in the real environment.
6.2. Experiment Setup
To test these hypotheses, a virtual model of a single occupancy office was created in an IVE. The
modeled office included the following features: light fixtures, three windows with blinds; a digital
screen; a blackboard next to the screen, on which the requests would appear; a chair; and a
conference table. An experiment was conducted over a period of three months (March 2015
through May 2015). Prior to running any experiments, the study was approved by the Institutional
Review Board (IRB). First, a pilot study was conducted with 21 participants to ensure that the
model, procedure, and the questionnaires were designed properly for the experiment. Based on the
findings of the pilot study, we modified the model, as well as the questionnaires that were
administered before and after the experiment, and conducted power analysis to determine the
appropriate sample size. The results of the power analysis showed that 108 participants in total
were needed to have close to 80 percent power to detect statistically significant effects in
differences between the three conditions. We tested 50 participants for each condition (n = 150
participants) resulting in 88 percent power. Participants were undergraduate and graduate students
`
94

at the University of Southern California. They participated in the experiment voluntarily or
received credit or incentives (i.e., pizza) for their participation.
6.2.1. Pre-Experiment Session
Prior to being immersed into the virtual environment, the participants were asked to complete a
personality questionnaire assessing the Big Five Personality Traits (personality test) [242] in order
to investigate if personality affected reactions to the requests. A personality test, usually involves
a standardized series of questions or tasks, used to describe or evaluate a subject's personality and
characteristics. This “Big five” theory incorporates five different variables (personality traits) into
a conceptual model for describing personality. The five personality traits include: extraversion
(how social and assertive a person is); agreeableness (how cooperative and trusting, a person is);
conscientiousness (how responsible and dependable, a person is); neuroticism (how anxious or
prone to depression, a person is); and openness (how imaginative and intellectual, a person is)
[242].
After completing the pre-test survey, the participants underwent training to become more familiar
with the virtual environment. They were instructed how to put on the head mounted display, adjust
it to a comfortable state, and how to work with the different buttons on the controller. During the
training, participants were immersed in the virtual environment and instructed with different tasks,
such as moving around the room, turning on and off lights, and opening and closing blinds (Figure
14). These assigned tasks were similar to what the participants had to perform during the actual
experiment. In addition, participants were asked to test different levels of lighting in the room. In
this way, they could experience how much light would be in the room if they choose to open the
blinds or turn on the lights. In addition, participants could see that there was no specific view
`
95

behind the windows (it was just the blue sky). Thus, any desire to see the view did not affect the
results.

Figure 14 – A participant during training
After the training, participants heard the following explanation: "You will be immersed in an office
room and you will be asked to sit on a chair and watch a video, which is a TED talk. Please watch
the video very carefully. You will be asked to answer questions about the video after the
experiment. You have the option to pause or resume the video whenever you want. During the
experiment, the building might communicate with you through the blackboard across from you
and request a favor. Please read the request carefully. You will be asked what the request was at
the end of the experiment. However, it is completely up to you whether or not to comply with the
request. Please try to act exactly as you would if you were in your office. Please let us know if you
experience any motion sickness so we can stop the experiment immediately."
6.2.2. Experiment Session
After the pre-experiment session, the participants were asked to put on the head mounted display
so that they could be immersed in the virtual environment. In the virtual office, the default lighting
`
96

setting was all the simulated artificial lights on and all the blinds closed (Figure 15a). Through
real-time rendering, four other lighting settings were created: (1) a bright room with all light bulbs
on while one of the blinds (left blind) was open and the other two were closed (Figure 15b); (2) a
bright room with all light bulbs on while one of the blinds (middle blind) was open and the other
two were closed (Figure 15c); (3) a bright room with all light bulbs on while one of the blinds
(right blind) was open and the other two were closed (Figure 15d); and (4) a bright room with all
light bulbs off while all the blinds were open (Figure 15e).

(a) (b) (c)
(d) (e)
Figure 15 – Different lighting levels in the room: (a) all light bulbs on and all blinds closed; (b) all light
bulbs on and the left blind open; (c) all light bulbs on and the middle blind open; (d) all light bulbs on and
the right blind open; (e) all light bulbs

`
97

Participants entered the virtual room, sat on the chair, and watched a video for 4 minutes and 43
seconds on the digital screen (Figure 16a). While participants were watching the video at the 90th
second, they were given a request in text, which was shown on the blackboard next to the digital
screen and accompanied by a beeping sound (Figure 16b). Participants were exposed to different
requests according to their assigned experimental condition and were given an opportunity to
interact with the environmental room (e.g., adjusting lighting by opening the blinds and turning
off the lights). In group 1, participants were asked to turn off the lights and open all of the blinds.
In group 2, the participants were asked to turn off the artificial lights (the building had already
opened the blinds for them, in order to establish reciprocity). In group 3, the participants were first
asked to open one of the blinds and if they complied, then they were asked to open the other blinds
and turn off the artificial lights.
(a) (b)
Figure 16 – How request was delivered to the participants in the immersive virtual environment: (a) a
participant watching a video in the immersive virtual environment; (b) Request delivered to the
participant while watching the video

`
98

We assessed whether participants complied with the buildings’ requests about energy conservation
behavior. All of the participants had the ability to pause the video, comply with the request, and
continue watching the video. All experiments were conducted on sunny and non-cloudy days
between 10 AM and 4 PM Los Angeles time to make sure that there were no effects due to
variations in lighting levels in the physical and virtual environment at the times that the experiment
was conducted. For example, if the experiment was conducted at night when the sky was dark or
when the outside weather was gloomy or rainy, participants might have an unconscious intuition
that if the blinds were to be open in the virtual space, they would not be able to get enough natural
light (regardless of how the IVE is designed).
6.2.3. Carry over effect
When participants completed watching the video in the virtual environment, they were asked to
remove the head mounted display. Then they went to a physical office room to take the post-test
survey. The lighting conditions in the physical office room were the same as in the virtual office
room (all artificial lights on/ blind closed). As the participants were completing the survey, the
computer screen next to them displayed the same text requests that were delivered to them in the
virtual environment in groups 1, 2, or 3, accompanied by the same beeping sound. We observed
participants' compliance/noncompliance in the physical office room to determine if their behavior
in the virtual environment carried over to the real environment (Figure 17).

`
99

(a) (b)
Figure 17 – How request was delivered to the participants in the physical office environment: (a) a
participant completing surveys in the physical office; (b) request delivered to the participant while
completing the survey
6.2.4. Post-Surveys
In the post-experiment session, participants filled out post-surveys. The post-surveys were
composed of four parts: (1) questionnaire about participants' age, gender, degree objective, and
major, etc.; (2) an IVE interaction questionnaire assessing perceptions of the virtual environment
(realism); (3) a survey to assess participants' intentions to use a similar system and behave in
similar fashion, if it were employed in their building, investigate the reasons behind their decisions,
and solicit suggestions for potential improvement; and (4) a questionnaire to assess participants'
environmental values (environmental concerns and daily environmental habits, such as turning
personal electronics off when not in use, etc.) and views (attitudes about recycling and usage of
eco-friendly products) adapted from Ishaswini et al. [235] measuring general environmental
concerns and knowledge of environmental issues. They were modified to assess factors related to
building occupants’ behaviors rather than consumers’ buying behaviors. After filling out the post
surveys, the participants were thanked and dismissed.
`
100

6.3. Results, and Discussion
We tested the research hypotheses by performing statistical analysis on the data collected from the
participants in the three groups. Chi-squared goodness of fit test ( 
2
), McNemar Chi-square test
( 
2
), ANOVA, t-test, and ordinary least square (OLS) linear regression were used to analyze the
data. The results of chi-squared goodness of fit test and McNemar Chi-square test were reported
with degrees of freedom and sample size in parentheses, the Pearson chi-square value ( 
2
) and the
significance level (p), ( 
2
(df, N) , p). The results of ANOVA are reported by between-groups
degrees of freedom and the within-groups degrees of freedom as well as the F statistics and the
significance level (p), (F (df), p). The results of t-test were reported by t statistics and the
significance level (p), (t (df), p). Ordinary least square (OLS) linear regression output was reported
with unstandardized slope ( ), with the t-test (t) and the corresponding significance level (p), ( ,
t(N- k -1), p), k equals the number of predictor variables [236]. The analyses are based on alpha
(  equals to 0.05. In order to interpret the results, p-values (p) are provided below and the null
hypotheses (H0) are rejected when p ≤ alpha. The results are also considered as marginally
significant when the p value is larger than 0.05 and smaller than 0.1 (0.05< p <0.1).
The results presented here are based on 150 participants (89 males and 61 females). Participants
were distributed equally between the groups (50 participants in each group). Participants were
undergraduate and graduate students at the University of Southern California enrolled in different
majors including: engineering (55%); psychology (26%); and other majors (19%) including
business, cinematic arts, communication, preventive medicine, etc. The majority of the participants
were 18 to 29 years old (91%) and the rest were around 30 to 39 years old (9%).
`
101

Using the IVE as a tool to collect data, one of our main concerns was whether the virtual
environment was an adequate representation of the physical environment. Therefore, we evaluated
the participants' sense of the realism of the environment through the IVE interaction questionnaire.
The results showed that participants rated the realism of the environment to be realistic on average
(M = 3.79, SD = 0.69) based on a five-point Likert scale (1-very unrealistic, 2- unrealistic, 3 -
neutral, 4- realistic, and 5-very realistic). The majority of participants (81%) rated the virtual
environment as realistic, 12% rated it to be neutral, and the rest (7%) believed that it was
unrealistic.
In order to test H1a and H1b, which suggested that the use of reciprocity or FITD engenders more
compliance than the direct request, we compared the participants' compliance with the pro-
environmental requests in the three experimental groups: (1) direct request; (2) reciprocity; and
(3) foot-in- the-door. Our analysis showed that reciprocity generated a significantly higher rate of
compliance (82%), followed by the direct request (62%), and FITD (50%). Statistical analysis,
using a chi-squared goodness of fit test, suggested accepting H1a, given that the percentage of
participants who complied with reciprocal requests was significantly greater than the percentage
of participants who complied with direct requests ( 
2
(1, N = 100) = 4.96, p = 0.026 < 0.05). Table
13 shows details of the analysis. In addition, the percentage of participants who complied with
FITD was not significantly different than the percentage participants who complied with a direct
request ( 
2
(1, N = 100) = 1.461, p = 0.23 > 0.05), rejecting H1b.

`
102

Table 13 – Direct request vs. reciprocal requests vs. FITD in virtual environment
( 
2
analysis - compliance vs. non-compliance)

We also examined the effects of the participants' personality and environmental values and views
on their compliance with the pro-environmental requests. We first checked the collected data to
make sure that participants were distributed evenly in the three experimental groups and found no
significant difference between the three experimental groups. Analyses investigating how
individuals’ environmental values and views are related to compliance revealed no significant
effects between environmental values views and their compliance. We tested the relation between
personality traits and the experiment results by using simple linear and logistic regression analyses
(personality traits predicting the dependent variable, compliance vs. non-compliance).
Participants’ level of openness (which refers to the degree, to which participants were open to new
experiences/new ways of doing things) and neuroticism (which refers to degree, to which
participants were anxious or impervious to what's going on around them) influenced their
compliance with the request. In general, participants with higher levels of openness were more
likely to comply with the requests (  = 1.047, t(148) = 2.745, p = 0.007 < 0.05). In addition,
participants with higher levels of neuroticism were (marginally) more likely to not comply with
the requests (  = - 0.664 t(148) = -1.686, p = 0.094 < 0.1). These results are consistent with the
findings in the health care domain suggesting that high neuroticism is significantly associated with
Groups Compliance Non-compliance Total
Direct request 31 (32. 33) 
2
= [0.05] 19 (17.67) 
2
= [0.10] 50
Rule of reciprocity 41 (32. 33) 
2
= [2.32] 9 (17.67) 
2
= [4.25] 50
FITD 25 (32. 33) 
2
= [1.66] 25 (17.67) 
2
= [3.04] 50
Total 97 53 150
`
103

non-compliance [243-245] and that openness to experience is positively associated with
compliance [246]. In addition, studies conducted by Milfont and Selby investigating how
personality traits are associated with environmental engagement showed that environmental values
and engagements are associated with openness and lower neuroticism [247].
Given that participants' personality traits influenced their compliance with the requests, we
investigated if these personality traits would influence the effects of different social influence
tactics. To do this, we ran separate regression analyses using personality traits as predictors of
participants' compliance with the request in the three experimental groups. The results revealed
that participants with higher levels of neuroticism were more likely to not comply with the direct
request (  = -1.776, t(98) = -2.285, p = 0.027 < 0.05). This could reflect that highly neurotic
individuals experience a wide range of negative emotions [248], leading to non-compliance. Use
of compliance-gaining techniques might mediate these effects, as results show no association
between neuroticism and non-compliance in the two other experimental groups. Our analysis also
showed significant relations between participants' level of openness and their compliance with the
request involving reciprocity (  = 1.736, t(98) = 3.303, p = 0.002 < 0.05). Individuals who scored
higher in openness, who are tolerant of new experiences and ideas and have active imagination,
might become more engaged with a system, in which buildings are communicating with them and
do a favor and ask something in return. These results confirm H2, suggesting participants’
personality traits would influence participants’ compliance with the social requests.
Participants were also asked to identify the reasons behind their compliance or non-compliance
with the request. Conducting the chi-square test of independence, the results showed no significant
differences between participants’ reasons to comply or not comply with the request among the
three experimental groups. The majority of participants in all of the groups indicated that they
`
104

complied because they did not feel like saying no. They also stated other reasons, mostly referring
to simplicity of the favor that they did not mind doing or having no particular preference in terms
of lighting. In addition, the main reasons for non-compliance were preference for artificial light,
not liking the way that the request was made, and consideration of the favor as an extra effort.
Although the majority of participants in the direct request and FITD groups (26% of the
participants who did not comply in the direct request and 28% of the participants who did not
comply in FITD) mentioned not liking the way that request was provided as the reason for their
non-compliance, none of the participants in the reciprocity condition mentioned this as a reason
for non-compliance. In addition, a large portion of participants in the direct request group and the
FITD group (36% of the participants who did not comply in the direct request and 32% of the
participants who did not comply in FITD) stated that the reason for their non-compliance was that
the request was an extra effort. However, fewer participants in the reciprocal requests group 2
(20% who did not comply in group 2) considered the request an extra effort. Other reasons for
noncompliance mentioned by the participants were mostly related to their desire to focus on the
video and not liking interruptions or multitasking.
Investigating the participants’ compliance in the physical environment, collecting data from 45
participants out of 50 participants in each group, our analysis showed that reciprocity generated a
significantly higher rate of compliance (73%), followed by the direct request (56%), and FITD
(47%). Statistical analysis, using a chi-squared goodness of fit test showed that the percentage of
participants who complied with reciprocal requests was significantly greater than the percentage
of participants who complied with direct requests ( 
2
(1, N = 90) = 0.86, p = 0.035< 0.05) and with
FITD ( 
2
(1, N = 90) = 4.85, p = 0.028 < 0.05). In addition, the percentage of participants who
complied with a direct request was not significantly different than the percentage participants who
`
105

complied with FITD ( 
2
(1, N = 90) = 1.67, p = 0.19 > 0.05). We also examined the possible
carryover effects, by which behavioral changes in the virtual part of the experiment were
transferred to the real environment. Using McNemar's test for paired nominal data, no significant
differences in individuals’ compliance emerged between the virtual room and physical room in
general ( 
2
(1, N = 135) = 80.011, p = 0.118 > 0.05). Thus, participants, who complied or did not
comply with the requests in the virtual environment acted in the same way in real life. This suggests
accepting H3, in which virtual environments provide an excellent venue for testing real-world
behavior change interventions.
We also asked the participants to rate their intentions to use a similar suggestion system, if it were
employed in their buildings in the future. The results are presented in Figure 18. We wanted to
understand how similar systems could be employed in the design of future buildings and operation
of existing buildings. Using ANOVA, the results showed that there were significant differences
between the participants who complied with the request and who did not comply (F (1, 148) =
8.635, p = 0.004 < 0.05). Participants, who complied with the requests, were more likely to use a
similar suggestion system than participants, who did not comply with the requests. Participants,
who did not comply, were more neutral about using the suggestion system. Most of the
participants, who did not comply with the request, said that they were likely to use a similar
suggestion system, mentioned being an extra effort and interruption as the reasons for not
complying with the request. Taking advantage of automation (which we implied indirectly to some
extend in group 2) could decrease the level of the effort and increase the compliance. The
participants also suggested some modifications to the system, such as using another modality like
voice to deliver the requests. Thus, there are potential ways to improve the suggestion system to
motivate those who did not comply with the requests.
`
106

Figure 18 – Participants’ intentions to use a similar suggestion system if it was employed in their building
in the future
In addition, the participants were asked if they felt that they would have responded in a similar
fashion to the buildings’ request if it were employed in their own office. The results are presented
in Figure 19. Sixty-two percent of the participants said that they would respond in similar fashion,
13% said that they would not respond in similar way and 25% were neutral. The results showed a
significant difference between participants who complied with the request and those who did not
comply in terms of their tendency to respond in the similar fashion (F (1, 148) = 5.466, p = 0.021
< 0.05). Participants, who complied with the requests, were more likely to behave in a similar
fashion in their own office compared with participants, who did not comply with the requests. The
results showed that a significant number of the participants, who did not comply with the request,
were unsure if they would behave in a similar fashion, and they provided comments stating that
they would comply with the request if there were changes in the suggestion system. For example,
26%
18%
12%
19%
0% 0% 0% 0%
42%
53%
32%
44%
42%
50%
24%
39%
13%
15%
28%
18%
32%
40%
44%
35%
13%
10%
16%
13%
11%
0%
16%
11%
6%
5%
12%
7%
16%
10%
16%
15%
0%
10%
20%
30%
40%
50%
60%
Direct Reciprocity FITD All Direct Reciprocity FITD All
Participants who complied with the request Participants who did not comply with the request
Very Likely Somewhat Likely Undecided Somewhat Unlikely Very Unlikely
`
107

they might comply if an authority, such as their boss were asking for the request or if the request
were delivered through voice instead of text.
Figure 19 – Participants intentions to respond in a similar fashion to the buildings’ request if it was
employed in their own office
We also asked participants if there was anything about the system that they would like to change.
Their suggested changes included adaption of different delivery styles rather than text, such as
through voice or by virtual person, as well as adaption of different delivery methods rather than a
screen on the wall, such as email or mobile applications. In addition, participants believed that the
suggestion system should take office activities into account to find the best timing for the request.
Participants believed that having another character provide the request to the user might enhance
the effectiveness of the request. Another suggestion included personalization of the system to the
74%
78%
64%
73%
53%
40%
36%
43%
3%
3%
8%
4%
26%
20%
32%
28%
13%
15%
24%
17%
21%
30%
24% 24%
10%
5%
4%
6%
0%
10%
8%
6%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
Direct Reciprocity FITD All Direct Reciprocity FITD All
Participants who complied with the request Participants who did not comply with the request
Yes No Not sure Other
`
108

user. For example, participants thought that the request could be more influential if it took user’s
age, gender, and position into consideration.
In this study, we examined the influence of two classic compliance gaining techniques, foot-in-
the-door and reciprocity, which are used in human communications to gain adherence on desired
behaviors. Previous studies in different domains reported that these two techniques increase
compliance with a request [108-110]. Although the results of the present study showed no
significant difference between the participants’ compliance with direct request and FTD,
reciprocity significantly increased compliance with the pro-environmental requests. What
differentiated our study from previous work done in the energy domain was that we focused on
communication intervention strategies, applying social influence methods rather than providing
feedback or information by visualizing energy data (charts directly representing an individual’s
energy use [249,250] or as indirect interventions e.g., ambient light providing feedback by
changing color depending on energy consumption [251] or virtual object, such as a digital pet with
dynamically changing state associated with energy consumption [252]). In addition, we adapted
compliance gaining techniques, which attempt to influence behavior without social pressure,
compared to influence that use social norms.
6.4. Conclusions
In this study, we investigated the influence of social influence techniques that are used in human
communication to gain adherence on building occupants' energy related behaviors. Specifically,
we examined the use of classic compliance-gaining strategies: foot-in-the-door and reciprocity.
We compared these to direct requests. We also investigated whether the results obtained through
IVE could be transferred to the real environments. Our analysis showed that reciprocity received
`
109

a significantly higher rate of compliance (82%), followed by direct request (62%), and FITD (50%)
in the virtual environment. Similarly, reciprocity received significantly higher compliance (73%)
in the real environment, followed by direct request (56%) and FITD (47%).
Investigating how different user personality traits could impact response to building requests, we
found that reciprocity induced greater compliance among participants with higher rate of openness.
In addition, participants with higher rates of neuroticism were less likely to comply with direct
requests. Suggesting the usefulness of IVE to evaluate compliance, there were no significant
differences in participants' compliance in the virtual room vs. physical room in the three
experimental groups. In addition, there were no significant differences between the participants’
preference of using a similar suggestion system in their own office or their intentions to respond
in similar way in the real world in the three experimental groups. The results also suggested that
there are potentials to increase the compliance by changing some aspects of the suggestion system,
such as its delivery style or communicator characteristics. However, we were not able to detect
any significant effects in terms of participants’ age, income levels, or environmental view and
values. This could be due to the fact that our participants were mostly college students, which was
a limitation of our study.
This chapter addresses Research Question 2-1a: “What are the effects of social influence methods
(i.e., foot in the door and reciprocity) on building occupants’ compliance with pro-environmental
requests (e.g., adaption of day lighting instead of artificial lighting)?”; and Research Question 2-
1b: “Do occupants’ characteristics (e.g., demographics, personality traits, and technology
readiness) impact the effects of social influence methods on building occupants’ compliance with
pro-environmental requests?”

`
110

Chapter 7. Exploring the Effects of Communication Delivery Style
and Communicator's Persona
Communication delivery styles are one of the important components of every communication
strategy that tends to influence users’ behaviors. They also impact users’ satisfaction with the
communication system. Communicator's persona is another important component of a successful
communication system, however, there is no consensus among scholars whether a communicator's
persona should portray a human or an inanimate object. It is specifically an unexplored area in
energy domain. In this chapter, we investigated the effects of different communication delivery
styles on influencing building occupants’ compliance with pro-environmental requests. In
addition, we investigated the persona that fits the purpose of the human-building communication
system.
In general, studies showed that people interact differently to media based on its relationalism,
which refers to the social aspect of media. Review of related work showed that there is no
consensus among scholars regarding the effectiveness of different delivery styles on enhancing
behavior change related communications. For example, in situations like persuasion and
negotiation which evoke social norms, verbal and nonverbal cues (e.g., vocal tone and gestures)
can be very important, therefore richer media promotes the effectiveness of the communication
[253,254]. However, in situations like health message reminder modality, leaner media promote
the effectiveness of the communication. As compliance with requests for pro-environmental
actions involves social norms, we expect that the effect of delivery style (i.e., avatar, voice, and
text) on compliance with such requests will be most akin to previous effects on social tasks (e.g.,
negotiation and persuasion). Specifically, prior research (reviewed in Section 2.2.3) demonstrates
`
111

that, when tasks require cooperation and persuasion, more social delivery styles (e.g., avatar and
voice) increase engagement compared to less social delivery styles (e.g., texting).
7.1. Research Methodology
In this study, we investigated the effects of different delivery styles, varying in their level of
relationalism on the effectiveness of persuasive pro-environmental requests. We increased the
relationalism of the delivery style by the inclusion of voice and human-like faces (avatars) to the
delivery of the pro-environmental request. Specifically, we examined the effectiveness of voice
and avatar and then, compared these delivery styles to text. We hypothesized that participants, who
were asked to perform a pro-environmental behavior (e.g., dim or turn off the lights or decrease
the temperature set point) using richer delivery styles would comply more and to the greater extent
than participants receiving the request via a leaner delivery style (e.g., text). Specifically, we
hypothesize that voice would result in greater compliance than text and avatar would result in
greater compliance than voice (H1).
In addition, considering the significant role of communicator’s “persona” or the character in the
level of trust shaping in the relationships, we investigated the influence of the communicator's
personal on the compliance with the pro-environmental requests. Communicator could represent
various personas in this context, but we focused on two possible personas that are more related to
buildings. Specifically, we examined if the compliance was influenced by whether the
communicator represented the building or took a human-like persona that represented the building
facility manager, who plays the role of human profession that is responsible for the building
operation and maintenance. Participants might prefer to communicate with the building itself or
building facility manager. Considering that it is an un-explored research area, we could not
`
112

hypothesize whether an agent representing a building or an agent representing a building facility
manager would engender more compliance. Therefore, we investigated the following question:
"Does building’s persona influence participants’ compliance with persuasive pro-environmental
requests?" In addition, we investigated the interaction between the communicator’s persona and
the effectiveness of the delivery styles. Participants might comply more with a request from a
building that is delivered through text rather than voice or avatar. On the other hand, participants
might comply more with a request from a facility manager that is delivered through richer delivery
styles, such as voice and avatar.
Despite the fact that the communication delivery styles used for the persuasive requests can
influence its effectiveness, these effects might be mediated by other factors (e.g., gender of the
communication delivery style [255,256], and user characteristics (e.g., age, gender, and personality
traits) [257]. Therefore, we also investigated the following questions:
 Does communicator's gender influences the effects of delivery styles on compliance?
 Do participants' characteristics influence the effects of delivery styles on compliance?
7.2. Experiment Setup
214 participants (98 males and 116 females), including graduate and undergraduate students, were
recruited. 81% of the participants were studying at the undergraduate level and 19% at the graduate
level. They either voluntarily participated in the experiment or received course credits for their
participation. The participants were enrolled in different majors including: engineering (31%);
psychology (40%); health (14%); and other majors (15%) including architecture, business, science,
communication, and cinematic arts. The majority of the participants were 18 to 29 years old (80%)
and the rest were around 30 to 39 years old (20%).
`
113

We conducted the experiment in IVEs (Immersive Virtual Environments) which gave us the ability
to easily manipulate the characteristics of the environment or the communication system to
investigate their influences on participants’ behaviors. We examined the influence of three delivery
styles - avatar, voice, and text - and two possible communicator personas - building facility
manager and building itself - on compliance with requests for pro-environmental actions. We also
considered the impact of communicator gender - male and female - on compliance with these
requests. Finally, we also examined other individual difference factors that may qualify the effects
of these experimental variables, such as participant gender and personality traits that might
influence the effects of different delivery styles on compliance.
In this study, participants were asked to decide whether or not to comply with pro-environmental
requests while completing a simulated office-related task (reading two different passages).
Participants were asked to comply with two different pro-environmental requests: change the light
and temperature settings to save energy. By using two different requests, we also investigated if
the content of the requests influenced the compliance with these requests.
If the participant decided to comply, they had to walk to the thermostat or light switch in the
virtual environment (which were both equal distance from the participant’s locations) and use the
control options to change the artificial light levels or temperature setpoints. All participants
performed two tasks and in each task, they were asked to read a passage. The reading levels were
kept consistent for the two passages to ensure there would not be any biases. To rule out confounds
due to order, order was counterbalanced and participants were randomly assigned to order for
communicator's gender (female first and male first) and request content (changing the light or the
temperature setpoint first) (Table 14). Altogether, this resulted in a 2 (persona: building and
building facility manager) X 3 (delivery style: text, voice, and avatar) X 4 (order) X 2
`
114

(communicator gender: male and female) X 2 (request content: changing the light and the
temperature set point) mixed factorial design, where the latter two factors were within-subjects
and the former three factors were between-subjects.
Table 14 – Persona – delivery style - request combinations that the participants will be assigned to
Group Persona
Delivery
Style
Task 1 Task 2

1

Building

Text
Female Text, Request 1* Male Text, Request 2**
Male Text, Request 1 Female Text, Request 2
Female Text, Request 2 Male Text, Request 1
Male Text, Request 2 Female Text, Request 1
Voice
Female Voice, Request 1 Male Voice, Request 2
Male Voice, Request 1 Female Voice, Request 2
Female Voice, Request 2 Male Voice, Request 1
Male Voice, Request 2 Female Voice, Request 1
Avatar
Female Avatar, Request 1 Male Avatar, Request 2
Male Avatar, Request 1 Female Avatar, Request 2
Female Avatar, Request 2 Male Avatar, Request 1
Male Avatar, Request 2 Female Avatar, Request 1

2

Building
Facility
Manager
Text
Female Text, Request 1 Male Text, Request 2
Male Text, Request 1 Female Text, Request 2
Female Text, Request 2 Male Text, Request 1
Male Text, Request 2 Female Text, Request 1
Voice
Female Text, Request 1 Male Text, Request 2
Male Text, Request 1 Female Text, Request 2
Female Text, Request 2 Male Text, Request 1
Male Text, Request 2 Female Text, Request 1
Avatar
Female Text, Request 1 Male Text, Request 2
Male Text, Request 1 Female Text, Request 2
Female Text, Request 2 Male Text, Request 1
Male Text, Request 2 Female Text, Request 1
*If I open the blinds for you to have natural light, would you please dim or turn off the artificial lights?; and
**If I open the window for you to have a breeze and fresh air, would you please increase the temperature setting on
the thermostat?
`
115

To conduct the experiment, first a single occupancy office space was created in Revit© and then
exported to 3ds Max
©
to add materials, furniture, texture, lighting, reflection, and shadows in order
to make the office space look more photo realistic. Then the model was imported to the Unity©
game engine. The Unity© game engine was used to program interactive options, such as
opening/closing blinds and turning on/off lights. A pilot study was conducted with a limited
number of participants to ensure the model, experimental procedure, and the questionnaires were
designed adequately. Based on the findings of the pilot study, we modified the immersive virtual
environment and the questionnaires that were administered during the experiment. In order to
eliminate any effects due to sky conditions, all experiment sessions were conducted on sunny and
non-cloudy days between 9 AM and 4 PM PST, when there was enough natural light outdoors so
that all of the participants experienced the same level of natural light during the experiment.
7.2.1. Pre-Experiment Session
Prior to participating in the experiment, informed consents were obtained from the participants. In
order to investigate if personality traits moderate the effects of delivery styles or persona on
compliance, participants completed a self-report measure of the big five personality traits, the Big
Five Inventory [242]. Aforementioned in the previous chapter, Big Five Inventory (BFI) is
designed to assess the five basic dimensions of personality including extraversion (gregarious,
assertive); agreeableness (cooperative, trusting); conscientiousness (responsible, dependable);
neuroticism (anxiousness); and openness (imaginative, intellectual). Across 44 items ranging from
1 (strongly disagree) to 5 (strongly agree), respondents were asked to rate themselves on all 5 of
these dimensions.
Participants underwent training to become familiar with the immersive virtual environment.
During the training, we showed the participants how to put the head mounted display (HMD) on
`
116

and adjust it to their most comfortable state and work with the different buttons on the controller.
The participants were immersed in the virtual environment and were instructed with different
tasks, such as moving around the room, dimming or turning off the lights, and opening and closing
blinds, changing the temperature set point (Figure 20).

Figure 20 – Training process: a participant is trained how to navigate and interact within the IVE. Figure
also illustrates what participants see through Oculus DK2 Head-Mounted Display
7.2.2. Experiment Session
After the training session, the experimenter explained to participants that they would be asked to
sit on a chair in an office and read two passages, which they should read very carefully as they
would be asked to answer questions about them. They were then told that, during the experiment,
the building (or the building facility manager) would communicate with them and request favors,
and they should pay attention to the requests carefully, as they will be asked what the requests
were at the end of the experiment. However, it is completely up to them whether to comply or not
with requests, and that they should act as they were in their own office. This explanation both
`
117

served to orient the participant to the upcoming task, as well as allow for the persona (building
facility manager vs building itself) to be manipulated.
While participants were reading the passages, a request was delivered to them (at 45 seconds after
starting the task) through one of the delivery styles (text, voice, or avatar) according to the
condition, to which they are assigned. It is important to note that the same voice with the same
social properties was used in the avatar and voice conditions and the only difference in these two
conditions was the inclusion of face in the avatar condition. Two different pro-environmental
requests were delivered to each participant: to turn down the lights, and to turn up the thermostat.
Specifically, participants were asked: (1) if I open the blinds for you to have natural light, would
you please dim or turn off the artificial lights?; and (2) if I open the window for you to have a
breeze and fresh air, would you please increase the temperature setting on the thermostat? These
requests included the contents of rule of reciprocity, which is a classic social influence method that
has been proven to be more influential than the other common social influence methods in one of
our previous studies [199]. The default lighting setting in the office room was the lights on while
the blind was closed. Participants could choose to comply with the pro-environmental request and
lower the level of lights while the blind was open. Likewise, participants could also choose to
comply with the other pro-environmental request and increase the temperature setpoint so energy
was saved on air-conditioning. Female/male avatars and female/male voices were used to deliver
the two requests in each of the delivery style conditions (Figure 21). In the text condition, we
demonstrated the gender of the communicator using these phrases "She said..." or "He said...” We
observed the participants’ compliance with the request and the lighting level and temperature
setpoint that they chose in the assigned conditions.
`
118

Figure 21 – Requests: (a) female text; (b) male text; (c) female voice; (d) male voice; (e) female avatar;
(f) male avatar

(a) (b)

(c) (d)

(e) (f)
`
119

7.2.3. Post-Experiment Session
The post-experiment session included several measures. First, participants completed a
questionnaire assessing their identification with environmentalists with 3 items (one asking if their
attitudes resemble environmentalists’ attitudes, another asking if they agree with
environmentalists, and a third asking if they are similar to environmentalists) ranging from 1
(strongly disagree) to 5 (strongly agree). Next, they completed a questionnaire assessing their
environmental setpoint preferences with 3 items (one asking about preference for lighting sources,
another asking about preference for lighting levels, and a third asking for preference about
temperature setpoint) ranging from 1 (strongly disagree) to 5 (strongly agree). Then they
completed the Allo-inclusive identity scale [258], which measures feelings of connectedness to
people, all living creatures, earth, and universe with 16 items (one target, such as a stranger or the
moon, per item) ranging from 1 to 7 using endpoints displayed in Appendix 2. After that,
participants took the Technology Readiness Index [259], which measures readiness to embrace
technology with 36 items ranging from 1 (strongly disagree) to 7 (strongly agree). Technology
readiness is composed of four parts including: (1) optimism (positive view of the benefits of
technology believing that it offers more control; (2) flexibility, and efficiency in their lives’ (3)
innovativeness (tendency to be technology pioneer and leader among others); and (4) discomfort,
having the feeling of being overwhelmed by technology and lack of having control over it).
Participants were also asked to rate each communicator on 3 items asking, respectively, how
affectionate, friendly, and likable the communicator seemed on scale ranging from 1 (Very
unaffectionate, unfriendly, or unlikable) to 5 (Very affectionate, friendly, or likable). Finally, they
completed the Simulator Sickness Questionnaire (SSQ), which measures 27 symptoms (e.g.,
fatigue, headache, and dizziness) with one item per symptom using a scale from 1 (none) to 4
`
120

(severe). The SSQ had also been completed in the pre-experiment session (after the BFI) so that
pre- and post-test scores could be compared to identify motion sickness caused by emersion in the
IVE.
Additionally, as a stimulus check, we wanted to ensure the IVE was an adequate representation of
the physical environment. Accordingly, at the end of the study, participants were asked to rate on
a scale from 1 (not at all) to 5 (very) how: (1) realistic the virtual environment was and, (2) easy it
was to perform the tasks (change the light and thermostat settings) in the virtual environment
compared to the physical environment. The entire experiment lasted approximately 40 minutes.
7.3. Results
Results of a repeated measure ANOVA examining the differences between before and after
exposure to IVE revealed that 22 participants experienced the symptoms of motion sickness; data
associated with these participants were excluded from the analysis. Accordingly, the results
presented below are based on data from the remaining 192 participants. Further statistical analysis
confirmed that there was no significant effect of order on compliance. Accordingly, we collapsed
across various order conditions in all subsequent analyses. Additional analyses were run to
consider the effect of request content (i.e., to change lighting vs temperature), and these analyses
showed no effect of request content, therefore all further analyses collapse across or ignore content.
Results of the stimulus check showed that participants perceived the environment to be sufficiently
realistic. Participants rated the realism of the environment to be realistic (M = 3.79, SD = 0.66). In
addition, participants rated it to relatively easy to change the light (M = 3.01, SD = 0.78) and
thermostat settings (M = 3.09, SD = 0.88) in the virtual environment compared to the physical
environment.
`
121

7.3.1. Influence of manipulated variables on compliance
To examine the effects of the delivery style and persona on compliance with pro-environmental
requests, we conducted a 3 (delivery style: text, voice, and avatar) × 2 (persona: building and
building facility manager) ANOVA on the number of times participants chose to comply with the
pro-environmental requests. The results showed a significant main effect of delivery style (F(2,
186) = 7.26, p = .001; see Table 15), such that participants complied marginally more when asked
by an avatar (M=1.41) than voice (M=1.17; t(189) = -1.78, p = .078), significantly more to avatar
than text (M=0.91; t(189) = -3.78, p < .001), and also complied significantly more when asked
using voice than text (t(189) = -2.01, p = .05). As expected, delivery styles with richer media did
evoke greater compliance than those with less rich media. Specifically, the richer the media, the
greater compliance with the pro-environmental request. There was also a significant main effect
of persona (F(1, 186) = 5.89, p = .016; see Table 15), such that participants complied more with a
request made by the building facility manager (M = 1.29) rather than the building itself (M = 1.03).
Because the persona of building facility manager is more human than a building, it may be better
able to evoke norms of compliance. This effect of persona did not depend on the style, in which
the request was delivered (text, voice, or avatar), as there was no interaction between persona and
delivery style (F(2, 186) = 0.07, p = .94).
Table 15 - Main effects of delivery style and persona on compliance
Dependent
Measure
Measure
Delivery Style and Persona
Compliance

Delivery
Style
Text
(n = 64)
Voice
(n = 64)
Avatar
(n = 64)

M SD M SD M SD F
.91 .79 1.17 .81 1.41 .63 7.26***
Persona
Building
(n = 96)
Building Facility Manager
(n = 96)

M SD M SD F
1.03 .76 1.29 .77 5.89*
Note. †p < .1; *p < .05; **p < .01; ***p < .001.
`
122

To test whether these effects were moderated by order, we conducted a 3 (delivery style: text,
voice, and avatar) × 2 (persona: building and building facility manager) × 2 (order: male
communicator first and female communicator first) ANOVA. The effects of delivery style and
persona remained with order of the communicator gender entered into the model (see Table 16),
and there were no significant interactions with order (Fs < 1.37, ps > .43).
Table 16 - Main and interaction effects of delivery style, persona, and order of communicator gender on
compliance
Dependent
Measure
Effects Measure F

Compliance

Main Effects
Delivery Style 7.19***
Persona 5.84*
Order .23

Interaction Effects
Delivery Style X Persona .07
Order X Delivery Style .23
Order X Persona .85
Order X Delivery Style X
Persona
1.37

Note. †p < .1; *p < .05; **p < .01; ***p < .001.
In addition to delivery style and persona, we also manipulated the gender of the communicator
(male vs female); unlike delivery style and persona, this factor was within-subjects. All
participants were exposed to both a male and female communicator with order counterbalanced.
As can be seen in Table 17, Chi-square tests revealed a significant effect of gender ( 
2
(1, N = 384)
= 11.647, p = 0.001).

`
123

Table 17 – Effect of communicator gender on compliance
Note. †p < .1; *p < .05; **p < .01; ***p < .001.

As displayed in Figure 22, participants were more likely to comply with the female than the male
communicator. Logistic regression analyses revealed that the effect of gender did not interact
with persona or delivery style conditions, nor with the order of either communicator gender
(male or female first) or message content (male or female request to reduce thermostat or lights;
see Table 18).

Figure 22 – Participants' compliance distribution based on communicator's gender

67%
49%
33%
51%
0%
10%
20%
30%
40%
50%
60%
70%
80%
Female Male
Compliance
Non-compliance
Communicator Compliance
(count)
Non-compliance
(count)
Total Chi-Square P - Value
Female Communicator 128 64 192
Male Communicator 95 97 192
Total 223 161 384 11.65*** .001
`
124

Table 18 – Main and interaction effects of communicator gender on compliance
Dependent
Measure
Equation Measure Wald

Compliance
Logistic Regression Communicator’s Gender 11.51***
Multiple Logistic Regression Text vs Voice 3.43†
Multiple Logistic Regression Avatar vs Voice 5.26*
Multiple Logistic Regression Text vs Avatar 16.34***
Multiple Logistic Regression
Communicator’s Gender X Text vs
Voice
.02
Multiple Logistic Regression
Communicator’s Gender X Avatar vs
Voice
.78
Multiple Logistic Regression
Communicator’s Gender X Text vs
Avatar
.68
Logistic Regression Persona 2.91†
Multiple Logistic Regression Communicator’s Gender X Persona .001
Logistic Regression Order .09
Multiple Logistic Regression Communicator’s Gender X Order .01
Logistic Regression Message Content .001
Multiple Logistic Regression
Communicator’s Gender X Message
Content
.03
Note. †p < .1; *p < .05; **p < .01; ***p < .001.
In addition to participants being more willing to comply with the female communicator, they may
view her more positively than her male counterpart. To test this possibility, we conducted a 2
(Persona) x 3 (Delivery Style) x 2 (Communicator Gender) mixed ANOVA on the average of how
affectionate, friendly, and likable the communicator seemed. The results revealed a significant
main effect of delivery style (F (2, 186) = 40.18, p < .001), such that participants in the avatar (M
= 3.60) and voice conditions (M = 3.37) rated the communicator more positively than those in the
`
125

text condition (M = 2.46). However, no significant main effects were found for persona (F(1, 186)
= 0.16, p = .69) or interaction of persona and delivery style (F(2, 186) = 0.53, p = .59). In addition,
the results showed significant effects of communicator's gender (F(1, 186) = 4.65, p = .03) such
that participants perceived the female communicator (M = 3.19) more positively than the male
communicator (M = 3.09). The analysis also revealed a marginally significant interaction of
persona and gender (F(1, 186) = 3.16, p = .08) such that participants perceived the female
communicator more positively when personified as a building facility manager, but, when
personifying the building, no significant difference was observed between the male and female
communicators (See Figure 23 and Table 19).

Figure 23 – Participants rating for communicator positivity

3.16
3.17
3.21
3.03
2.9
2.95
3
3.05
3.1
3.15
3.2
3.25
Female Male
Building
Building Facility
Manager
`
126

Table 19 – Main effects of demographic variables on compliance
Dependent
Measure
Measure
Delivery Style and Persona
Compliance

Delivery Style
Text
(n = 64)
Voice
(n = 64)
Avatar
(n = 64)

M SD M SD M SD F
2.46 .96 3.37 .61 3.6 .66 40.18
***

Persona

Building
(n = 96)
Building Facility Manager
(n = 96)

M SD M SD F
3.16 .91 3.12 .89 .16

Communicator’s
Gender
Female
(n = 192)
Male
(n = 192)

M SD M SD F
3.19 .97 3.09 .94 4.65*
Note. †p < .1; *p < .05; **p < .01; ***p < .001.

7.3.2. Influence of individual difference variables on compliance
We used regression analyses to examine whether participants’ demographic (i.e., gender, age,
education level, identification with environmentalism, and room set point preferences) had any
effect on compliance or moderated the aforementioned impacts of delivery style and persona.
There were no main effects on compliance for participant gender, age, education level,
identification with environmentalists, or environmental set point preferences (  < .12, ps > .10
Table 20).
`
127

Table 20 – Interaction effects of demographic variables on compliance
Dependent
Measure
Equation
Measure Beta t
Compliance

Multiple Linear
Regression
Text vs Voice -.16 -2.01*
Avatar vs Voice .14 1.77†
Text vs Avatar -.31 - 3.78***

Linear
Regression
Persona .17 2.36*
Participants’ Gender -.03 -.38
Participants’ Age -.07 -.97
Participants’ Education -.02 -.27
Participants’ Environmental Value .02 .24
Participants’ Lighting Preferences .03 .66
Participants’ Temperature Set point -.12 -1.64
Note. †p < .1; *p < .05; **p < .01; ***p < .001.

Furthermore, none of these demographic variables moderated the effects of delivery style or
persona (  < .08, ps > .65); Table 21).
Table 21 – Main and interaction effects of Technology Readiness and Allo-inclusive Identity on
compliance
Dependent
Measure
Between
Subject
Measure
Measure Beta t

Participant’s Gender X Text vs Voice .02 .17
Participant’s Gender X Avatar vs Voice -.04 -.29
Participant’s Gender X Text vs Avatar .06 .46
Participant’s Age X Text vs Voice -.03 -.26
Participant’s Age X Avatar vs Voice -.05 -.52
Participant’s Age X Text vs Avatar .02 .16
Participant’s Education X Text vs Voice -.04 -.34
`
128

Note. †p < .1; *p < .05; **p < .01; ***p < .001.

We also examined whether participants’ individual difference or personality factors (i.e., Allo-
inclusive identity, technology readiness, and big 5 personality traits) had any effect on compliance
or moderated the aforementioned impacts of delivery style and persona. Regression analyses
showed no significant main or interaction effects of any of the Allo-inclusive identity or
technology readiness subscales (  < .16, ps > .14; Table 22). Likewise, we found no significant
main effects of the big 5 personality traits (  < .11, ps > .12); however, two interaction effects
emerged (see Table 23). First, there was a significant interaction between delivery style and
extraversion. Specifically, the interaction term for extraversion and the avatar (0) vs. text (1)

Compliance
Delivery
Style
Participant’s Education X Avatar vs Voice -.05 -.44
Participant’s Education X Text vs Avatar .01 .10
Participant’s Environmental Value X Text vs
Voice
-.04 -.41
Participant’s Environmental Value X Avatar vs
Voice
-.08 -.77
Participant’s Environmental Value X Text vs
Avatar
.04 .34
Participants’ Lighting Preferences X Text vs
Voice
.07 .38
Participants’ Lighting Preferences X Avatar vs
Voice
.01 .06
Participants’ Lighting Preferences X Text vs
Avatar
.06 .34
Participants’ Temperature Setpoint X Text vs
Voice
-.03 -.21
Participants’ Temperature Setpoint X Avatar vs
Voice
.03 .23
Participants’ Temperature Setpoint X Text vs
Avatar
-.06 -.43

Persona
Participants’ Gender X Persona .05 .38
Participants’ Age X Persona -.01 -.05
Participants’ Education X Persona .04 .36
Participant’s Environmental X Persona -.03 -.41
Participants’ Lighting Preferences X Persona .01 .04
Participants’ Temperature Setpoint X Persona -.05 -.38
`
129

dummy-coded variable was significant (  = -.17, p = .05), and simple effects tests revealed that
individuals who were more extroverted complied to a marginally greater extent than those who
were less extroverted when the requests were made by an avatar (  = .18, p = .089), but not in the
voice (  = .05, p = .72) or text (  = -.16, p = .23) conditions. Predicting compliance from persona
and personality, the results showed that there was a marginally significant interaction between
persona and openness (  = -.19, p = .078), such that individuals who were more open complied to
a greater extent than those who were less open when the building facility manager persona was
used (  = 0.21, p = .03) but not when the building persona was used (  = -.05, p = .65). No other
interaction effects emerged for personality traits (  < .16, ps > .18; Table 23).
Table 22 – Main and interaction effects of personality variables on compliance
Dependent
Measure
Between Subject
Measure
Equation
Measure Beta t

Compliance

Technology Readiness
(Main Effects)

Linear
Regression
OPT -.06 -.81
INS .04 .57
DIS -.003 -.04
INN -.01 -.19
Allo-inclusive Identity
(Main Effects)
Linear
Regression
Al - People .11 1.48
Al - World -.005 -.07

Technology Readiness
(Interaction with Delivery
Style)

Multi Linear
Regression
OPT X Text vs Voice .08 .76
OPT X Avatar vs Voice .13 1.28
OPT X Text vs Avatar -.06 -.56
INS X Text vs Voice -.01 -.16
INS X Avatar vs Voice -.12 -1.32
INS X Text vs Avatar .11 1.08
DIS X Text vs Voice -.11 -1.14
DIS X Avatar vs Voice -.03 -.33
DIS X Text vs Avatar -.08 -.77
INN X Text vs Voice -.10 -.94
INN X Avatar vs Voice -.06 -.69
INN X Text vs Avatar -.01 -.10

Al - People X Text vs
Voice
.06 .58
`
130

Note. †p < .1; *p < .05; **p < .01; ***p < .001.

Table 23 – Main effects of delivery style, persona, and communicator gender on how affectionate,
friendly, and likable the communicator seemed

Allo-inclusive identity
(Interaction with Delivery
Style )
Multi Linear
Regression
Al - People X Avatar
vs Voice
-.02 -.22
Al - People X Text vs
Avatar
.09 .86
Al - World X Text vs
Voice
-.08 -.79
Al - World X Avatar vs
Voice
-.12 -1.2
Al - World X Text vs
Avatar
.04 .43

Technology Readiness
(Interaction with Persona)

Multi Linear
Regression
OPT X Persona .16 1.65
INS X Persona .05 .47
DIS X Persona .16 1.45
INN X Persona .11 1.10
Allo-inclusive identity
(Interaction with Persona)
Multi Linear
Regression
Al - People X Persona .01 .98
Al - World X Persona .09 .87
Dependent
Measure
Between
Subject
Measure
Equation Measure Beta t

Compliance

Personality
Traits
(Main Effects)

Linear
Regression
Extraversion .07 .96
Agreeableness .05 .67
Consciousness -.06 -.86
Neuroticism -.02 -.27
Openness .11 1.58

Delivery
Styles

Multi Linear
Regression
Extraversion X Text vs Voice -.11 -1.10
Extraversion X Avatar vs Voice .09 .76
Extraversion X Text vs Avatar -.17 -1.98*
Agreeableness X Text vs Voice .16 1.35
Agreeableness X Avatar vs Voice .16 1.32
Agreeableness X Text vs Avatar .01 .07
Consciousness X Text vs Voice .05 .53
Consciousness X Avatar vs Voice .12 1.17
Consciousness X Text vs Avatar -.07 -.69
`
131

Note. †p < .1; *p < .05; **p < .01; ***p < .001.

7.4. Discussion
Overall, these results suggest that delivery styles can influence the effectiveness of pro-
environmental requests in the context of building-occupant communication and support the
hypothesis that participants would comply more with requests using richer delivery style (avatar
followed by voice) rather than leaner delivery style (text). This finding is in line with previous
work showing that richer delivery styles include more verbal and nonverbal cues (e.g., vocal tone
and gestures) and these factors could help in evoking social norms and increase cooperation and
adherence [253,254] more than less rich delivery styles. Indeed, our results suggest that the richer
the media, the greater the compliance with pro-environmental requests. These results are also
consistent with other studies showing that visual presence of the communicator (virtual human)
leads to more persuasive and affective outcomes than a voice alone or textual message [260,261].
We also found that persona of the communicator influences the effectiveness of the
communication. Framing the communicator as the building facility manager rather than the
building itself increased compliance with pro-environmental requests. Because the persona of
Neuroticism X Text vs Voice .02 .25
Neuroticism X Avatar vs Voice .03 .34
Neuroticism X Text vs Avatar -.01 -.09
Openness X Text vs Voice -.04 -.40
Openness X Avatar vs Voice -.06 -.55
Openness X Text vs Avatar .02 .15

Persona

Multi Linear
Regression
Extraversion X Persona -.06 -.55
Agreeableness X Persona -.04 -.39
Consciousness X Persona .14 1.40
Neuroticism X Persona -.15 -.43
Openness X Persona -.16 -1.77†
`
132

building facility manager is more human than a building, it may be better able to evoke norms of
compliance. Accordingly, it might be harder to refuse a request from a human (building facility
manager) than from an inanimate object (building itself). Indeed, this result seems consistent with
the results of other studies that showed virtual representation is more persuasive and effective
when it is perceived to be controlled by a human rather than a machine [143]. Additionally, agents
that are more believable are more influential [142], and participants might have felt that the agent
who portrays a facility manager was more believable than one portraying the building itself.
Additionally, the results showed that the gender of the communicator affected the effectiveness of
delivery styles on increasing compliance with pro-environmental requests. The female
communicator evoked more compliance compared to the male communicator. Several studies have
shown the preference for female communicators starts as early as the womb. Studies show that
infants prefer their mother’s voice more than their father’s voice [262,263]. Furthermore, when
social norms are involved (as in the present context of pro-environmental requests), female voices
seem to prove to be more persuasive than male voices. Indeed, although studies showed that
effectiveness of communicator gender might be mediated by other factors such as context of the
task, the results of our study and similar studies suggested that female communicators are more
influential in contexts that involve social influence [256,264].
Along these lines, participants perceived the female communicator to be more affectionate,
friendly, and likable than her male counterpart (at least in the building manager condition).
According to Carli [265], “people generally evaluate women more favorably than men and like
them more.” The results of our study and similar studies confirmed that this fact can be applied to
the virtual humans meaning that female avatars are perceived to be more likable than male avatars,
especially when participants believed that they were interacting with an avatar that is more human
`
133

[266,267]. Indeed, our analyses also revealed that participants perceived the avatar to be more
affectionate, friendly, and likable than voice or text, and voice more so than text. In general terms,
these results parallel our findings for compliance, as participants complied more with the avatar
than voice or text, and more with the female than male communicator. Participants’ perceptions of
the delivery style and communicator are therefore in line with their decision to comply (or not).
In addition, the results showed that participants who were more extroverted were more likely to
comply with pro-environmental requests delivered by the avatars, but not with voice or text. As in
other work [268], more extroverted individuals tend to prefer richer delivery styles; perhaps they
experience greater “fit” with this style, given their outgoing, loud nature. Accordingly, they find
the social cues embedded in voice or text less motivating. Additional research should attempt to
replicate and further understand this finding. Additionally, individuals who were more open
complied to a greater extent than those who were less open when the building facility manager
persona was used. People who are higher in openness tend to be more tolerant of new technologies;
perhaps our finding reflects tolerance for a technology that allows a building facility manager to
communicate with them through an online agent. In contrast to this, having a system that simply
pretends it’s the building may seem less novel. Further research could attempt to identify the
underlying mechanism driving more open individuals to have an especially strong preference for
the building manager persona.
7.5. Conclusion
In this study, we investigated the effectiveness of different delivery styles as well as the impact of
communicator's persona on influencing building occupants’ compliance with pro-environmental
requests. The results of our study indicated that the effects of persuasive pro-environmental request
`
134

could be moderated by delivery style, communicator persona, and also communicator gender. In
general, avatars were found to be more effective than voice and text for promoting compliance
with pro-environmental requests. Additionally, participants complied with requests from female
communicator more than male communicator. Finally, computer agents were more effective when
personifying the building facility manager rather than the building itself. These results could be
used to build effective communication between buildings and their occupants to support a desired
behavior such as compliance with pro-environmental requests or reminders.
This chapter addresses Research Question 2-2a: " What are the effects of incorporating social
surface cues (i.e., voice and face) to the delivery of persuasive requests on building occupants’
compliance with pro-environmental requests? "; Research Question 2-2b: " Do occupants’
characteristics (e.g., demographics, personality traits, and technology readiness) impact the effects
of different delivery styles on building occupants’ compliance with pro-environmental requests?”;
Research Question 3-1a: “What are the effects of the different personas (i.e., building facility
manager and building itself) on building occupants’ compliance with pro-environmental
requests?”; and Research Question 3-1b: “Do occupants’ characteristics (e.g., demographics,
personality traits, and technology readiness) impact the effects of persona on building occupants’
compliance with pro-environmental requests?”
`
135

Chapter 8. Exploring the Effects of Social Dialog and
Communicator's Persona
Relational agents establish social-emotional relationships with their users. These relationships lead
to communicative efficiency, better learning outcomes, and improved acceptance of a request
[269]. Therefore, computer agents in applications, in which the goal is to induce a change in user
behavior could be much more effective if they build empathetic relationships with their users.
Research has shown the potential of relational agents to establish a trusting relationship with users
through simple liable verbal and nonverbal techniques [139]. In this study, we investigated the
influence of the social dialog as a verbal technique on the persuasive effects of pro-environmental
requests. However, the effects of social dialog might be mediated by other characteristics of the
communicator such as the persona, which might impact the individuals’ willingness to engage in
a social dialog. Therefore, in this chapter we investigated if the effects of social dialog were
influenced with the persona that communicator represented. We also assessed the effects relating
to the respondents’ characteristics including demographic characteristics (i.e., age, gender,
ethnicity, education, and income), personality traits (i.e., extroversion, agreeableness,
consciousness, neuroticism, and openness), and technology readiness (i.e., optimism,
innovativeness, discomfort, and insecurity).
In spite of the fact that some of the behavior change interventions reported promising results over
one-time interaction, it is not clear whether their effects were maintained over more interactions.
Therefore, we also investigated the effects of pro-environmental requests over one more
interaction with building occupants.
`
136

8.1. Research Methodology
In this study, we investigated the effects of social dialog on the outcomes of pro-environmental
requests, given the influence of two possible personas that are more related to the buildings (i.e.,
building manager vs. building itself). Building on the previous studies showing the potential
capabilities of interactive communications, we predicted to observe more positive effects when
engaging users in social dialogs. We tested the following hypotheses: Hypothesis 1: respondents
will comply more with pro-environmental requests when engaging in a social dialog rather than a
monolog; Hypothesis 2: persona of the communicator will influence the effects of social dialog on
promoting compliance with pro-environmental requests; and Hypothesis 3: respondents’
characteristics will impact the effects of social dialog, and persona on promoting compliance with
the pro-environmental requests. We tested these hypotheses using an on-line survey (Study 1a)
and then, we examined some parts of our findings in a physical office space (Study 1b) to
investigate if there is any difference in our findings from the on-line survey and physical
environment. In addition, to make sure that more interactions with the agents would not affect the
influence of pro-environmental behaviors, we repeated the experiment a week later and
investigated the changes in compliance after one more interaction (Study 1b). Most of the previous
studies did not investigate the effects of interventions over repetitive interactions and it is not clear
whether behavioral changes were maintained. Therefore, we investigated the following question:
how does compliance with pro-environmental requests vary over more interactions with users? In
the following sections, we provided the research procedures and results for the two parts of the
study (Study 1a and Study 1b).
`
137

8.2. Study 1a
An online questionnaire was developed to collect data from a convenience sample of 200
respondents living in the United States and having the experience of working in an office space.
Participants were either recruited from the University of Southern California psychology
department subject pool or Amazon’s Mechanical Turk (www.MTurk.com), which is an online
marketplace used to recruit subjects to undertake tasks [270]. Participants from the subject pool
received course credits and participants from MTurk were paid $3 for their participation.
The experimental design in the survey was based on a 2 (communication mode: monolog vs.
dialog) x 2 (persona: building manager vs. building) between-subjects factorial design. The on-
line survey was composed of multiple questions including demographic questions, measure of
personality traits, measure of environmental values, and technology readiness index. In addition,
a single occupancy virtual office environment was created and embedded in the on-line survey.
The virtual office was rendered using WebGL and it was embedded in the on-line survey (Figure
24). Respondents were required to interact with the virtual office and answer questions according
to what they did and the reasons behind their behaviors. A pilot study was conducted with a limited
number of participants to ensure that the training video and the virtual environment embedded in
the survey were working on different computer systems and questions were designed adequately.
Based on the findings of the pilot study, we modified the immersive virtual environment and the
questionnaires that were included in the survey.
`
138

Figure 24 – Virtual office space embedded in the on-line survey
We collected data from 249 respondents, among which 200 responses were complete and valid.
The other 49 respondents either did not complete the entire survey or did not provide correct
responses when we compared their responses with the codes embedded in the survey. Respondents
included 93 females and 107 males. They were from different ethnicities: 65% White, 10%
Hispanic, 17% Asian, and 8% Black. Respondents’ education levels varied from High school
(28%), to some college courses (34%), and bachelor and more (38%). Respondents were also of
different ages: 18-24 years old (35%), 25-34 (44%), and 35 years old and older (21%). Their
income varied from under $20,000 (32%), $20,000-$50,000 (36%), and $50,000 and more (32%).
8.2.1. Pre-Experiment Session
First, the respondents had to answer demographic questions investigating their age, gender,
ethnicity, education, and income. Then, respondents had to complete self-report measure of the
big five personality traits. Big Five Inventory is designed by John & Srivastava [242] to evaluate
the five basic dimensions of personality including extraversion (how social, energetic, and
assertive a person is); agreeableness (how cooperative, positive, and trusting, a person is);
conscientiousness (how responsible and dependable, and orderly, a person is); neuroticism (how
anxious, furious or prone to depression, a person is); and openness (how imaginative, independent-
`
139

minded, and intellectual, a person is). Respondents were asked to rate their level of extraversion,
agreeableness, conscientiousness, neuroticism, and openness across 44 items ranging from 1
(strongly disagree) to 5 (strongly agree).
Then, the respondents were asked to watch a video to become familiar with the virtual office. In
the video, we showed them how to move around the virtual office space, dim or turn off the lights,
change the temperature set point, and read a passage. When respondents were done with the video,
we asked them to interact with the environment to practice what they learnt in the video.
8.2.2. Experiment Session
After the training, the respondents read a text, which explained what they should do in the virtual
environment. The procedure was as follows: respondents had to sit on a chair in a virtual office
and read a passage. While they were in the virtual environment, a virtual human communicated
with them and requested some favors. The virtual human was representing either the building
manager or the building. The virtual representations in both persona conditions were a pre-
recorded digital representation of a female human, which was found to be more effective than the
male virtual human in one of our previous studies [271]. Respondents were told to pay attention
to the requests carefully, as they would be asked what the requests were at the end of the
experiment. However, it was noted that it was completely up to the respondents whether to comply
or not with the requests, and that they had to act as if they were in their own office. While
respondents were reading the passage, a request was delivered to them. The request included the
contents of rule of reciprocity, which is a classic social influence method that we found it to be
effective in promoting pro-environmental behaviors in another study [199]. Respondents were
required to read two passages and a different pro-environmental request was delivered to them
while they were reading each passage (Figure25). We kept the reading levels consistent for the
`
140

two passages to avoid any biases. First request was “if I open the blinds for you to have natural
light, would you please dim or turn off the artificial lights?” and the second one was “if I open the
window for you to have a breeze and fresh air, would you please increase the temperature setting
on the thermostat?” For the first part, the default lighting setting in the office room was the lights
on while the blind was closed. If the respondents chose to comply with the first pro-environmental
request, they had to go to the light switch and lower the level of lights or turn them off while the
blind was open so energy was saved on lighting. Similarly, if the respondents decided to comply
with the second pro-environmental request, they had to go to thermostat and increase the
temperature set point so energy was saved on air-conditioning. We recorded the participants’
compliance with the requests and the lighting level and temperature set point that they chose by
giving them a code when they were ready to exit the environment. They were asked to report the
code in the post-experiment questionnaire.
(a) (b)
Figure 25 – Pro-environmental requests delivered to the participant in the VE: (a) before the request was
delivered to the participant; (b) while the request was being delivered to the participant (right)

We examined the role of social dialog, given different personas on the persuasiveness of virtual
humans by directly comparing the compliance with pro-environmental requests in different
`
141

conditions. To manipulate persona, participants in the building manager persona condition were
told they were interacting with the building manager and participants in the building persona
condition were told that they were interacting with the building. In the dialog condition, we used
small talk to execute a social dialog. The virtual human first initiated small talk with the respondent
by introducing herself and asked the respondents their name and how they felt. The content of the
requests in the social dialog condition included:
- "Hi, how are you?”
- […]
- "I'm Ellie, representing your building. What's your name?”
- […]
- "I'm glad to see you here. If I open the blinds for you to have natural light, would you please dim
or turn off the artificial lights?"
And
- "Hi, how are you?”
- […]
- "I'm Ellie. I’m your building manager. What's your name?”
- […]
- "I'm glad to see you here. If I open the blinds for you to have natural light, would you please dim
or turn off the artificial lights?"
In the monolog condition, the virtual human just delivered the request to the respondents:
- “If I open the blinds for you to have natural light, would you please dim or turn off the artificial
lights?”
`
142

8.2.3. Post-Experiment Session
After interacting with the virtual environment, the respondents were asked to answer some
questions regarding their actions within the virtual environment including compliance/non-
compliance with the requests and the reasons behind their decisions. We also performed a stimulus
check to ensure that the virtual game was an adequate representation of the physical environment.
In this questionnaire, respondents were asked to rate how realistic the virtual environment was and
how easy it was to perform the tasks (change the light level and temperature set point) compared
to the physical environment. They also rated their sense of presence in the virtual office compared
to the real office. These questionnaires were designed to be rated on a scale from 1 (not at all) to
5 (very). It took the respondents approximately 40 minutes to complete the whole survey. Then,
respondents answered the questions assessing their environmental values to make sure that there
were no significant differences between respondents’ environmental values in different conditions.
These questions asked if participants agree with environmentalists’ attitudes and values, if
participants’ attitudes resemble environmentalists’ attitudes, and if participants were acting similar
to environmentalists. The answers were based on a scale ranging from 1 (strongly disagree) to 5
(strongly agree). Then, we assessed respondents’ environmental set point preferences including
preference for light levels and temperature set points.
Finally, respondents were asked to complete the Technology Readiness Index [259,272], which
assessed respondents’ tendency to embrace technology when accomplishing their tasks across 16
items ranging from 1 (strongly disagree) to 5 (strongly agree). Technology readiness Index is
composed of four subgroups including: optimism (positive assessment of the benefits of
technology, believing that it offers more control, flexibility, and efficiency in people’s lives);
innovativeness (propensity to be technology pioneer and leader among others); discomfort (having
`
143

the feeling of being overwhelmed by technology and lack of having control over it); and insecurity
(distrust of technology and uncertainty about its capabilities as well as concerns about its potential
harms) [259].
8.2.4. Results and Discussion
We analyzed the results of the on-line study investigating the effects of the social dialog on
compliance with pro-environmental requests under the influence of the communicator’s persona.
We conducted power analysis to calculate the appropriate sample size and the results showed that
128 responses in total were needed to have close to 80 percent power to detect statistically
significant effects in differences between the different conditions. We collected 200 responses
resulting in 94 percent power.
First, we tested how respondents were distributed in different conditions in terms of their
characteristics and made sure that there were not significant differences between respondents’
characteristics that might affect the compliance with pro-environmental requests. The results of
statistical analysis showed no significant differences between individuals’ characteristics:
demographic variables (social dialog: Fs < 1.12, ps > .29; and persona: Fs < .98, ps > .32),
personality traits (social dialog: Fs < .41, ps > .68; and persona: Fs < 1.80, ps > .18), technology
readiness (social dialog: Fs < 2.09, ps > .15; and persona: Fs < .39, ps > .53 ), environmental values
(social dialog: F = .31, p = .58; and persona: Fs = .61, ps = .44 ), and lighting preferences (social
dialog: F=.01, p = .94; and persona: F = 2.27, p = .13 ) and temperature set point preferences (social
dialog: F = .2.04, p = .16; and persona: F = .17, p = .69 ). Then, we checked if the virtual
environment was an adequate representation of the physical environment performing a stimulus
check. The results showed that respondents perceived the virtual environment to be realistic (M =
3.71, SD = 0.85). Respondents also rated it to be neither easier nor harder to change the light and
`
144

thermostat settings (M = 2.90, SD = 0.94) in the virtual environment compared to the physical
environment. In addition, respondents rated their sense of presence in the virtual environment to
be similar to their sense of presence in the physical environment (M = 3.56, SD = 0.90).
We calculated ANOVAs with the two independent variables including social dialog and persona
and a dependent variable including compliance. We conducted a 2 (communication mode: social
dialog vs. monolog) x 2 (persona: building vs. building manager) ANOVA on the number of times
respondents chose to comply with the pro-environmental requests.
We identified a marginally significant effect for the social dialog (F(1, 196) = 3.33, p = .07). These
findings support our predictions that the involvement of social dialog is an effective strategy for
inducing compliance with pro-environmental requests. Respondents that were engaged in social
dialog complied slightly more with the pro-environmental requests than the respondents that were
engaged in the monolog (Figure 26). Therefore, social modes of communication that increase the
interactions between buildings and occupants are more effective. Our results are consistent with
the findings of other studies suggesting that engaging users in social dialog promotes more
behavior change than engaging them in a monolog. Particular mode of communication might
represent the way that we deal with strangers or with friends and acquaintances. Particularly,
people associate monolog with strangers and dialog with closer relationships. Therefore, engaging
individuals in a dialog, encourages them to treat a stranger as an acquaintance and become more
willing to comply with the stranger’s requests [157].
However, when we investigated the effects of social dialog for different personas, we found that
the effect of social dialog was dependent on the persona (F(1, 196) = 4.88, p = .03). There were
significant effects of building persona on the social dialog, meaning that when respondents thought
`
145

that they were interacting with the building, engaging in the social dialog resulted in more
compliance than engaging in the monolog (F(1, 98) = 7.29, p = .01). However, there were no
effects of social dialog in the building manager persona and when respondents had the perception
of interacting with the building manager, engaging in the social dialog vs. monolog did not affect
the results (F(1, 98) =.08, p = .77). These results might imply that forming a strong relationship
within one short conversation between two humans is difficult and more interactions are required
to maximize the impacts of the social dialog [273]. However, an in-animate object that engages
users in a social dialog makes the interaction more social and users will become more susceptible
to social influence even within one short conversation and social influence induces changes in
individuals’ behaviors [143,274].

Figure 26 – Compliance (count) for independent variables, persona and social dialog

We also conducted regression analyses to detect any possible effect(s) of demographic variables,
personality traits, and technology readiness on the effectiveness of the manipulated variables
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
Building Building
Manager
Monolog Dialog Building -
Monolog
Building -
Dialog
Building
Manager -
Monolog
Building
Manager -
Dialog
Persona Communication
Mode
Interactions
`
146

including social dialog and persona (Table 24). The results of the statistical analyses using linear
regression showed that gender of the respondents influenced the effects of the social dialog (  = -
.58, p = .05), suggesting that females complied more with pro-environmental requests when the
requests engaged them in social dialog. The reason for this observation might be the fact that
females are more extroverted than males [275,276] and extroverted people prefer to be engaged in
social interactions more than introverts [242]. In addition, income level significantly influenced
the effects of persona (  = .55, p = .01) and social dialog (  = .47, p = .02), suggesting that
respondents with higher income were more likely to comply with the pro-environmental request
when they were interacting with the building manager or were engaged in social dialog.
Individuals with higher income level are more sociable and they generally have larger social
networks and frequent interactions with people [277]. Therefore, more social interactions
including interacting with a virtual representation representing a human (instead of an inanimate
object) and engaging in social dialog rather than a monolog might result in more adherence.
Table 24 - Effects of demographic variables on effectiveness of persona and delivery style
Dependent
Measure
Independent
Measure
Measure Beta t

Persona
Persona x Gender -.36 -1.18
Persona x Ethnicity .05 .30
Persona x Age .09 .43
Persona x Education .01 .07

Compliance
Persona x Income .55 2.53*

Social Dialog
Social Dialog x Gender -.58 -2.01*
Social Dialog x Ethnicity -.01 -.06
Social Dialog x Age .34 1.64
`
147

Social Dialog x Education .06 .32
Social Dialog x Income .47 2.44*

Persona x Social
Dialog
Persona x Social Dialog x Gender .71 1.67
Persona x Social Dialog x Ethnicity .002 .01
Persona x Social Dialog x Age -.27 -.92
Persona x Social Dialog x Education -.15 -.54
Persona x Social Dialog x Income .25 1.74
Note. †p < .1; *p < .05; **p < .01; ***p < .001.
We also explored the effects of respondents’ personality traits on the effects of social dialog,
persona, and their interaction on compliance with pro-environmental requests (Table 25). The
results showed that respondents with higher level of extroversion were more likely to comply with
pro-environmental requests when they were delivered by building manager engaging them in
social dialog rather than other tested conditions (  = .55, p = .01). These results suggested that
having a human engaging users in a social dialog increased compliance by extroverts. Extroversion
is one of the indicators of an individual’s comfort level with face-to-face interactions and an
individual’s sociability [242,273], therefore, individuals with higher levels of extroversion trust
the communication systems more when the interaction is more social (i.e., interacting with a
human engaging users in a small talk) [273].
Table 25 – Effects of personality traits on effectiveness of persona and delivery style
Dependent
Measure
Independent
Measure
Measure Beta t
Compliance

Persona
Persona x Extroversion -.16 -1.08
Persona x Agreeableness -.11 -.60
Persona x Consciousness -1.46 -.90
Persona x Neuroticism -.02 -.17
Persona x Openness .20 1.29

Social Dialog x Extroversion -.04 -.24
Social Dialog x Agreeableness .01 .06
`
148

Dependent
Measure
Independent
Measure
Measure Beta t
Social Dialog Social Dialog x Consciousness -.05 -.27
Social Dialog x Neuroticism -.15 -.99
Social Dialog x Openness -.16 -.93

Persona x Social
Dialog
Persona x Social Dialog x Extroversion .55 2.44*
Persona x Social Dialog x Agreeableness -.02 -.07
Persona x Social Dialog x Consciousness .15 .61
Persona x Social Dialog x Neuroticism .25 1.08
Persona x Social Dialog x Openness -.05 -.23
Note. †p < .1; *p < .05; **p < .01; ***p < .001.
Investigating the impacts of respondents’ technology readiness on the influence of social dialog
and persona on compliance with pro-environmental requests, the results showed significant effects
of individuals’ level of insecurity on the social dialog (  = .34, p = .02), suggesting that
respondents with higher level of insecurity complied more with social dialog rather than monolog.
(Table 26). Insecurity refers to distrust of technology. When virtual representations engage users
in a social dialog, they build a trusting relationship with the users and make the users treat them as
someone, which can increase their trust [278].
In addition, respondents with higher level of innovativeness and insecurity were more likely to
comply with pro-environmental requests when they were delivered by the building manager
(innovativeness:  = .32, p = .6 and insecurity:  = .30, p = .07) rather than the building.
Individuals’ personal innovativeness reflects their willingness to change [279] and an innovative
individual is enthusiastic about new experiences [280]. Therefore, the experience of
communicating with the people in an organization (e.g., a building manager) while they are
represented by a virtual human might make the individuals with higher levels of innovativeness
more willing to comply with pro-environmental requests. In addition, more positive effects of the
building manager persona and social dialog might be justified by the fact that security and trust
towards humans are higher than towards machines [281].
`
149

Table 26 – Effects of technology readiness on effectiveness of persona and delivery style
Dependent
Measure
Independent
Measure
Measure Beta t

Compliance

Persona
Persona x Optimism -1.96 -1.16
Persona x Innovativeness .32 1.87†
Persona x Discomfort .04 .21
Persona x Insecurity .30 1.81†

Social Dialog
Social Dialog x Optimism -.13 -.75
Social Dialog x Innovativeness -.07 -.41
Social Dialog x Discomfort -.21 -1.39
Social Dialog x Insecurity .34 2.34*

Persona x
Social Dialog
Persona x Social Dialog x Optimism .21 .87
Persona x Social Dialog x Innovativeness -.07 -.28
Persona x Social Dialog x Discomfort .23 1.03
Persona x Social Dialog x Insecurity -1.91 -.88
Note. †p < .1; *p < .05; **p < .01; ***p < .001.
8.3. Study 1b
We also conducted an experiment in a physical office space to examine if there were any
differences between our findings of the on-line survey and physical environment and if the
repetition of the request over more interactions would impact the compliance.
Fifty participants (23 males and 27 females), including graduate and undergraduate students from
the University of Southern California, were recruited. Seventy-eight percent of the participants
were studying at the undergraduate level and 22% at the graduate level. The majority of the
participants were 18 to 24 years old (88%) and the rest were around 25 to 34 years old (12%).
`
150

Majority of the participants were white (44%), 16% were Hispanic or Latino, 22% were Asian,
and 18% were black. Participants’ income varied from under $20,000 (70%), $20,000-$50,000
(24%), and $50,000 and more (6%).
8.2.1. Pre-Experiment Session
We repeated the same procedure as the on-line survey in the physical environment. We
manipulated two conditions including monolog and building manager persona. Although social
dialog resulted in more compliance in the on-line survey, we did not manipulate it in the physical
environment since we were also investigating the effects of the repetition of the requests over more
interactions and social dialog needed to be framed real time and based on existing situations and
we could not use the same content (e.g., agent introducing itself and ask the user’s name) for two
consecutive interactions between the same people. In addition, we manipulated the building
manager persona, which was found to have more positive effects than the building persona when
individuals were engaged in a monolog. Prior to beginning the experiment, the participants were
asked to complete the demographic questionnaire as well as the personality questionnaire, which
assessed the Big Five Personality Traits.
8.2.2. Experiment Session
After completing the pre-test survey, participants were asked to read two passages and answer
questions about what they read. They were also told that, during the experiment, the building
manager might communicate with them and request some favors. However, it was emphasized
that it was completely up to the participants to decide whether to comply or not comply with the
requests. While participants were reading the passage, a request was delivered to them through a
female virtual human. It is important to note that the same female virtual human was used in both
on-line survey and the physical environment (Figure 27). Two different pro-environmental
`
151

requests were delivered to each participant: to turn off the lights, and to turn off the air conditioner.
Specifically, participants were asked: (1) if I turn on the desk lamp for you, would you please turn
off the artificial lights?; and (2) If I turn on the fan for you, would you please turn off the air
conditioner? Although the requests asked for similar favors in both the on-line survey and physical
environment, the contents of the requests were slightly different in these two conditions due to
limitations existing in the physical environment. For example, it was not possible to automate the
blind or the window in the existing office space. However, in our previous study, we found that
the content of the pro-environmental requests did not impact the compliance [271]. The default
lighting setting in the office room was the lights on while the desk lamp was off. Participants could
choose to comply with the pro-environmental request and turn off the lights while the desk lamp
was on to save energy on lighting. Likewise, participants could choose to comply with the other
pro-environmental request and turn off the air conditioner while the fan was on to save energy on
air-conditioning.
(a) (b)
Figure 27 – Pro-environmental requests delivered to the participant in the physical environment: (a)
before the request was delivered to the participant; (b) while the request was being delivered to the
participant
`
152

8.2.3. Post Experiment Session
In the post-experiment session, participants were asked to complete questionnaires assessing their
environmental values, their environmental lighting and temperature setpoint preferences, and
Technology Readiness. The entire experiment lasted approximately 40 minutes. We also
investigated the effectiveness of the pro-environmental requests over two time interactions with a
1-week gap between the interactions. Participants were asked to return one week later for the
second part of the experiment and the same procedure as the first time was repeated. We observed
the changes in compliance in the first and second week.
8.2.4. Results
In the physical environment, we investigated the compliance with an agent representing the
building manager, which engaged participants in a monolog. In order to investigate the differences
between the compliance in the on-line survey and physical environment, we compared the on-line
responses from the respondents in the same condition (building manager persona and monolog) as
those who experienced the physical environment. The results of the statistical analysis using
independent sample t-test showed that there were no significant differences between participants'
compliance with the pro-environmental requests in the on-line survey (M = 1.26, SD = .80, N =
50) and physical environment (M = 1.5, SD = .68, N = 50), t(98) = -1.61, p = .11).
We also examined if the participant's compliance would change over more interactions and if the
repetitiveness of the requests would decrease or increase the ability of the system to motivate
participants to perform pro-environmental actions. Thirty out of fifty participants accepted to
participate in the second session of the experiment and came to the physical office space a week
after the first session and went through the same procedure. We conducted paired sample t-test to
examine the changes in the compliance rate in the first and second sessions of the experiment. The
`
153

results of the paired sample t-test showed that there were marginal significant differences between
the participants' compliance rate in the first and second week (t (29) = -1.97, p = .06) and more
compliance with the pro-environmental requests was observed in the second week (M = 1, SD =
.83, N = 30) compared to the first week (M = .73, SD = .69, N = 30). These results suggested that
more interactions with the users might increase the effectiveness of the interventions. The results
are consistent with the findings of the studies suggesting that constant exposure to interventions
might improve the outcomes over more interactions [282,283]. Frequent contacts with individuals
for social purposes can build trust in relationships since individuals can observe the
communicator’s behavior in multiple situations and obtain information that helps them predict the
communicator’s behavior. Frequent contact also results in building closer interpersonal
relationships [284]. Although, in this study, we investigated the changes in the compliance after
one more interaction a week later, there is a need to investigate the changes in the compliance over
longer periods of time.
8.4. Conclusions
To date, little attention has been paid to the behavior change intervention strategies aiming to foster
pro-environmental behaviors in building use. Substantial opportunities exist to work with building
occupants in promoting a wide range of pro-environmental behaviors (e.g., adaption of natural
lighting instead of artificial lighting and natural ventilation instead of mechanical ventilation).
Behavior change may be essential to transition to a sustainable future, but behavior change
intervention strategies have yet to become central to the development of initiatives to foster
sustainable behaviors in buildings [285]. In this chapter, we presented the results of a study
designed to determine the effective communication strategies (i.e., social dialog) used in behavior
change interventions to promote pro-environmental behaviors in buildings. Specifically, we
`
154

explored the effects of social dialog, given two different possible personas associated with
buildings. Our results showed that an interactive social dialog persuaded users to perform more
pro-environmental actions, especially when the requests were delivered by an agent representing
the building. However, these strategies were not equally effective across all types of people. For
example, people with higher income were more persuaded by strategies that included either social
dialog or human-like persona; people with higher level of extroversion were more motivated by
social agents that represented a human-like persona and used social dialog; people with higher
level of insecurity were motivated more with a social dialog; ;and people with higher level of
innovativeness and insecurity were persuaded more when they thought they were interacting with
a human (building manager). Our results also showed no significant differences between our
findings through the on-line survey and our observations in the physical environment. In addition,
when another interaction happened a week later, more positive effects were observed, suggesting
that more interactions can improve the outcomes of interventions.
This chapter addresses Research Question 3-2a: “What are the effects of social dialog on building
occupants’ compliance with pro-environmental requests?”; Research Question 3-2b “Do
occupants’ characteristics (e.g., demographics, personality traits, and technology readiness) impact
the effects of social dialog on building occupants’ compliance with pro-environmental requests?”;
and Research Question 3-3a: “Does persona of the communicator influence the effects of social
dialog on promoting compliance with pro-environmental requests?”; Research Question 3-3b: “Do
occupants’ characteristics (e.g., demographics, personality traits, and technology readiness) impact
the interaction effects of persona and social dialog on building occupants’ compliance with pro-
environmental requests?”; and Research Question 4-1: “Does occupants’ compliance with the pro-
environmental requests change over one more interaction with the building?”
`
155

Chapter 9. Limitations and Potential Future Work
Although our behavior change interventions (LEED branding and direct communication) have
achieved promising results in promoting pro-environmental behaviors in buildings, the work bears
a number of limitations that need to be noted. First, there are some limitations associated with the
sample size and subject diversity during different studies and some caution is needed when
generalizing the reported results. In majority of our studies, the participants were USC students,
who were highly educated and mostly around 18-29 years old. Therefore, the results could not be
generalized to the entire population and a large sample size with much more diversity is needed to
obtain results that could be generalized to larger populations.
We mentioned user characteristics impact the effects of intervention strategies. Although we found
some of these effects in our studies, we were not able to find all the significant effects and one
reason might be the fact that our sample sizes were not large enough. Based on a general rule, for
logistic regression-based analysis, more than 10 observations per degree of freedom is required
[286] and the degrees of freedom are based on the number of variables and their possible levels.
We measured many individual differences including demographics (e.g., age, gender, ethnicity,
education, and income), personality traits (e.g., extroversion, agreeableness, consciousness,
neuroticism, and openness), technology readiness (e.g., optimism, innovativeness, discomfort, and
insecurity), lighting and temperature set point preferences, environmental value, and other
dispositional traits. Therefore, a very large sample size including hundreds of people would be
needed to enable us to have sufficient power to test all these variables. Considering that research
`
156

suggests that interventions could be more successful when they are personalized
7
to the users’
characteristics [287-292], message design, delivery styles, and communicator characteristics could
be more effective when personalized to one specific user [293]. If we are able to find the users’
characteristics that have significant effects on the outcomes of interventions, we could personalize
the interventions in this way. Modeling of behavior with user characteristics can be done in the
future studies when there is access to a large number of people with greater diversity in individual
differences.
Another limitation of this study is associated with the research methodology. There are some
factors that influence occupants’ decisions in terms of lighting preference in an actual office, such
as environment related factors including the time of day, which influence the amount of available
day lighting, and building related factors including design and layout of the building (e.g., position
of the window). However, the experimental condition should be the same for all of the participants
and it is not possible to provide similar conditions for all participants in the physical environment.
Therefore, we used virtual environments to simulate an office space and explore participants’
reactions to the interventions under different scenarios. Virtual environment gave us the ability to
focus on the variable of interest (e.g., social influence method, delivery styles, persona, and social
dialog) and eliminated the influence of the potential confounding variables such as time of the day
and various sky conditions. Although VEs provide many advantages for conducting experiments
involving human behavior and users act in similar ways within VEs as they do in physical
environments, there still exist a number of limitations that come with such technologies. For
example, users had to use a controller to navigate in the virtual environment instead of walking

7
Personalization, an approach, which makes use of data about a specific individual and personalize an
intervention for an individual based on his/her characteristics, is an effective technique to encourage
behavioral influences [174,287,302,303].
`
157

around. They also had to open or close the blinds by clicking a button on the controller or keyboard
instead of pulling down the strings, which changes the amount of effort needed to complete a task
and might affect user’s sense of presence in the environment. However, these differences could be
improved with higher quality devices (e.g., tracking gloves, tracking sensors, and etc.) to improve
participants’ sense of presence within a virtual environment. Another limitation is that while real-
life experimental settings allow for collection of behavioral data throughout a day or a longer
period of time, yet due to motion sickness, we could not keep users longer than 30 to 60 minutes
within immersive virtual environments. However, according to previous studies including ours
where IVEs were used for investigating the effects of social messages [199] and design factors
[191,294] in an office environment, we believe the limitations of IVE will not be significant
enough to have an influence on the users’ interactions within the virtual space.
In addition, participants could not stay in the IVEs for long periods of time due to possibility of
participants' motion sickness. For example, investigating the influence of LEED branding on the
participants' lighting preferences, the assigned task was reading a passage, which had the potential
to divert the participants' attention from energy efficiency. However, examining the influence of
the LEED branding on the environmental behaviors, the assigned task included placing scrap paper
in one of the storage bins, which was more directly associated with the act of recycling. To avoid
this issue, the experiment could have included various tasks that could help us investigate the
influence of the LEED branding on the environmental behavior more indirectly. Yet, considering
the possibility of participants' motion sickness in the virtual environment, we were not able to
increase the duration of the experiment. However, we can design experiments in a way that more
indirectly investigate the influence of LEED branding on the occupants’ recycling behaviors.
`
158

This study also reported promising results for one or two time interactions. However, since
occupants and buildings have continuous interactions, there is a need to investigate the changes in
the behavior over longer periods of time and more interactions. More frequent interactions might
cause fatigue in the users and it may negatively affect the compliance. Incorporating more
relational strategies like real time verbal and non-verbal behaviors (e.g., listening behavior and
facial expressions) might mediate these negative effects. Non-verbal behaviors, such as increased
and direct gaze, smiling, pleasant facial expressions and facial animation in general, and nodding,
play a significant role in creating trusting relationships. Nonverbal behavior could help in
conveyance of the propositional information, expression of emotions, self-presentation, and for
communication of interpersonal attitudes [295]. Since we used a pre-recorded speech for all the
participants, we were not able to investigate the effects of real time verbal responses reflecting
user’s reactions. In the future research, we can incorporate real time verbal and non-verbal
behaviors in design of our interventions and monitor the changes over longer period of time. In
addition, some relational tactics (e.g., social dialog) might be more effective in long-term
interventions, therefore, there is a need to investigate the difference between short-term
compliance and long-term adherence.
It is also important to note that in our studies, we used a single occupancy office space, however,
people’s behavior might change in the presence of others [296]. Therefore, it is important to
examine the effects of these interventions in other architectural settings such as multi-occupancy
or open-plan offices as well as other types of spaces. In this dissertation, we focused on strategies
aiming to promote occupants’ pro-environmental behaviors in buildings. These strategies can also
be applied to achieve other joint goals of building and occupants including improving safety,
security and public health in built environments. In design of pro-environmental requests, we did
`
159

not take advantage of the emotional benefits of pro-environmental behaviors such as their impact
on the environment and human wellbeing, these strategies could potentially be used in other
contexts such as safety and comfort. In future studies, the effects of human-building
communication on achieving other joint goals of buildings and their occupants could be studied.

`
160

Chapter 10. Conclusions
Behavior change is essential to transition to a sustainable future, but behavior change intervention
strategies have yet to become central to the development of initiatives to foster sustainable
behaviors in buildings [285]. This dissertation explores effective behavior change strategies
aiming to promote pro-environmental behaviors in buildings. We investigated the effects of LEED
branding as an indirect behavior change strategy as well as direct communication on promoting
pro-environmental behaviors in buildings. The following communication components were
identified in literature as main components of a successful communication: intended audience,
intended behavior, message design, delivery style, and communicator [72,73]. Previous work
showed that occupant behavior is one of the significant contributors to building energy demand
and pro-environmental behaviors offer a significant potential for cost effective energy savings
[297-299]. Therefore, in this dissertation, we focused on the factors that are influenced the most
by occupant behaviors and have significant impacts on buildings’ energy consumption, such as
lighting operations and thermostat adjustments.
We explored effective strategies for design of other components of the communication including
message design, delivery style and communicator. We hypothesized that interventions would be
more persuasive if the interaction was seen as more social and a relationship was built between
buildings and their occupants. Therefore, we investigated the consequences of making human-
building communication more social by incorporating social features such as social influence
methods (e.g., foot in the door and reciprocity) and social-cues (e.g., face and voice) into the design
of interventions. In addition, we investigated the effects of relational strategies (e.g., persona and
social dialog) on making a relationship with users. Considering that people apply social rules to
social and relational agents and react differently to them based on their own characteristics (e.g.,
`
161

demographics and personality traits), we also examined the effects of user characteristics on the
outcomes of these interventions.
We tested the hypotheses and evaluated our findings in both virtual and physical environments
and found similar effects in both environments. Our results demonstrated LEED branding
motivated occupants to act more pro-environmentally and communication was more persuasive
when the interaction was more social and relational strategies were adapted. In fact, we found out
that rules that apply to human-human interactions applied also to human-building interactions. For
example, more positive outcomes were observed with reciprocal requests; requests that were
delivered through face to face interaction (avatar) rather than voice and text; while participants had
the perception of interacting with a human-like persona (i.e., building manager) rather than an in-
animate object (i.e. building); and when participants were engaged in a social dialog (i.e., small
talk) rather than a monolog. However, these strategies were not equally effective across all types
of people and their effects varied for people with different characteristics. For example, the effects
of LEED branding were greater for participants with higher environmental values and views and
extroverts were more willing to comply with pro-environmental requests when they had the
perception of interacting with a human and were engaged in a social dialog. However, these effects
were not the same for the introverts.
Although in our studies, we found promising results over one-time interaction, it is not clear
whether the effects of interventions would be the same over more interactions. Therefore, we
conducted another experiment, in which another interaction happened a week later. The results
showed more positive effects during the second interaction, suggesting that more interactions can
improve the outcomes of interventions.
`
162

In sum, we investigated effective design strategies for successful behavior change interventions
aiming to promote pro-environmental behaviors as well as the factors that could impact the effects
of these interventions. The findings of this dissertation represent useful design choices for
persuasive technologies aiming to promote pro-environmental behaviors.
`
163

References
[1] N. E. Klepeis, W. C. Nelson, W. R. Ott, J. P. Robinson, A. M. Tsang, P. Switzer, J. V. Behar,
S. C. Hern, W. H. Engelmann. The National Human Activity Pattern Survey (NHAPS): a
resource for assessing exposure to environmental pollutants. Journal of exposure analysis and
environmental epidemiology. 11 (2001) 231-252.
[2] W. Guo, M. Zhou. Technologies toward thermal comfort-based and energy-efficient HVAC
systems: A review. Systems, Man and Cybernetics. SMC 2009. 2009, pp. 3883-3888.
[3] D. Steinberg, M. Patchan, C. Schunn, A. Landis. Determining adequate information for green
building occupant training materials. Journal of Green Building. 4 (2009), 143-150.
[4] D. Yan, W. O’Brien, T. Hong, X. Feng, H. B. Gunay, F. Tahmasebi, A. Mahdavi. Occupant
behavior modeling for building performance simulation: Current state and future challenges.
Energy and Buildings. 107 (2015) 264-278.
[5] L. Sawyer, P. De Wilde, S. Turpin-Brooks. Energy performance and occupancy satisfaction:
A comparison of two closely related buildings. Facilities. 26 (2008) 542-551.
[6] J. H. Scofield. Do LEED-certified buildings save energy? Not really…. Energy and
Buildings. 41(2009) 1386-1390.
[7] W. Chang, T. Hong. Statistical analysis and modeling of occupancy patterns in open-plan
offices using measured lighting-switch data. Building Simulation. (2013) 23-32.
[8] R. V. Andersen, J. Toftum, K. K. Andersen, B. W. Olesen. Survey of occupant behaviour and
control of indoor environment in Danish dwellings. Energy and Buildings. 41 (2009) 11-16.
[9] F. Haldi. A probabilistic model to predict building occupants’ diversity towards their
interactions with the building envelope. Proceedings of the International IBPSA Conference,
Chambery, France, 2013.
[10] K. Papakostas, B. Sotiropoulos. Occupational and energy behaviour patterns in Greek
residences. Energy and Buildings. 26 (1997) 207-213.
`
164

[11] C. M. Clevenger, J. Haymaker. The impact of the building occupant on energy modeling
simulations. Joint International Conference on Computing and Decision Making in Civil and
Building Engineering, Montreal, Canada, 2006, pp. 1-10.
[12] S. Creer, S. Cunningham, M. Hawley, P. Wallis. Describing the interactive domestic robot
setup for the sera project. Applied Artificial Intelligence. 25 (2011) 445-473.
[13] R. M. Anderson, M. M. Funnell. Negotiating behavior changes with patients who have
diabetes: negotiation or coercion? Diabetes Management. 2 (2012) 41-46.
[14] P. Aggleton. Behavior change communication strategies. AIDS Education and Prevention :
Official Publication of the International Society for AIDS Education. 9 (1997) 111-123.
[15] United Nations Environment Programme. Buildings Can Play a Key Role in Combating
Climate Change2007, Last accessed: March 31, available at:
http://www.unep.org/Documents.Multilingual/Default.asp?DocumentID=502&ArticleID=55
45&l=en.
[16] M. A. Khan, M. Z. Khan, K. Zaman, L. Naz. Global estimates of energy consumption and
greenhouse gas emissions. Renewable and Sustainable Energy Reviews. 29 (2014) 336-344.
[17] L. Yang, H. Yan, J. C. Lam. Thermal comfort and building energy consumption
implications–A review. Applied Energy. 115 (2014) 164-173.
[18] eia, U.S. Energy Information Administration. Total Energy. Last updated at 2017, Accessed
at October 2017. Available at:
https://www.eia.gov/totalenergy/data/browser/?tbl=T02.01#/?f=A&start=1949&end=2016&chart
ed=3-6-9-12
[19] eia, U.S. Energy Information Administration. Energy Consumption by Consumption and
Source. Last accessed: November 2014, available at:
http://www.eia.gov/forecasts/aeo/er/excel/aeotab_2.xlsx.
`
165

[20] U.S. Department of Energy. Building Energy Data Book, Commercial Sector
Characteristics. Last accessed: February 2012, available at:
http://buildingsdatabook.eren.doe.gov/TableView.aspx?table=3.2.2.
[21] L. Pérez-Lombard, J. Ortiz, C. Pout. A review on buildings energy consumption
information. Energy and Buildings. 40 (2008) 394-398.
[22] R. A. Buswell, Menezes, Anna Carolina Kossmann de, A. Cripps, D. Bouchlaghem.
Analysis of electricity consumption for lighting and small power in office buildings. In CIBSE
Techni-cal Symposium, DeMontfort University, Leicester, UK, 6th and 7th September, 2011.
[23] F. Jazizadeh, S. Ahmadi-Karvigh, B. Becerik-Gerber, L. Soibelman. Spatiotemporal lighting
load disaggregation using light intensity signal. Energy and Buildings. 69 (2014) 572-583.
[24] Bureau of Labor Statistics. American Time Use Survey Summary - Economic News
Release. Last accessed: August 2017, available at:
https://www.bls.gov/news.release/atus.nr0.htm.
[25] T. Hong, H. Lin. Occupant Behavior: Impact on Energy Use of Private Offices. Ernest
Orlando Lawrence Berkeley National Laboratory, Berkeley, CA, USA, 2013.
[26] M. J. Coleman, K. N. Irvine, M. Lemon, L. Shao. Promoting behaviour change through
personalized energy feedback in offices. Building Research & Information. 41(2013) 637-651.
[27] O. Nisiforou, S. Poullis, A. Charalambides. Behaviour, attitudes and opinion of large
enterprise employees with regard to their energy usage habits and adoption of energy saving
measures. Energy and Buildings. 55 (2012) 299-311.
[28] O. Masoso, L. Grobler. The dark side of occupants’ behaviour on building energy use.
Energy and Buildings. 42 (2010) 173-177.
[29] M. Bonte, F. Thellier, B. Lartigue. Impact of occupant's actions on energy building
performance and thermal sensation. Energy and Buildings. 76 (2014) 219-227.
`
166

[30] W. Parys, D. Saelens, H. Hens. Impact of occupant behavior on lighting energy use.
Proceedings of Building Simulation. 2009, 1143-1150.
[31] A. Tzempelikos. The impact of manual light switching on lighting energy consumption for a
typical office building. International High Performance Buildings Conference, July 2010.
[32] N. Wang, J. Zhang, X. Xia. Energy consumption of air conditioners at different temperature
set points. Energy and Buildings. 65 (2013) 412-418.
[33] H. Sarantis. Business guide to paper reduction. (2002).
[34] G. Peschiera, J. E. Taylor. The impact of peer network position on electricity consumption
in building occupant networks utilizing energy feedback systems. Energy and Buildings. 49
(2012) 584-590.
[35] A. R. Carrico, M. Riemer. Motivating energy conservation in the workplace: An evaluation
of the use of group-level feedback and peer education. Journal of Environmental Psychology. 31
(2011) 1-13.
[36] R. Katzev, H. R. Mishima. The use of posted feedback to promote recycling. Psychological
reports. 71(1992) 259-264.
[37] B. E. Porter, F. C. Leeming, W. O. Dwyer. Solid waste recovery: A review of behavioral
programs to increase recycling. Environment and Behavior. 27 (1995), 122-152.
[38] J. Froehlich, L. Findlater, J. Landay. The design of eco-feedback technology. Proceedings of
the SIGCHI Conference on Human Factors in Computing Systems, 2010, pp. 1999-2008.
[39] W. Abrahamse, L. Steg, C. Vlek, T. Rothengatter. A review of intervention studies aimed at
household energy conservation. Journal of Environmental Psychology. 25 (2005) 273-291.
[40] T. Segal. Rethiking Energy: The importance of Education. Education Week. Last accessed:
November 2013, available at:
http://blogs.edweek.org/edweek/reimagining/2013/11/rethinking_energy_in_education.html.
`
167

[41] R. A. Winett, J. H. Kagel, R. C. Battalio, R. C. Effects of monetary rebates, feedback, and
information on residential electricity conservation. Journal of Applied Psychology. 63 (1987),
73.
[42] L. McClelland, S. W. Cook. Energy conservation effects of continuous in-home feedback in
all-electric homes. Journal of Environmental Systems. 9 (1979), 169-173.
[43] J. C. Yang, K. H. Chien, T. C. Liu. A digital game-based learning system for energy
education: an energy conservation pet. Turkish Online Journal of Educational Technology-
TOJET. 11 (2012) 27-37.
[44] A. Gustafsson, C. Katzeff, M. Bang. Evaluation of a pervasive game for domestic energy
engagement among teenagers. Computers in Entertainment (CIE). 7 (2009) 54.
[45] L. Van de Velde, W. Verbeke, M. Popp, G. Van Huylenbroeck. The importance of message
framing for providing information about sustainability and environmental aspects of energy.
Energy Policy. 38 (2010) 5541-5549.
[46] R. A. Winett, I. N. Leckliter, D. E. Chinn, B. Stahl, S. Q. Love. Effects of television
modeling on residential energy conservation. Journal of applied behavior analysis. 18 (1985) 33-
44.
[47] R. B. Hutton, D. L. McNeill. The value of incentives in stimulating energy conservation.
Journal of Consumer Research. (1981) 291-298.
[48] P. C. Stern, E. Aronson, J. M. Darley, D. H. Hill, E. Hirst, W. Kempton, T. J. Wilbanks. The
effectiveness of incentives for residential energy conservation. Evaluation review. 10 (1986)
147-176.
[49] J. H. Van Houwelingen, W. F. Van Raaij. The effect of goal-setting and daily electronic
feedback on in-home energy use. Journal of consumer research. 16 (1989) 98-105.
[50] S. J. Kantola, G. J. Syme, N. A. Campbell. Cognitive dissonance and energy conservation.
Journal of Applied Psychology. 69 (1984) 416.
`
168

[51] L. J. Becker. Joint effect of feedback and goal setting on performance: A field study of
residential energy conservation. Journal of applied psychology. 63 (1978) 428.
[52] G. N. Dixon, M. B. Deline, K. McComas, L. Chambliss, M. Hoffmann. Using comparative
feedback to influence workplace energy conservation: A case study of a university campaign.
Environment and Behavior. 47 (2015) 667-693.
[53] P. Oliver, G. Marwell, R. Teixeira. A theory of the critical mass. I. Interdependence, group
heterogeneity, and the production of collective action. American journal of Sociology. 91(1985)
522-556.
[54] J. Axsen, K. S. Kurani. Social influence, consumer behavior, and low-carbon energy
transitions. Annual Review of Environment and Resources. 37 (2012) 311-340.
[55] A. Boztepe. Green marketing and its impact on consumer buying behavior. European
Journal of Economic and Political Studies. 5 (2012) 5-21.
[56] K. Sammer, R. Wüstenhagen. The influence of eco‐labelling on consumer behaviour–
Results of a discrete choice analysis for washing machines. Business Strategy and the
Environment. 15 (2006) 185-199.
[57] J. Pickett-Baker, R. Ozaki. Pro-environmental products: marketing influence on consumer
purchase decision. Journal of consumer marketing. 25 (2008) 281-293.
[58] E. SW Chan. Gap analysis of green hotel marketing. International Journal of Contemporary
Hospitality Management. 25 (2013) 1017-1048.
[59] H. Han, Y. Kim. An investigation of green hotel customers’ decision formation: Developing
an extended model of the theory of planned behavior. International Journal of Hospitality
Management. 29 (2010) 659-668.
[60] J. Lee, L. Hsu, H. Han, Y. Kim. Understanding how consumers view green hotels: how a
hotel's green image can influence behavioural intentions. Journal of Sustainable Tourism. 18
(2010) 901-914.
`
169

[61] H. Han, L. J. Hsu, J. Lee. Empirical investigation of the roles of attitudes toward green
behaviors, overall image, gender, and age in hotel customers’ eco-friendly decision-making
process. International Journal of Hospitality Management. 28 (2009) 519-528.
[62] N. S. Mokhtar Azizi, S. Wilkinson, E. Fassman. Strategies for improving energy saving
behaviour in commercial buildings in Malaysia. Engineering, Construction and Architectural
Management. 22 (2015) 73-90.
[63] N. S. M. Azizi, S. Wilkinson, E. Fassman. An analysis of occupants response to thermal
discomfort in green and conventional buildings in New Zealand. Energy and Buildings. 104
(2015) 191-198.
[64] R. Cole. Green buildings and gray occupants.e AIA‐USGBC Conference on Mainstreaming
Green, Chattanooga, TN, October 14‐16, 1999.
[65] H. Herring, R. Roy. Technological innovation, energy efficient design and the rebound
effect. Technovation. 27 (2007) 194-203.
[66] J. D. Khazzoom. Economic implications of mandated efficiency in standards for household
appliances. The Energy Journal. 1 (1980) 21-40.
[67] J. R. Catlin, Y. Wang. Recycling gone bad: When the option to recycle increases resource
consumption. Journal of Consumer Psychology. 23 (2012) 122-127.
[68] S. Bamberg, G. Möser. Twenty years after Hines, Hungerford, and Tomera: A new meta-
analysis of psycho-social determinants of pro-environmental behaviour. Journal of
Environmental Psychology. 27 (2007) 14-25.
[69] N. Mazar, C. B. Zhong. Do green products make us better people? Psychological science.
21 (2010) 494-498.
[70] L. E. Bolton, J. B. Cohen, P. N. Bloom. Does marketing products as remedies create “get
out of jail free cards”? Journal of Consumer Research. 33 (2006) 71-81.
`
170

[71] M. Frondel. An introduction to energy conservation and the rebound effect.
Energy Technology and Policy. 2 (2004) 203-208.
[72] w. McGuire. Theoretical Foundations of Campaigns. Public communication campaigns.
Ronald Rice, & Charles, Beverly Hills, California, Sage, 1981, pp. 41-70.
[73] H. D. Lasswell. The structure and function of communication in society. The
communication of ideas. 37 (1948) 215-228.
[74] M. W. Kreuter, S. M. McClure. The role of culture in health communication. Annual
Review of Public Health. 25 (2004) 439-455.
[75] Resource NSW. Waste Reduction in Office Buildings. Last accessed: October 2017.
Available at: http://recyclingnearyou.com.au/documents/DEC_WasteMgrs.pdf.
[76] Statistic Brain. Paper use statistics. Last updated: October 2016. Last accessed: October
2017. Available at: http://www.statisticbrain.com/paper-use-statistics/.
[77] V. Fabi, S. P. Corgnati, R. V. Andersen, M. Filippi, B. W. Olesen. Effect of occupant
behaviour related influencing factors on final energy end uses in buildings. Proceedings of
Climamed11 Conference, Madrid, June 2011.
[78] S. Ghaemi, G. Brauner. User behavior and patterns of electricity use for energy saving.
Internationale Energiewirtschaftstagung an der TU Wien, IEWT. 2009.
[79] M. Moezzi, C. Hammer, J. Goins, A. Meier. Behavioral strategies to bridge the gap between
potential and actual savings in commercial buildings. Contract Number: 09-327. Sacramento: Air
Resources Board. 2014.
[80] O. G Santin. Occupant behaviour in energy efficient dwellings: evidence of a rebound
effect. Journal of Housing and the Built Environment. 28 (2013) 311-327.
[81] A. I. Rubin, B. L. Collins, R. L. Tibbott. Window Blinds as a Potential Energy Saver: A
Case Study. US Department of Commerce, National Bureau of Standards. (1978).
`
171

[82] V. Fabi, R. V. Andersen, S. Corgnati, B. W. Olesen. Occupants' window opening behaviour:
A literature review of factors influencing occupant behaviour and models. Building and
Environment. 58 (2012) 188-198.
[83] V. Fabi, R. V. Andersen, S. Corgnati, B. W. Olesen. A methodology for modelling energy-
related human behaviour: Application to window opening behaviour in residential buildings.
Building Simulation. (2013) 415-427.
[84] P. Tuohy, H. Rijal, M. Humphreys, J. Nicol, A. Samuel, J. Clarke. Comfort driven adaptive
window opening behaviour and the influence of building design. Proceedings of Building
Simulation 2007, 10th IBPSA Conference, 2007.
[85] G. Wood, M. Newborough. Dynamic energy-consumption indicators for domestic
appliances: environment, behaviour and design. Energy and Buildings. 35 (2003) 821-841.
[86] R. Yao, K. Steemers. A method of formulating energy load profile for domestic buildings in
the UK. Energy and Buildings. 37 (2005) 663-671.
[87] D. Wang, C. C. Federspiel, F. Rubinstein. Modeling occupancy in single person offices.
Energy and Buildings. 37 (2005) 121-126.
[88] Y. Agarwal, B. Balaji, R. Gupta, J. Lyles, M. Wei, T. Weng. Occupancy-driven energy
management for smart building automation. Proceedings of the 2nd ACM Workshop on
Embedded Sensing Systems for Energy-Efficiency in Building. 2010, 1-6.
[89] J. A. Davis III, D. W. Nutter. Occupancy diversity factors for common university building
types. Energy and Buildings. 42 (2010) 1543-1551.
[90] L. Klein, J. Kwak, G. Kavulya, F. Jazizadeh, B. Becerik-Gerber, P. Varakantham, M.
Tambe. Coordinating occupant behavior for building energy and comfort management using
multi-agent systems. Automation in Construction. 22 (2012) 525-536.
[91] E. Azar, C.C. Menassa. Agent-based modeling of occupants and their impact on energy use
in commercial buildings. Journal of Computing in Civil Engineering. 26 (2011) 506-518.
`
172

[92] C. A. Webber, J. A. Roberson, M. C. McWhinney, R. E. Brown, M. J. Pinckard, J. F. Busch.
After-hours power status of office equipment in the USA. Energy. 31 (2006) 2823-2838.
[93] M. Sanchez, C. Webber, R. Brown, J. Busch, M. Pinckard, J. Roberson. Space heaters,
computers, cell phone chargers: How plugged in are commercial buildings? Lawrence Berkeley
National Laboratory. (2007).
[94] A. Al-Mumin, O. Khattab, G. Sridhar. Occupants’ behavior and activity patterns influencing
the energy consumption in the Kuwaiti residences. Energy and Buildings. 35 (2003) 549-559.
[95] W. Abrahamse, L. Steg. Factors related to household energy use and intention to reduce it:
The role of psychological and socio-demographic variables. Human Ecology Review. 18(2011)
30-40.
[96] R. B. Adam Burke Measuring End-Use Technological and Behavioral Waste to Prioritize
and Improve Program Design. International Energy Program Evaluation Conference, 2013,
Chicago.
[97] M. Eilers, J. Reed, T. Works. Behavioral aspects of lighting and occupancy sensors in
private offices: a case study of a university office building. ACEEE 1996 Summer Study on
Energy Efficiency in Buildings, 1996.
[98] T. A. Nguyen, M. Aiello. Energy intelligent buildings based on user activity: A survey.
Energy and Buildings. 56 (2013) 244-257.
[99] C. Morgan, R. de Dear. Weather, clothing and thermal adaptation to indoor climate. Climate
Research. 24 (2003) 267-284.
[100] K. C Parsons. The effects of gender, acclimation state, the opportunity to adjust clothing
and physical disability on requirements for thermal comfort. Energy and Buildings. 34 (2002)
593-599.
[101] American Geosciences Institute (AGI). How does recycling save energy? Last updated:
2017. Last accessed: October 2017. Available at: https://www.americangeosciences.org/critical-
issues/faq/how-does-recycling-save-energy.
`
173

[102] MPS. The Strategy of Persuasion. Last accessed: November 2014, available at:
https://hseminarrhetoric.files.wordpress.com/2011/10/mps_16_the_strategy_of_persuasion.pdf
[103] L. R. Wheeless, R. Barraclough, R. Stewart, R. Bostrom. Compliance-gaining and power
in persuasion. Communication yearbook. 7 (1983) 105-145.
[104] C. A. Scott. Modifying socially-conscious behavior: The foot-in-the-door technique.
Journal of Consumer Research. (1977) 156-164.
[105] J. L. Freedman, S. C. Fraser. Compliance without pressure: the foot-in-the-door technique.
Journal of personality and social psychology. 4 (1966) 195.
[106] P. H. Reingen. On inducing compliance with requests. Journal of Consumer Research.
(1978) 96-102.
[107] N. Guéguen. Foot-in-the-door technique and computer-mediated communication.
Computers in Human Behavior. 18 (2002) 11-15.
[108] N. Gueguen, M. Marchand, A. Pascual, M. Lourel. Foot-in-the-door technique using a
courtship request: a field experiment 1. Psychological reports. 103 (2008) 529-534.
[109] Guéguen, N., Jacob, C. Fund-raising on the web: the effect of an electronic foot-in-the-
door on donation. CyberPsychology & Behavior. 4 (2001) 705-709.
[110] F. Girandola. Sequential requests and organ donation. The Journal of social psychology.
142 (2002) 171-178.
[111] Guadagno, R. E., Liberman, C. Social influence online: The six principles in action.
Casing persuasive communication. Dubuque, IA: Kendall Hunt. (2013).
[112] A. W. Gouldner. The norm of reciprocity: A preliminary statement. American Sociological
Review. (1960) 161-178.
[113] N. Summers. TNW. The art of persuasion: How to pull in more customers with scarcity,
reciprocity and other tactics. 2013. Last updated: November 2013, Last accessed: May 2014,
`
174

available at: https://thenextweb.com/insider/2013/11/02/art-persuasion-scarcity-reciprocity-
quality-products-can-help-pull-customers/.
[114] D. T. Regan. Effects of a favor and liking on compliance. Journal of experimental social
psychology. 7 (1971) 627-639.
[115] M. A. Whatley, J. M. Webster, R. H. Smith, A. Rhodes. The effect of a favor on public and
private compliance: How internalized is the norm of reciprocity? Basic and Applied Social
Psychology. 21 (1999) 251-259.
[116] Y. Takeuchi, Y. Katagiri, C. I. Nass, B. Fogg. Social response and cultural dependency in
human-computer interaction. Proceedings of PRICAI. (1998) 114–123.
[117] C. Nass, Y. Moon. Machines and mindlessness: Social responses to computers. Journal of
Social Issues. 56 (2002) 81-103.
[118] Y. Moon. Intimate exchanges: Using computers to elicit self-disclosure from consumers.
Journal of consumer research. 26 (2000) 323-339.
[119] B. Fogg, C., Nass. How users reciprocate to computers: an experiment that demonstrates
behavior change. CHI'97 extended abstracts on Human factors in computing systems. 1997, pp.
331-332.
[120] J. M. Burger, J. Sanchez, J. E. Imberi, L. R. Grande. The norm of reciprocity as an
internalized social norm: Returning favors even when no one finds out. Social Influence. 4
(2009) 11-17.
[121] A. Gaudeul, C. Peroni. Reciprocal attention and norm of reciprocity in blogging networks.
Gaudeul, Alexia and Peroni, Chiara, Reciprocal Attention and Norm of Reciprocity in Blogging
Networks. JENA Economic Research Papers #2010-020. (2010). Available at
SSRN: https://ssrn.com/abstract=1577062 or http://dx.doi.org/10.2139/ssrn.1577062.
[122] E. Williams. Experimental comparisons of face-to-face and mediated communication: A
review. Psychological bulletin. 84 (1977) 963.
`
175

[123] M. Davies. Cooperative Problem Solving: An Exploratory Study. British Post Office.
(1971).
[124] B. Champness, M. Davies. The Maier pilot experiment. Unpublished Communication
Studies Group paper nº.E/710301/CH. 1971.
[125] J.A. Short. Effects of medium of communication on experimental negotiation. Human
Relations. 27 (1974) 225-234.
[126] N. Bos, J. Olson, D. Gergle, G. Olson, Z. Wright. Effects of four computer-mediated
communications channels on trust development. Proceedings of the SIGCHI conference on
human factors in computing systems, 2002, pp. 135-140.
[127] C. Jensen, S. D. Farnham, S. M. Drucker, P. Kollock. The effect of communication
modality on cooperation in online environments. Proceedings of the SIGCHI conference on
Human Factors in Computing Systems, 2000, pp. 470-477.
[128] S. Chaiken, A.H. Eagly. Communication modality as a determinant of persuasion: The role
of communicator salience. Journal of personality and social psychology. 45 (1983) 241.
[129] C. de Melo, P. J. Carnevale, and J. Gratch. Humans vs. Computers: The Effect of
Perceived Agency on People’s Decision Making. Marshall School of Buisiness and Institute for
Creative Technologies. University of Southern California. (2013).
[130] L. Gong. How social is social responses to computers? The function of the degree of
anthropomorphism in computer representations. Computers in Human Behavior. 24 (2008) 1494-
1509.
[131] M. L. Greaney, E. Puleo, K. Sprunck-Harrild, G. G. Bennett, M. A. Cunningham, M. W.
Gillman, M. Coeling, K. M. Emmons. Electronic reminders for cancer prevention: factors
associated with preference for automated voice reminders or text messages. Preventive medicine.
5 (2012) 151-154.
`
176

[132] J. Crawford, E. Larsen-Cooper, Z. Jezman, S. C. Cunningham, E. Bancroft. SMS versus
voice messaging to deliver MNCH communication in rural Malawi: assessment of delivery
success and user experience. Global health, science and practice. 2 (2014) 35-46.
[133] K. L. Nowak, J. Watt, J. B. Walther. Computer mediated teamwork and the efficiency
framework: Exploring the influence of synchrony and cues on media satisfaction and outcome
success. Computers in Human Behavior. 25 (2009) 1108-1119.
[134] B. Reeves, C. Nass. The Media Equation: how people treat computers, televisions and new
media like real people and places. Cambridge University Press. (1996).
[135] C. Nass. Etiquette equality: exhibitions and expectations of computer politeness.
Communications of the ACM. 47 (2004) 35-37.
[136] I. Jonsson, M. Zajicek, H. Harris, C. Nass. Matching in-car voice with driver state: Impact
on attitude and driving performance. Proceedings of the Third International Driving Symposium
on Human Factors in Driver Assessment, Training and Vehicle Design. 2005, pp. 173-180.
[137] I. Jonsson, C. Nass, H. Harris, L. Takayama. Thank you, I did not see that: in-car speech
based information systems for older adults.CHI'05 Extended Abstracts on Human Factors in
Computing Systems. 2005, pp. 1953-1956.
[138] A. Baylor. Beyond butlers: Intelligent agents as mentors. Journal of Educational
Computing Research. 22 (2000) 373-382.
[139] T. Bickmore, A. Gruber, R. Picard. Establishing the computer–patient working alliance in
automated health behavior change interventions. Patient education and counseling. 59 (2005) 21-
30.
[140] D. DeVault, R. Artstein, G. Benn, T. Dey, A. Egan, E. Fast. SimInterviewer: A Virtual
human interviewer for health decision support. Proceedings of Automated Agents and
Multiagent Systems, Paris, France, 2014.
[141] T. W. Bickmore, R. W. Picard. Establishing and maintaining long-term human-computer
relationships. ACM Transactions on Computer-Human Interaction (TOCHI). 12 (2005) 293-327.
`
177

[142] J. Bates. The role of emotion in believable agents. Communications of the ACM. 37
(1994) 122-125.
[143] J. Fox, S. J. Ahn, J. H. Janssen, L. Yeykelis, K. Y. Segovia, J. N. Bailenson. Avatars
versus agents: a meta-analysis quantifying the effect of agency on social influence. Human–
Computer Interaction. 30 (2015) 401-432.
[144] R. Schroeder. Social interaction in virtual environments: Key issues, common themes, and
a framework for research. Anonymous The Social Life of Avatars, Springer, 2002, pp.1-18.
[145] J. Blascovich, J. Loomis, A. C. Beall, K. R. Swinth, C. L. Hoyt, J. N. Bailenson.
Immersive virtual environment technology as a methodological tool for social psychology.
Psychological inquiry. 13 (2002) 103-124.
[146] S. Lim, B. Reeves. Computer agents versus avatars: Responses to interactive game
characters controlled by a computer or other player. International Journal of Human-Computer
Studies. 68 (2010) 57-68.
[147] S. Y. Okita, J. Bailenson, D. L. Schwartz. The mere belief of social interaction improves
learning. Proceedings of the Cognitive Science Society, 2007, pp. 1355-1360
[148] E. J. de Visser, F. Krueger, P. McKnight, S. Scheid, M. Smith, S. Chalk, R. Parasuraman.
The world is not enough: Trust in cognitive agents. Proceedings of the Human Factors and
Ergonomics Society Annual Meeting, Los Angeles, CA: Sage Publications, 2012, pp. 263-267.
[149] C. L. Hoyt, J. Blascovich, K. R. Swinth. Social inhibition in immersive virtual
environments. Presence: Teleoperators and Virtual Environments. 12 (2003) 183-195.
[150] C. M. Kennedy, J. Powell, T. H. Payne, J. Ainsworth, A. Boyd, I. Buchan. Active
assistance technology for health-related behavior change: an interdisciplinary review. Journal of
medical Internet research. 14 (2012) e80.
[151] S. Garrod, M.J. Pickering. Why is conversation so easy? Trends in cognitive sciences. 8
(2004) 8-11.
`
178

[152] J. Seikkula, D. Trimble. Healing elements of therapeutic conversation: Dialogue as an
embodiment of love. Family process. 44 (2005) 461-475.
[153] G. Cowan, A. Arsenault. Moving from monologue to dialogue to collaboration: The three
layers of public diplomacy. The Annals of the American Academy of Political and Social
Science. 616 (2008) 10-30.
[154] J. K. Burgoon, D. B. Buller, K. Floyd. Does participation affect deception success? Human
Communication Research. 27 (2001) 503-534.
[155] J. Cassell, T. Bickmore, M. Billinghurst, L. Campbell, K. Chang, H. Vilhjálmsson, H. Yan.
Embodiment in conversational interfaces: Rea. Proceedings of the SIGCHI Conference on
Human Factors in Computing Systems, Pittsburgh, Pennsylvania, USA, May 1999, pp. 520-527.
[156] D. J. Howard. The influence of verbal responses to common greetings on compliance
behavior: The foot‐in‐the‐mouth effect. Journal of Applied Social Psychology. 20 (1990) 1185-
1196.
[157] D. Dolinski, M. Nawrat, I. Rudak. Dialogue involvement as a social influence technique.
Personality and Social Psychology Bulletin. 27 (2001) 1395-1406.
[158] D. Dolinski, T. Grzyb, J. Olejnik, S. Prusakowski, K. Urban. Let's dialogue about penny:
Effectiveness of dialogue involvement and legitimizing paltry contribution techniques. Journal of
Applied Social Psychology. 35 (2005) 1150-1170.
[159] W. J. Turner, T. Hong. A technical framework to describe occupant behavior for building
energy simulations. BECC Conference, Sacramento, CA, USA, July 2013.
[160] S. Z. Attari, M. L. DeKay, C. I. Davidson, W. Bruine de Bruin. Public perceptions of
energy consumption and savings. Proceedings of the National Academy of Sciences of the
United States of America, 2010, pp. 16054–16059.
[161] R. Mourik, E. Heiskanen. Past 10 year of Best and Bad Practices in Demand Management:
a Meta-analysis of 27 Case Studies Focusing on Conditions Explaining Success and Failure of
`
179

Demand-side Management. Deliverable 4 of Changing Behaviour Project (GA 213217) Co-
funded by the European Commission within the Seventh Framework Programme. (2009).
[162] K. Bachus, L.Van Ootegem. Determinants of energy saving behaviour by households.
Research paper INESPO, HIVA-K.U.Leuven. (2011).
[163] O. Sapci, T. Considine. The link between environmental attitudes and energy consumption
behavior. Journal of Behavioral and Experimental Economics. 52 (2014) 29-34.
[164] J. Thøgersen, F Ölander. Human values and the emergence of a sustainable consumption
pattern: A panel study. Journal of Economic Psychology. 23 (2002) 605-630.
[165] D. G. Karp. Values and their effect on pro-environmental behavior. Environment and
Behavior. 28 (1996) 111-133.
[166] N. S. M. Azizi, S. Wilkinson. Motivation Factors in Energy Saving Behaviour between
Occupants in Green and Conventional Buildings—Malaysia Case Study. International Journal of
Environmental Science and Development. 6 (2015) 491.
[167] I. G. Monfared, S. Sharples. Occupants’ perceptions and expectations of a green office
building: a longitudinal case study. Architectural Science Review. 54 (2011) 344-355.
[168] L. Lan, Z. Lian, W. Liu, Y. Liu. Investigation of gender difference in thermal comfort for
Chinese people. European journal of applied physiology. 102 (2008) 471-480.
[169] L. Schellen, M. Loomans, M. d. de Wit, B. Olesen, W. van Marken Lichtenbelt. The
influence of local effects on thermal sensation under non-uniform environmental conditions—
Gender differences in thermophysiology, thermal comfort and productivity during convective
and radiant cooling. Physiology & Behavior. 107 (2012) 252-261.
[170] L. Schellen, W. van Marken Lichtenbelt, M. Loomans, J. Toftum, M. De Wit. Differences
between young adults and elderly in thermal comfort, productivity, and thermal physiology in
response to a moderate temperature drift and a steady‐state condition. Indoor air. 20 (2010) 273-
283.
`
180

[171] S. Karjalainen. Gender differences in thermal comfort and use of thermostats in everyday
thermal environments. Building and Environment. 42 (2007) 1594-1603.
[172] H. Liao, T. Chang. Space-heating and water-heating energy demands of the aged in the
US. Energy Economics. 24 (2002) 267-284.
[173] A. Paivio. Personality, success and failure, and preferred degree of public exposure.
Bulletin of the Maritime Psychological Association. 10 (1961) 50-58.
[174] C. Nass, K.M. Lee. Does computer-synthesized speech manifest personality? Experimental
tests of recognition, similarity-attraction, and consistency-attraction. Journal of Experimental
Psychology: Applied. 7 (2001) 171.
[175] D. Albarracín, J. C. Gillette, A. N. Earl, L. R. Glasman, M. R. Durantini, M. A test of
major assumptions about behavior change: a comprehensive look at the effects of passive and
active HIV-prevention interventions since the beginning of the epidemic. Psychological Bulletin.
131 (2005) 856-897.
[176] M. Shen, Q. Cui, L. Fu. Personality traits and energy conservation. Energy Policy. 85
(2015) 322-334.
[177] S. J. G. Ahn, J. Fox, K. R. Dale, J. A. Avant. Framing virtual experiences effects on
environmental efficacy and behavior over time. Communication Research. 42 (2015) 839-863.
[178] S. J. Ahn. Embodied experiences in immersive virtual environments: effects on pro-
environmental attitude and behavior. Doctoral dissertation, Stanford University, Palo Alto, CA.
2011.
[179] J. N. Bailenson, J. Blascovich, A. C. Beall, J. M. Loomis. Interpersonal distance in
immersive virtual environments. Personality & social psychology bulletin. 29 (2003) 819-833.
[180] C. Midden, J. Ham. The persuasive effects of positive and negative social feedback from
an embodied agent on energy conservation behavior. AISB 2008 Convention Communication,
Interaction and Social Intelligence, Aberdeen, Scotland, 2008.
`
181

[181] J. O. Bailey, J. N. Bailenson, J. Flora, K. C. Armel, D. Voelker, B. Reeves. The Impact of
Vivid Messages on Reducing Energy Consumption Related to Hot Water Use. Environment and
Behavior. 47 (2015) 570-592.
[182] J. M. Loomis, J. J. Blascovich, A. C. Beall. Immersive virtual environment technology as a
basic research tool in psychology. Behavior Research Methods, Instruments, & Computers. 31
(1999) 557-564.
[183] J. W. Smith. Immersive virtual environment technology to supplement environmental
perception, preference and behavior research: A review with applications. International journal
of environmental research and public health. 12 (2015) 11486-11505.
[184] T. S. Mujber, T. Szecsi, M. S. Hashmi. Virtual reality applications in manufacturing
process simulation. Journal of Materials Processing Technology. 155 (2004) 1834-1838.
[185] J. Psotka. Immersive training systems: Virtual reality and education and training.
Instructional science. 23 (1995) 405-431.
[186] A. Rizzo, J. G. Buckwalter, C. van der Zaag, U. Neumann, M. Thiébaux, C. Chua, A. van
Rooyen, L. Humphrey, P. Larson. Virtual environment applications in clinical neuropsychology.
Virtual Reality, 2000. Proceedings. IEEE, New Brunswick, NJ, USA, March 2000, pp. 63-70.
[187] N. Yee, J. N. Bailenson, K. Rickertsen. A meta-analysis of the impact of the inclusion and
realism of human-like faces on user experiences in interfaces. Proceedings of the SIGCHI
conference on Human factors in computing systems, San Jose, California, USA, 2007, 1-10.
[188] S. J. G. Ahn, J. N. Bailenson, D. Park. Short-and long-term effects of embodied
experiences in immersive virtual environments on environmental locus of control and behavior.
Computers in Human Behavior. 39 (2014) 235-245.
[189] A. Heydarian, J. P. Carneiro, D. Gerber, B. Becerik-Gerber, T. Hayes, W. Wood.
Immersive virtual environments: experiments on impacting design and human building
interaction. In Rethinking Comprehensive Design: Speculative Counterculture, Proceedings of
the 19th International Conference on Computer-Aided Architectural Design Research in Asia
(CAADRIA), 2014, pp. 729-738.
`
182

[190] C. Chan, C. Weng. How Real Is the Sense of Presence in A Virtual Environment?:
Applying Protocol Analysis for Data Collection. Digital Opportunities: Proceedings of the 10th
International Conference on Computer-Aided Architectural Design Research in Asia, April, New
Delhi, India, 2005.
[191] A. Heydarian, J. P. Carneiro, D. Gerber, B. Becerik-Gerber, T. Hayes, W. Wood.
Immersive virtual environments versus physical built environments: A benchmarking study for
building design and user-built environment explorations. Automation in Construction. 54 (2015)
116-126.
[192] M. Slater, S. Wilbur. A framework for immersive virtual environments (FIVE):
Speculations on the role of presence in virtual environments. Presence: Teleoperators and virtual
environments. 6 (1997) 603-616.
[193] D. A. Bowman, A. Datey, Y. S. Ryu, U. Farooq, O. Vasnaik. Empirical comparison of
human behavior and performance with different display devices for virtual environments.
Proceedings of the human factors and ergonomics society annual meeting, 2002, pp. 2134-2138.
[194] B. G. Witmer, M. J. Singer. Measuring presence in virtual environments: A presence
questionnaire. Presence: Teleoperators and virtual environments. 7 (1998) 225-240.
[195] M. Slater, V. Linakis, M. Usoh, R. Kooper, G. Street. Immersion, presence, and
performance in virtual environments: An experiment with tri-dimensional chess. ACM virtual
reality software and technology (VRST), ACM Press, New York, 1996, pp. 163.
[196] K. Kim, M. Z. Rosenthal, D. J. Zielinski, R. Brady. Effects of virtual environment
platforms on emotional responses. Computer methods and programs in biomedicine. 113 (2014)
882-893.
[197] S. Walkowiak, T. Foulsham, A. F. Eardley. Individual differences and personality
correlates of navigational performance in the virtual route learning task. Computers in Human
Behavior. 45 (2015) 402-410.
[198] P. Figueroa, W. F. Bischof, P. Boulanger, H. J. Hoover. Efficient comparison of platform
alternatives in interactive virtual reality applications. Efficient comparison of platform
`
183

alternatives in interactive virtual reality applications. International journal of human-computer
studies. 62 (2005) 73-103.
[199] S. Khashe, A. Heydarian, B. Becerik-Gerber, W. Wood. Exploring the effectiveness of
social messages on promoting energy conservation behavior in buildings. Building and
Environment. 102 (2016) 83-94.
[200] R. S. Kennedy, N. E. Lane, K. S. Berbaum, M. G. Lilienthal. Simulator sickness
questionnaire: An enhanced method for quantifying simulator sickness. The international journal
of aviation psychology. 3 (1993) 203-220.
[201] M. Slater, M. Usoh, A. Steed. Depth of presence in virtual environments. Presence:
Teleoperators & Virtual Environments. 3 (1994) 130-144.
[202] T. Schubert, F. Friedmann, H. Regenbrecht. Embodied presence in virtual environments.
Anonymous Visual Representations and Interpretations, Springer, 1999, pp. 269-278.
[203] T. Schubert, F. Friedmann, H. Regenbrecht. The experience of presence: Factor analytic
insights. Presence: Teleoperators and virtual environments. 10 (2001) 266-281.
[204] M. Slater, J. McCarthy, F. Maringelli. The influence of body movement on subjective
presence in virtual environments. Human factors. 40 (1998) 469-477.
[205] B. Keshavarz, H. Hecht. Visually induced motion sickness and presence in videogames:
The role of sound. Proceedings of the Human Factors and Ergonomics Society Annual Meeting,
56 (2012) 1763-1767.
[206] C. J. Jerome, B. Witmer. Immersive tendency, feeling of presence, and simulator sickness:
formulation of a causal model. Proceedings of the Human Factors and Ergonomics Society
Annual Meeting, 46 (2002) 2197-2201.
[207] R. M. Baños, C. Botella, I. Rubió, S. Quero, A. García-Palacios, M. Alcañiz. Presence and
emotions in virtual environments: The influence of stereoscopy. CyberPsychology & Behavior.
11 (2008) 1-8.
`
184

[208] K. Mania, A. Chalmers. The effects of levels of immersion on memory and presence in
virtual environments: A reality centered approach. CyberPsychology & Behavior. 4 (2001) 247-
264.
[209] C. Lo Priore, G. Castelnuovo, D. Liccione, D. Liccione. Experience with V-STORE:
considerations on presence in virtual environments for effective neuropsychological
rehabilitation of executive functions. CyberPsychology & Behavior. 6(2003) 281-287.
[210] R. P. McMahan, D. Gorton, J. Gresock, W. McConnell, D. A. Bowman. Separating the
effects of level of immersion and 3D interaction techniques. Proceedings of the ACM
symposium on Virtual reality software and technology, Limassol, Cyprus, November 2006, pp.
108-111.
[211] S. Persky, J. Blascovich. Immersive virtual video game play and presence: Influences on
aggressive feelings and behavior. Presence: Teleoperators and Virtual Environments. 17 (2008)
57-72.
[211] S. Sharples, S. Cobb, A. Moody, J. R. Wilson. Virtual reality induced symptoms and
effects (VRISE): Comparison of head mounted display (HMD), desktop and projection display
systems. Displays. 29 (2008) 58-69.
[213] C. Regan. An investigation into nausea and other side-effects of head-coupled immersive
virtual reality. Virtual Reality. 1 (1995) 17-31.
[214] J Barrett. Side effects of virtual environments: A review of the literature. Technical Report,
No. N9505/21/155. May 2004.
[215] D. R. Baltzley, R. S. Kennedy, K. S. Berbaum, M. G. Lilienthal, D. W. Gower. The time
course of post flight simulator sickness symptoms. Aviation, Space, and Environmental
Medicine. 60 (1989) 1043-1048.
[216] J. Brade, M. Lorenz, M. Busch, N. Hammer, M. Tscheligi, P. Klimant. Being there again–
Presence in real and virtual environments and its relation to usability and user experience using a
mobile navigation task. International Journal of Human-Computer Studies. 101 (2017) 76-87.
`
185

[217] M. Schuemie, B. Abel, C. van der Mast, M. Krijn, P. Emmelkamp. The effect of
locomotion technique on presence, fear and usability in a virtual environment. Proceeding of
Euromedia, Toulouse, France, 2005, pp. 129-135,.
[218] T. Vogt, R. Herpers, C. D. Askew, D. Scherfgen, H. K. Strüder, S. Schneider. Effects of
exercise in immersive virtual environments on cortical neural oscillations and mental state.
Neural plasticity. (2015).
[219] E. Nystad, A. Sebok. A comparison of two presence measures based on experimental
results. Proceedings of the seventh international workshop on presence. (2004) 266-273.
[220] P. Banerjee, G. M. Bochenek, J. M. Ragusa. Analyzing the relationship of presence and
immersive tendencies on the conceptual design review process. Journal of Computing and
Information Science in Engineering (Transactions of the ASME). 2 (2002) 59-64.
[221] D. Weibel, B. Wissmath. Immersion in computer games: The role of spatial presence and
flow. International Journal of Computer Games Technology. (2011) 6.
[222] J. L. Kirk. Sustainable Environments and Pro-Environmental Behavior. Master’s thesis,
University of Nebraska, Lincoln, Lincoln, Nebraska. (2010).
[223] S. Wilkinson, M. Rashid, K. Spreckelmeyer, N. J. Angrisano. Green buildings,
environmental awareness, and organizational image. Journal of Corporate Real Estate. 14 (2012)
21-49.
[224] J. Cidell. A political ecology of the built environment: LEED certification for green
buildings. Local Environment. 14 (2009) 621-633.
[225] U.S. Green Building Council (USGBC). USGBC Directory. Last accessed: July 2015,
available at: http://www.usgbc.org/projects.
[226] W. H. Golove, J. H. Eto. Market barriers to energy efficiency: a critical reappraisal of the
rationale for public policies to promote energy efficiency. LBL-38059.Berkeley, CA: Lawrence
Berkeley National Laboratory. (1996).
`
186

[227] P. J. May, C. Koski. State environmental policies: analyzing green building mandates.
Review of policy research. 24 (2007) 49-65.
[228] J. Heerwagen. Green buildings, organizational success and occupant productivity. Building
Research & Information. 28 (2000) 353-367.
[229] Y. S. Lee, D. A. Guerin. Indoor environmental quality related to occupant satisfaction and
performance in LEED-certified buildings. Indoor and Built Environment. 18 (2009) 293-300.
[230] G. Newsham, B. Birt, C. Arsenault, L. Thompson, J. Veitch, S. Mancini, A. Galasiu, B.
Gover, I. Macdonald, G. Burns. Do green buildings outperform conventional buildings? Indoor
environment and energy performance in North American offices. National Research Council
Canada. 329 (2012) 1-71
[231] Z. M. Gill, M. J. Tierney, I. M. Pegg, N. Allan. Low-energy dwellings: the contribution of
behaviours to actual performance. Building Research & Information. 38 (2010) 491-508.
[232] S. Browne, I. Frame. Green buildings need green occupants. Corporate Social-
Responsibility and Environmental Management. 6 (1999) 80-85.
[233] P. Hartmann, V. Apaolaza Ibáñez. Green value added. Marketing Intelligence & Planning.
24 (2006) 673-680.
[234] Z. Staff. The lighting handbook. Austria: Zumtobel. (2004).
[235] S. K. Datta. Pro-environmental concern influencing green buying: A study on Indian
consumers. International Journal of Business and management. 6 (2011) 124-133.
[236] J. Kahn. Reporting statistics in APA style. Last accessed: June 2015, Available at:
http://my.ilstu.edu/~jhkahn/apastats.html.
[237] T. W. Bickmore, C. Sidner. Towards Plan-based Health Behavior Change Counseling
Systems. AAAI Spring Symposium: Argumentation for Consumers of Healthcare, 2006, pp. 14-
18.
`
187

[238] J. C. Mowen, R.B. Cialdini. On implementing the door-in-the-face compliance technique
in a business context. Journal of Marketing Research. (1980) 253-258.
[239] G. Y. Yun, H. Kim, J. T. Kim. Effects of occupancy and lighting use patterns on lighting
energy consumption. Energy and Buildings. 46 (2012) 152-158.
[240] K. V. Rhoads, R.B. Cialdini. The business of influence: Principles that lead to success in
commercial settings. The persuasion handbook: Developments in theory and practice. (2002)
513-542.
[241] D. Malhotra, M.H. Bazerman. Psychological influence in negotiation: An introduction long
overdue. Journal of Management. 34 (2008) 509-531.
[242] O.P. John, S. Srivastava. The Big Five trait taxonomy: History, measurement, and
theoretical perspectives. Handbook of personality: Theory and research. 2 (1999) 102-138.
[243] T. M. Umaki, M. R. Umaki, C. M. Cobb. The psychology of patient compliance: a focused
review of the literature. Journal of periodontology. 83 (2012) 395-400.
[244] R. R. McCrae, O.P. John. An introduction to the five-factor model and its applications.
Personality: critical concepts in psychology. Journal of personality. 60 (1992) 175-215.
[245] M. Emilsson, I. Berndtsson, J. Lotvall, E. Millqvist, J. Lundgren, A. Johansson, E. Brink.
The influence of personality traits and beliefs about medicines on adherence to asthma treatment.
Primary care respiratory journal. 20 (2011) 141-147.
[246] R. C. Hilliard, B. W. Brewer, A. E. Cornelius, J. L. Van Raalte. Big Five Personality
Characteristics and Adherence to Clinic-Based Rehabilitation Activities After ACL Surgery: A
Prospective Analysis. The open rehabilitation journal. 7 (2014) 1-5.
[247] T. L. Milfont, C.G. Sibley. The big five personality traits and environmental engagement:
Associations at the individual and societal level. Journal of Environmental Psychology. 32
(2012) 187-195.
`
188

[248] S. E. Hampson. Personality processes: mechanisms by which personality traits "get outside
the skin". Annual Review of Psychology. 63 (2012) 315-339.
[249] S. Karjalainen. Consumer preferences for feedback on household electricity consumption.
Energy and Buildings. 43 (2011) 458-467.
[250] R. K. Jain, J. E. Taylor, G. Peschiera. Assessing eco-feedback interface usage and design
to drive energy efficiency in buildings. Energy and Buildings. 48 (2012) 8-17.
[251] S. Maan, B. Merkus, J. Ham, C. Midden. Making it not too obvious: the effect of ambient
light feedback on space heating energy consumption. Energy Efficiency. 4 (2011) 175-183.
[252] H. Chen, C. Lin, S. Hsieh, H. Chao, C. Chen, R. Shiu, S. Ye, Y. Deng. Persuasive
feedback model for inducing energy conservation behaviors of building users based on
interaction with a virtual object. Energy and Buildings. 45 (2012) 106-115.
[253] I. Chen, F. Tseng. The relevance of communication media in conflict contexts and their
effectiveness: A negotiation experiment. Computers in Human Behavior. 59 (2016) 134-141.
[254] A. R. Dennis, J. S. Valacich. Rethinking media richness: Towards a theory of media
synchronicity. Systems Sciences. HICSS-32. Proceedings of the 32nd Annual Hawaii
International Conference, Maui, HI, USA, 1999.
[255] E. J. Lee, C. Nass, S. Brave. Can computer-generated speech have gender?: an
experimental test of gender stereotype. CHI'00 extended abstracts on Human factors in
computing systems, The Hague, The Netherlands, 2000, pp. 289-290.
[256] C. Nass, Y. Moon, N. Green. Are Machines Gender Neutral? Gender‐Stereotypic
Responses to Computers With Voices. Journal of Applied Social Psychology. 27 (1997) 864-
876.
[257] D. J. Reid, F.J. Reid. Text or talk? Social anxiety, loneliness, and divergent preferences for
cell phone use. CyberPsychology & Behavior. 10 (2007) 424-435.
`
189

[258] M. R. Leary, J. M. Tipsord, E. B. Tate. Allo-inclusive identity: Incorporating the social and
natural worlds into one's sense of self. In H. A. Wayment & J. J. Bauer (Eds.), Decade of
behavior. Transcending self-interest: Psychological explorations of the quiet ego. (2008) 137-
147.
[259] A. Parasuraman. Technology Readiness Index (TRI) a multiple-item scale to measure
readiness to embrace new technologies. Journal of service research. 2 (2000) 307-320.
[260] A. L. Baylor. Promoting motivation with virtual agents and avatars: role of visual presence
and appearance. Philosophical transactions of the Royal Society of London. Series B, Biological
sciences. 364 (2009) 3559-3565.
[261] A. Baylor, J. Ryu, E. Shen. The effects of pedagogical agent voice and animation on
learning, motivation and perceived persona. EdMedia: World Conference on Educational Media
and Technology, Honolulu, Hawaii, USA, 2003, pp. 452-458.
[262] B. S. Kisilevsky, S. M. Hains, C. A. Brown, C. T. Lee, B. Cowperthwaite, S. S. Stutzman,
M. L. Swansburg, K. Lee, X. Xie, H. Huang. Fetal sensitivity to properties of maternal speech
and language. Infant Behavior and Development. 32 (2009) 59-71.
[263] G.Y. Lee, B.S. Kisilevsky. Fetuses respond to father's voice but prefer mother's voice after
birth. Developmental psychobiology. 56 (2014) 1-11.
[264] E. A. Plant, A. L. Baylor, C. E. Doerr, R. B. Rosenberg-Kima. Changing middle-school
students’ attitudes and performance regarding engineering with computer-based social models.
Computers & Education. 53 (2009) 209-215.
[265] L. L. Carli. Gender, interpersonal power, and social influence. Journal of Social Issues. 55
(1999) 81-99.
[266] R. E. Guadagno, J. Blascovich, J. N. Bailenson, C. Mccall. Virtual humans and persuasion:
The effects of agency and behavioral realism. Media Psychology. 10 (2007) 1-22.
`
190

[267] P. Kulms, N. Krämer, J. Gratch, S. Kang. It’s in their eyes: A study on female and male
virtual humans’ gaze. Intelligent Virtual Agents, 10th International Conference, Reykjavik,
Iceland, 2011, pp. 80-92
[268] D. R. Dunaetz, T. C. Lisk, M. M. Shin. Personality, Gender, and Age as Predictors of
Media Richness Preference. Advances in Multimedia. (2015).
[269] J. Gratch, N. Wang, A. Okhmatovskaia, F. Lamothe, M. Morales, R.J. van der Werf, L. P.
Morency. Can virtual humans be more engaging than real ones? In: Jacko J.A. (eds) Human-
Computer Interaction. HCI Intelligent Multimodal Interaction Environments. HCI 2007. Lecture
Notes in Computer Science, Springer, Berlin, Heidelberg, 2007, pp. 286-297.
[270] M. Buhrmester, T. Kwang, S. D. Gosling. Amazon's Mechanical Turk: A New Source of
Inexpensive, Yet High-Quality, Data? Perspectives on psychological science : a journal of the
Association for Psychological Science. 6 (2011) 3-5.
[271] S. Khashe, G. Lucas, B. Becerik-Gerber, J. Gratch. Buildings with Persona: Towards
Effective Building-Occupant Communication. Computers in Human Behavior. 75 (2017) 607-
618.
[272] A. Parasuraman, C.L. Colby. An updated and streamlined technology readiness index: TRI
2.0. Journal of service research. 18 (2015) 59-74.
[273] T. Bickmore, J. Cassell. Social dialogue with embodied conversational agents. Advances
in natural multimodal dialogue systems. 30 (2005) 23-54.
[274] R. B. Cialdini. Influence: Science and Practice, 4th ed. Champaign IL: Omegatype
Typography, Inc. (2001).
[275] Y. J. Weisberg, C. G. Deyoung, J. B. Hirsh. Gender Differences in Personality across the
Ten Aspects of the Big Five. Frontiers in psychology. 2 (2011)178.
[276] A. R. Krentzman, K. J. Brower, J. A. Cranford, J. C. Bradley, E. A. Robinson. Gender and
extroversion as moderators of the association between Alcoholics Anonymous and sobriety.
Journal of studies on alcohol and drugs. 73 (2012) 44-52.
`
191

[277] J. Veroff, E. A. M. Douvan, R. A. Kulka. The inner American: A self-portrait from 1957 to
1976. New York: Basic Books. (1981).
[278] T. Bickmore, J. Cassell. Relational agents: a model and implementation of building user
trust. Proceedings of the SIGCHI conference on Human factors in computing systems, Seattle,
WA, USA, 2001, pp. 396-403.
[279] H. T. Hurt, K. Joseph, C. D. Cook. Scales for the measurement of innovativeness. Human
Communication Research. 4 (1977) 58-65.
[280] J. B. Thatcher, P. L. Perrewe. An empirical examination of individual traits as antecedents
to computer anxiety and computer self-efficacy. MIS quarterly. MIS Quarterly. 26 (2002) 381-
396.
[281] A. Bogdanovych, H. Berger, S. Simoff, C. Sierra. Travel agents vs. online booking:
Tackling the shortcomings of nowadays online tourism portals. Information and communication
technologies in tourism. (2006) 418-428.
[282] S. E. Jackson, A. Steptoe, J. Wardle. The influence of partner’s behavior on health
behavior change: the English Longitudinal Study of Ageing. JAMA internal medicine. 175
(2015) 385-392.
[283] T.A. Falba, J.L. Sindelar. Spousal concordance in health behavior change. Health services
research. 43 (2008) 96-116.
[284] P.M. Doney, J. P. Cannon. An examination of the nature of trust in buyer - seller
relationships. Journal of Marketing. 61 (1997) 35-51.
[285] D. McKenzie-Mohr. Fostering sustainable behavior: An introduction to community-based
social marketing. 3
rd
Edition. New society publishers. (2011).
[286] J. E. Barlett, J. W. Kotrlik, C. C. Higgins. Organizational research: Determining
appropriate sample size in survey research. Information technology, learning, and performance
journal. 19 (2001) 43.
`
192

[287] W. Abrahamse, L. Steg, C. Vlek, T. Rothengatter. The effect of tailored information, goal
setting, and tailored feedback on household energy use, energy-related behaviors, and behavioral
antecedents. Journal of Environmental Psychology. 27 (2007) 265-276.
[288] F. C. Bull, C. Holt, M. Kreuter, E. Clark, D. Scharff. Understanding the effects of printed
health education materials: which features lead to which outcomes? Journal of health
communication. 6 (2001) 265-280.
[289] L.S. Suggs. A 10-year retrospective of research in new technologies for health
communication. Journal of health communication. 11 (2006) 61-74.
[290] L. J. Trevena, A. Barratt, P. Butow, P. Caldwell. A systematic review on communicating
with patients about evidence. Journal of evaluation in clinical practice. 12 (2006) 13-23.
[291] P. Ryan, D.R. Lauver. The efficacy of tailored interventions. Journal of Nursing
Scholarship. 34 (2002) 331-337.
[292] D. A. Hanauer, K. Wentzell, N. Laffel, L. M. Laffel. Computerized Automated Reminder
Diabetes System (CARDS): e-mail and SMS cell phone text messaging reminders to support
diabetes management. Diabetes technology & therapeutics. 11 (2009) 99-106.
[293] C. Steiner, E. D. Mekler, K. Opwis. Personalised persuasion–What are the most effective
user data for persuasion profiling? Bachelor’s Thesis, University of Basel, Department of
Psychology, Basel, Switzerland. (2013).
[294] A. Heydarian, J. P. Carneiro, D. Gerber, B. Becerik-Gerber. Immersive virtual
environments, understanding the impact of design features and occupant choice upon lighting for
building performance. Building and Environment. 89 (2015) 217-228.
[295] M. Argyle. Bodily communication (2nd ed.). London: Routledge. Brannigan, C. R., &
Humphries, D. A. Human non-verbal behavior, a means of communication. In N. G. Blurton-
Jones (Ed.), Ethological studies of child behavior. Cambridge: Cambridge University Press.
(1972) 37-64.
`
193

[296] T. Bereczkei, B. Birkas, Z. Kerekes. The presence of others, prosocial traits,
machiavellianism. Social Psychology. 41 (2015) 238-245.
[297] Y. Chen, X. Liang, T. Hong, X. Luo. Simulation and visualization of energy-related
occupant behavior in office buildings. Building Simulation. 10 (2017) 785–798.
[298] D. Mora, C. Carpino, M. De Simone. Behavioral and physical factors influencing energy
building performances in Mediterranean climate. Energy Procedia. 78 (2015) 603-608.
[299] Y. Duan, B. Dong. The Contribution of Occupancy Behavior to Energy Consumption In
Low Income Residential Buildings. Proceedings of the 3rd International High Performance
Buildings Conference, West Lafayette, USA, 2014.
[300] EPA Environmental Protection Agency. Climate Change: Basic Information. Last
accessed: May 2016. Available at: http://www.epa.gov/climatechange/basics/.
[301] EPA Environmental Protection Agency. Definition on Green Building. Last accessed:
April 2016. Available at: www.epa.gov/greenbuilding/pubs/about.htm.
[302] E. Maslowska, E. Smit, B. van den Putte. Examining the (in) effectiveness of personalized
communication. Proceedings of the 10th International Conference on Research in Advertising,
M. Eisend and T. Langner, eds., Berlin, Germany: European Advertising Academy, 2011, pp. 1–
12.
[303] R. G. McFarland, G. N. Challagalla, T. A. Shervani. Influence tactics for effective adaptive
selling. Journal of Marketing. 70 (2006) 103-117.

`
194

Appendixes
Appendix 1: Green buildings and LEED certification
Over the past century, human activities have released large amounts of carbon dioxide and other
greenhouse gases into the atmosphere. Greenhouse gases act like a blanket around the Earth,
trapping energy in the atmosphere and causing it to warm up. This phenomenon is called the
greenhouse effect and is natural and necessary to support life on Earth. However, the build up of
greenhouse gases can change Earth's climate and result in dangerous effects to human health and
welfare and to ecosystems [289]. It is estimated that at present, buildings contribute as much as
one third of total global greenhouse gas emissions, primarily through the use of fossil fuels during
their operational phase. Therefore, buildings have great impact on the natural environment, human
health, and the economy. One solution is adopting green building strategies, by which we can
maximize both economic and environmental performance of buildings. Green building design is
the practice of creating structures and using processes that are environmentally responsible and
resource-efficient throughout a building's life-cycle from design to construction, operation,
maintenance, renovation and deconstruction [290].
Green buildings are designed to reduce the overall impact of the built environment on human
health and natural environment by:
 Efficiently using energy, water, and other resources;
 Protecting occupant health and improving employee productivity;
 Reducing waste, pollution and environmental degradation.
For example, green buildings incorporate sustainable materials (e.g., reused, recycled-content, or
made from renewable resources) and dedicate places for collection of recyclable materials in order
`
195

to reduce the need for extracting, refining, and processing raw materials and reduce energy
consumption and greenhouse gas emissions. Green buildings create healthy indoor environments
by improving indoor air quality and reducing exposure levels for indoor pollutants (e.g., reduced
product emissions). Green buildings also decrease energy consumption by taking advantage of
natural lighting and ventilation and using efficient appliances. In addition, green buildings feature
landscaping that reduces water usage (e.g., by using native plants that survive without extra
watering).
To recognize how green a building is, LEED or Leadership in Energy & Environmental Design, a
green certification program, is implemented in the United States. LEED provides building owners
and operators with a framework for identifying and implementing practical and measurable green
building designs, construction, operations, and maintenance solutions. If a building project is
considered as LEED certified, it means that the building is truly "green" and has achieved its
environmental goals. There are four levels of certification, depending on the number of points a
project earns during certification: LEED certification, Silver LEED certification, Gold LEED
certification and Platinum LEED certification (the highest degree of certification). There are thirty
LEED certified, sixty Silver LEED certified, fifty five Gold LEED certified, and five Platinum
certified buildings in Los Angeles.
The building in our study is Gold LEED certified. Offices are decorated with reused furniture,
which is made from certified wood (wood products that come from responsibly managed forests).
The walls are painted with organic chemicals. The floors are carpeted with LEED labeled materials
(materials that meet the LEED certification requirements). The rooms have dedicated bins to
separate the recyclable trashes from non-recyclable ones. The building also takes advantage of
`
196

LED lighting and automated control systems, which improve energy efficiency and may result in
up to 30% electricity savings.”

Appendix 2: Allo-inclusive identity scale

Abstract (if available)

Abstract There are technological approaches that try to improve the performance of building systems. However, these technological improvements are only successful when implemented and conformed correctly by building occupants. In fact, occupants and their behaviors have significant impacts on a building's performance. In this research, we investigated behavior change intervention strategies that influence the way occupants behave in buildings. We specifically focused on pro-environmental behavior in commercial buildings. Pro-environmental behaviors refer to occupants' activities such as reducing energy consumption and recycling which reduce environmental negative impacts. We investigated effective design strategies for successful behavior change interventions aiming to promote pro-environmental behaviors as well as the factors that could impact the effects of these interventions. First, we explored the effects of Leadership in Energy and Environmental Design (LEED) branding on occupants' pro-environmental behaviors in buildings. LEED branding, in the context of this research, is defined as introducing the building as being LEED certified and emphasizing the aspects of the LEED building's green features to its users. This strategy took advantage of functional attributes and emotional benefits of LEED certified buildings to influence occupants' behaviors. We also investigated the effects of direct communication on occupants' behavior. We explored approaches to transform buildings into interactive living spaces that communicate with their occupants and create conditions for behavior change. We investigated the effects of incorporating social and relational features into the design of different components of the behavior change intervention. Social features included adapting social influence methods more commonly seen in face-to-face communication, as well as the surface social cues (e.g., face and voice). In addition, relational features included adaption of human-like persona, as well as interactive approaches such as social dialog. For our investigations, we used a single occupancy office, which was either a physical or a virtual environment (VE). Our results demonstrated LEED branding motivated occupants to act more pro-environmentally and communication was more persuasive when the interaction was more social and relational strategies were adapted. In fact, we found out that rules that apply to human-human interactions applied also to human-building interactions. Our findings emphasize the importance of occupants' perception of buildings. These results can be used to create an interactive environment, which influences occupants' relationships with the building. Our findings can also provide useful design choices for persuasive technologies aiming to promote pro-environmental behaviors.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Understanding human-building interactions through perceptual decision-making processes

PDF

Understanding human-building-emergency interactions in the built environment

PDF

Towards health-conscious spaces: building for human well-being and performance

PDF

Behavioral form finding using multi-agent systems: a computational methodology for combining generative design with environmental and structural analysis in architectural design

PDF

Towards social virtual listeners: computational models of human nonverbal behaviors

PDF

User-centric smart sensing for non-intrusive electricity consumption disaggregation in buildings

PDF

Computational modeling of human behavior in negotiation and persuasion: the challenges of micro-level behavior annotations and multimodal modeling

PDF

A framework for research in human-agent negotiation

PDF

Computational foundations for mixed-motive human-machine dialogue

PDF

Decoding information about human-agent negotiations from brain patterns

PDF

A framework for comprehensive assessment of resilience and other dimensions of asset management in metropolis-scale transport systems

PDF

Transitions in the collective behavior of microswimmers confined to microfluidic chips

PDF

Role enactment in interactive media: a role-play perspective

PDF

Artificial intelligence for low resource communities: Influence maximization in an uncertain world

PDF

Healthy mobility: untangling the relationships between the built environment, travel behavior, and environmental health

PDF

Developing an agent-based simulation model to evaluate competition in private health care markets with an assessment of accountable care organizations

PDF

Application of data-driven modeling in basin-wide analysis of unconventional resources, including domain expertise

PDF

The environmental and genetic determinants of cleft lip and palate in the global setting

Asset Metadata

Creator Khashe, Saba (author)

Core Title Enabling human-building communication to promote pro-environmental behavior in office buildings

School Viterbi School of Engineering

Degree Doctor of Philosophy

Degree Program Civil Engineering

Publication Date 01/26/2020

Defense Date 11/14/2017

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

Tag behavior change intervention,Communication,Compliance,delivery style,human-building interaction,OAI-PMH Harvest,persona,personalization,relational agents,social influence method,virtual environments

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Becerik-Gerber, Burcin (committee chair), Gratch, Jonathan (committee member), Lucas, Gale (committee member), Soibelman, Lucio (committee member)

Creator Email sabakhashe@gmail.com,skhashe@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-466512

Unique identifier UC11266711

Identifier etd-KhasheSaba-5973.pdf (filename),usctheses-c40-466512 (legacy record id)

Legacy Identifier etd-KhasheSaba-5973.pdf

Dmrecord 466512

Document Type Dissertation

Rights Khashe, Saba

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