Content uploaded by Stephan Diederich
Author content
All content in this area was uploaded by Stephan Diederich on Sep 26, 2019
Content may be subject to copyright.
This is the author’s version of a work that has been published in the following outlet:
Diederich, S.; Lichtenberg, S.; Brendel, A.B.; Trang, S. (2019): Promoting Sustainable Mobility Beliefs
with Persuasive and Anthropomorphic Design: Insights from an Experiment with a Conversational
Agent, Proceedings of International Conference on Information Systems (ICIS), Munich, Germany.
Chair of Information Management
Smart Mobility Research Group (SMRG)
Prof. Dr. Lutz M. Kolbe
Platz der Göttinger Sieben 5
37073 Göttingen – Germany
www.uni-goettingen.de/im
Dr. Alfred Benedikt Brendel
Humboldtallee 3
37073 Göttingen – Germany
www.uni-goettingen.de/smrg
Please note: The copyright is owned by the author and / or the publisher.
Commercial use is not allowed.
This work is licensed under a Creative Commons Attribution-NonCommercial-
NoDerivatives 4.0 International License.
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 1
Promoting Sustainable Mobility Beliefs with
Persuasive and Anthropomorphic Design:
Insights from an Experiment with a
Conversational Agent
Completed Research Paper
Stephan Diederich
University of Göttingen
Humboldtallee 3, 37073
Göttingen, Germany
stephan.diederich@
stud.uni-goettingen.de
Sascha Lichtenberg
University of Göttingen
Humboldtallee 3, 37073
Göttingen, Germany
sascha.lichtenberg@
uni-goettingen.de
Alfred Benedikt Brendel
University of Göttingen
Humboldtallee 3, 37073
Göttingen, Germany
abrende1@uni-goettingen.de
Simon Trang
University of Göttingen
Platz der Göttinger Sieben 5
37073 Göttingen, Germany
strang@uni-goettingen.de
Abstract
Sustainable mobility behavior is increasingly relevant due to the vast environmental
impact of current transportation systems. With the growing variety of transportation
modes, individual decisions for or against specific mobility options become more and
more important and salient beliefs regarding the environmental impact of different
modes influence this decision process. While information systems have been recognized
for their potential to shape individual beliefs and behavior, design-oriented studies that
explore their impact, in particular on environmental beliefs, remain scarce. In this
study, we contribute to closing this research gap by designing and evaluating a new
type of artifact, a persuasive and human-like conversational agent, in a 2x2 experiment
with 225 participants. Drawing on the Theory of Planned Behavior and Social Response
Theory, we find empirical support for the influence of persuasive design elements on
individual environmental beliefs and discover that anthropomorphic design can
contribute to increasing the persuasiveness of artifacts.
Keywords: Sustainable mobility, persuasive design, conversational agent
Introduction
Sustainable mobility behavior becomes more and more important nowadays due to the extensive
environmental impact of current transportation systems (Vergragt and Brown 2007) as well as the
increasing variety of transportation modes (Willing et al. 2017). On the individual level, people in urban
areas now have various mobility modes at their disposal, in particular shared ones, such as for cars,
bicycles or e-scooters (Rodrigue et al. 2013). However, it is still up to the individual to decide which
(sustainable) mobility mode to use, a decision that is influenced not only by pragmatic questions of
availability and travel time but also by the salient beliefs one holds regarding the different modes and
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 2
their environmental impact (Donald et al. 2014). According to the Belief-Action-Outcome framework by
Melville (2010), this belief precedes an individual’s action (e.g. to choose a car or bicycle for travel), which
ultimately leads to a certain outcome (e.g. adding to CO2 emissions or not). Information systems (IS)
have been recognized for their potential to influence individual beliefs and actions towards sustainable
behavior (vom Brocke et al. 2013; Butler 2011; Elliot 2011; Malhotra et al. 2013). In the context of
environmental sustainability, studies investigate how IS can promote sustainable mobility behavior with
the use of e-bikes (Flüchter and Wortmann 2014; Froehlich et al. 2009), encourage efficient energy use
(Corbett 2013a; Fitzpatrick and Smith 2009; Loock et al. 2013), or create transparency for different types
of environmentally friendly actions (Pitt et al. 2011). However, studies that explore the opportunity to
promote eco-friendly beliefs and behavior through IS design are still scarce (Henkel and Kranz 2018;
Seidel et al. 2017), focus primarily on action-formation (Brendel et al. 2017) and the question of how
individual environmental beliefs can be influenced through IS design in Melville's research agenda (2010,
RQ5) remains largely unanswered.
In this study, we contribute to closing this research gap by designing and evaluating an artifact that we
envisage to positively influence individual sustainability beliefs regarding the use of shared e-bikes in an
organizational setting. In contrast to prevalent studies that present persuasive designs of mobile apps or
websites in the context of environmental sustainability, we design a conversational agent (CA) with which
users interact through natural language. CAs, driven by machine learning and natural language
processing, currently attract strong interest in research and practice alike (McTear 2017; Oracle 2016),
both in the form of digital assistants (e.g. Siri or Alexa) as well as chatbots on company websites and
social media (Baier et al. 2018). From a theoretical perspective, CAs represent an interesting phenomenon
as users show substantial social responses to the human-like representation and behavior of such agents,
which in turn offers new opportunities for the anthropomorphic design of such agents (Seeger et al. 2018)
and their persuasiveness (Adler et al. 2016). In this study, we combine insights from human-like CA
design with prescriptive knowledge from persuasive systems for environmental sustainability to explore
the following research question: How can persuasive and human-like design of conversational agents
positively shape individual beliefs towards sustainable mobility behavior?
To address this question, we design a conversational agent and investigate to which extent it impacts
individual behavioral, normative, and control beliefs (Ajzen 1985, 1991) in the context of organizational e-
bike sharing by means of an online experiment with 225 participants and a 2x2 between-subjects design.
Our study makes three main contributions: First, we empirically demonstrate how the design of an
artifact can shape individual beliefs for sustainable mobility at the example of e-bikes, thus augmenting
the prevalent research focus on action-formation. Second, we present a conversational agent as a new
technological artifact that triggers social responses by users and offers new opportunities for
anthropomorphic design. Third, we show how prescriptive knowledge from persuasive system design and
human-like design can be combined to build an artifact that exhibits a greater degree of persuasiveness.
The remainder of this paper follows the communication schema for experimental research proposed by
Dennis and Valacich (2001): We first describe individual environmental beliefs and their formation,
persuasive system design and anthropomorphic CA design as well as Social Response Theory as the
research background of our study. Then, we develop our hypotheses and present our research design,
comprising the data collection procedure and sample, the experimental conditions as well as the measures
used in the questionnaire. Afterwards, we present our results, discuss theoretical and practical
implications as well as outline limitations and opportunities for future research before concluding with
some final remarks.
Research Background and Related Work
Every individual action, such as choosing the car or bicycle for short distance travel, is influenced by
personal beliefs. In the context of sustainability, these beliefs can encourage environmentally friendly
behavior, in particular when individuals believe that they can take efficient countermeasures to mitigate a
threat to personally valued objects like oceans or rainforests (Henkel and Kranz 2018; Steg et al. 2005).
Individual beliefs are influenced by a variety of societal and organizational factors (Melville 2010).
Societal factors include for example family life or political discourses that shape our beliefs about the state
of the planet and the impact of human behavior on it. Organizational factors comprise aspects, such as
corporate vision statements, sustainability campaigns or information systems that create transparency
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 3
regarding the environmental impact of the organization and promote sustainable behavior. Against this
background, the Theory of Planned Behavior (Ajzen 1985, 1991) helps to understand the relation between
individual environmental beliefs and actions (De Groot and Steg 2007; Melville 2010). According to Ajzen
(1991), three types of salient beliefs influence individual action: First, behavioral beliefs link behavior to
outcome and describe that if an outcome is favorable from the individual’s perspective she or he will have
a positive attitude towards performing the behavior, which increases the likelihood of actually performing
it. For example, if one believes contributing to the reduction of CO2 emissions, even on a very small scale,
is a favorable outcome, she or he will have a generally positive attitude towards riding a bicycle instead of
taking a car. The second type of belief, normative beliefs, considers that humans are social beings and
suggests that the key people around an individual with their behavioral expectations influence individual
behavior. Staying with the example of bicycles as a sustainable mobility mode, behavioral expectations
could stem from colleagues at work that expect one to use a bicycle for short distance travel. Third,
control beliefs, comprise an individual’s belief about factors that may facilitate or hinder performing a
specific behavior. The concept of perceived behavioral control is related to self-efficacy (Bandura 1977,
1996) and includes for example if one feels to make a worthwhile contribution to reducing CO2 emissions
by taking a bicycle for a short ride or whether it contributes to individual health. In the following, we will
use this distinction between behavioral, normative, and control beliefs for a more granular
understanding of the relation between our artifact’s design and individual environmental beliefs.
Persuasive Design for Pro-Environmental Beliefs and Behavior
The idea that technology can influence human beliefs and behavior has been established for around two
decades now and is based on the paradigm of computers as social actors (Fogg and Nass 1997; Nass and
Moon 2000). Research in the area of persuasive design seeks to account for the social response people
show to computers (Reeves and Nass 1996) and uses different design elements to shape user perception
and promote desired behavior. Fogg (2003) distinguishes five types of social cues in persuasive design:
physical (e.g. face, movement), psychological (e.g. humor, empathy), language (e.g. interactive language,
spoken language), social dynamics (cooperation, praise), and social roles (e.g. teammate, guide).
Designers are thus equipped with cues that can be used for persuasive design in a variety of application
domains, such as education, health or environmental sustainability (Langrial et al. 2012).
In the domain of green IS, persuasive design offers the opportunity to promote individual and
organizational environmentally friendly beliefs and behavior (Brauer et al. 2016; Shevchuk and Oinas-
Kukkonen 2016), thus contributing to more sustainable processes and practices (Dedrick 2010; Gholami
et al. 2016; Marett et al. 2012). While design-oriented studies in this domain are limited (Brendel et al.
2017; Henkel and Kranz 2018), different researchers have started to explore persuasive design for green
IS and indicate a positive impact on individual behavior. For example, Loock et al. (2013) design a web
portal to motivate customers to reduce their electricity consumption with the use of goal setting. Their
results show a positive impact of default goal setting on energy conservation as long as the default goal is
not set too low or too high with regard to a self-set goal and identify a moderating effect of feedback on
goal attainment. Similarly, Fitzpatrick and Smith (2009) explore technology-enabled feedback on
domestic energy consumption and present design concerns to build systems that provide effective
consumption feedback to promote behavior change. Extending this fact-orientated feedback on
sustainability behavior, Flüchter and Wortmann (2014) investigate the role of additional normative
feedback on the use of e-bikes for commutes. Flüchter and Wortmann (2014) design an information
system that facilitates a competition for e-bike use and creates transparency regarding the user’s own
ranking in this challenge. The authors argue that while commuting competitions might be useful to
encourage sustainable user behavior, they might also lead to unintended negative effects regarding
individual intrinsic motivation (decreased enjoyment of e-bikes) and perceived autonomy (increased
feeling of being controlled in the competition). Shevchuk and Oinas-Kukkonen (2016) review these and
other applications in published research to identify elements of persuasive design used in such artifacts.
These design elements fall into the categories of primary task support (e.g. self-monitoring or
personalization), dialogue support (e.g. praise, rewards, or reminders), credibility support (e.g.
trustworthiness or expertise), and social support (social facilitation, cooperation, or competition).
Admitting that Green IS is still a rather young domain, Shevchuk and Oinas-Kukkonen (2016) find that
few studies explore design in the context of sustainability and its impact on individual beliefs and actions
despite its relevance and potential (Malhotra et al. 2013).
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 4
Anthropomorphic Conversational Agent Design and Social Response Theory
Studies that present prescriptive knowledge for systems which promote pro-environmental belief and
behavior change primarily focus on web-based systems (e.g. Loock et al. (2013)) or mobile applications
(e.g. Pitt et al. (2011)) with traditional, graphical user interfaces. With substantial technological progress
in the areas of natural language processing and machine learning, so-called conversational agents (CAs)
currently attract strong interest in both theory and practice (Diederich, Brendel, and Kolbe 2019; McTear
et al. 2016; Oracle 2016). Based on the fundamental idea of human-computer interaction via natural
language, CAs are now studied in various application areas, such as customer service (e.g. (Diederich,
Brendel, Lichtenberg, et al. (2019), Wünderlich and Paluch (2017)), marketing and sales (e.g. Hanus and
Fox (2015)), healthcare (e.g. Sebastian and Richards (2017)), or education (e.g. Graesser et al. 2017; Le
and Wartschinski (2018)) and different forms of CAs have emerged. Humans can now interact with CAs
using written (e.g. chatbots) or spoken natural language (e.g. personal assistants like Siri or Microsoft
Cortana) and CAs can be disembodied (Araujo 2018), have a virtual embodiment (Wünderlich and Paluch
2017) or a physical embodiment, such as in the form of service robots (Stock and Merkle 2018). The
human-like characteristics of CAs, both in form, such as a human name, and function, for example
participating in a dialogue with turn-taking, trigger mindless social responses by users (Feine et al. 2019;
Verhagen et al. 2014) as postulated in Social Response Theory (Fogg and Nass 1997; Nass and Moon
2000) where the intensity of such responses varies according to the degree of human-likeness of the CA
(Gong 2008). Emerging studies on CA design find that a higher degree of humanness is associated with
different positive effects, such as on service encounter satisfaction (Gnewuch et al. 2018), perceived social
presence (Pereira et al. 2014), trustworthiness (Araujo 2018) and persuasiveness (Gong 2008).
For the human-like design of CAs, Seeger et al. (2018) present a conceptual framework that comprises
three dimensions (Table 1): The first dimension, human identity, includes anthropomorphic cues
regarding the representation of the agent, for example by means of an avatar (Gong 2008) or
demographic characteristics like a human name (Cowell and Stanney 2005) or gender (Nunamaker et al.
2011). The second dimension, verbal cues, comprises the choice of words and sentences and the way the
CA converses with a user. Exemplary cues in this dimension include for example self-references (“I think
that…” (Sah and Peng 2015)), self-disclosure of artificial thoughts and emotions (“In my experience,
riding the bicycle to work instead of taking the car does make a difference for the environment”
(Schuetzler et al. 2018)) or variability in syntax and word choice (Seeger et al. 2018) instead of repeating
the same sentences in a conversation. The third dimension, non-verbal cues includes non-verbal
communication that conveys information about an individual’s attitudes or emotional state (Ekman and
Friesen 1969). Anthropomorphic cues in this dimensions include dynamic response times depending on
message length and complexity to indicate thinking (Gnewuch et al. 2018) or blinking dots to simulate the
typing of responses by the artificial agent (de Visser et al. 2016) as well as the use of emoticons to express
emotions in a dialogue (Mayer et al. 2006).
Anthropomorphic Design Dimensions
Human identity
Verbal cues
Non-verbal cues
• Avatar (Gong 2008)
• Human name
(Cowell and Stanney
2005)
• Gender
(Nunamaker et al.
2011)
• Self-references (Sah and Peng 2015)
• Self-disclosure
(Schuetzler et al. 2018)
• Personal introduction and greeting (Cafaro
et al. 2016)
• Variability in syntax and word choice
(Seeger et al. 2018)
• Politeness through words
(Mayer et al. 2006)
• Dynamic response times
(Gnewuch et al. 2018)
• Blinking dots
(de Visser et al. 2016)
• Politeness through
emoticons (Mayer et al.
2006)
Table 1. Dimensions for Anthropomorphic CA Design (Seeger et al. 2018) and Cues
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 5
Hypotheses Development
Our research seeks to contribute to a better understanding of the relation between CA design and
individual beliefs regarding sustainable mobility with a focus on two dimensions: persuasive and
anthropomorphic design. Specifically, we hypothesize that persuasive (Shevchuk and Oinas-Kukkonen
2016) and human-like (Seeger et al. 2018) design elements in conversational agents contribute to stronger
behavioral, normative, and control beliefs (Ajzen 1991) of individuals regarding the environmental
sustainability of e-bike use.
Drawing on the computers as social actors paradigm (Fogg and Nass 1997; Nass and Moon 2000), it has
been found that technology influences individual beliefs and behavior (Fogg 2003). Similar to human
persuasion approaches, technology can shape user attitudes and behavior through different approaches
ranging from humor, empathy, and praise to using social dynamics like cooperation and competition.
Oinas-Kukkonen and Harjumaa (2009) organize these persuasive elements in a framework that provides
prescriptive knowledge by means of specific design principles and exemplary instantiations. The
researchers propose four categories that group design elements into primary task support (e.g. tailoring of
information and personalized design), dialogue support (e.g. praise or rewards), credibility support (e.g.
trustworthiness or expertise), and social support (e.g. social facilitation or normative influence) (Oinas-
Kukkonen and Harjumaa 2009; Shevchuk and Oinas-Kukkonen 2016). These design elements have the
potential to reinforce, change or shape beliefs and behavior of individuals by increasing the
persuasiveness of artifacts (Lehto et al. 2012). Thus, we formulate our first hypothesis as follows:
H1: Persuasive design elements have a positive impact on perceived persuasiveness of the agent.
Furthermore, conversational agents as a specific type of IT artifact with which users interact through
natural language offer various possibilities for human-like design. Equipping the agent with a human
identity including a name, gender and avatar (Cowell and Stanney 2005; Gong 2008), verbal cues like
expressions of emotions (Wang et al. 2008) or the use of self-references (Sah and Peng 2015), and non-
verbal cues, such as dynamic response delays for thinking and typing (Gnewuch et al. 2018) can
contribute to users perceiving the agent as more human-like despite its apparent artificial nature. Against
this background, we hypothesize that such cues contribute to users perceiving the agent as more human-
like:
H2: Human-like design cues have a positive impact on perceived humanness of the agent.
Emerging studies that explore the human-like design of conversational agents suggest that perceived
humanness can increase the persuasiveness of the agent. For example, Harjunen et al. (2018) find that
participants in an experiment accepted virtual offers more when the agent showed human-like behavior
by smiling or touching with a haptic glove. Similarly, Adler et al. (2016) show that a CA which showed
(positive) emotions, exhibits a higher degree of perceived persuasiveness than a CA that does not make
emotionally-loaded statements. Thus, we hypothesize:
H3: Perceived humanness has a positive impact on perceived persuasiveness of the agent.
Next, we consider the relation between persuasiveness and individual behavioral beliefs. As persuasive
design elements, such as praise for e-bike use in combination with positively highlighting the individual
environmental contribution (Oinas-Kukkonen and Harjumaa 2009), link behavior (e-bike use) to an
actual, favorable outcome (contribution to our environment), we hypothesize that persuasiveness has a
positive impact on behavioral beliefs in the fourth hypothesis:
H4: Perceived persuasiveness positively impacts individual behavioral beliefs regarding the
environmental impact of e-bike use.
Additionally, we suggest that persuasive design can increase individual normative beliefs by showing the
sustainable behavior of normative referents of the individual user, such as family, friends, and colleagues.
Elements from the persuasive design category of social support (Oinas-Kukkonen and Harjumaa 2009),
such as social comparison, social facilitation or competition/cooperation, can make collective desired
behavior transparent and indicate the underlying, implicit beliefs by highlighting activities of other users
(Flüchter and Wortmann 2014). Thus, we propose that perceived persuasiveness positively affects
individual normative beliefs regarding the use of e-bikes:
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 6
H5: Perceived persuasiveness positively impacts individual normative beliefs regarding the
environmental impact of e-bike use.
Finally, we propose that persuasive design can positively contribute to stronger control beliefs in the
context of environmental sustainability by highlighting factors facilitating and diminishing factors that
inhibit performance of sustainable behavior. For example, a conversational agent with a persuasive design
can provide tailored information (Shevchuk and Oinas-Kukkonen 2016) by highlighting the decreased
commute time for business travelers with a e-bike compared to a car during rush hour and emphasize the
positive impact on individual health. In this example, decreased commute time for business travelers with
little time and contributing to individual health, serve as facilitating factors for commuting with the e-bike
instead of the car. Hence, we hypothesize that persuasive design of a CA by means of tailored information
(Oinas-Kukkonen and Harjumaa 2009) positively impacts individual control beliefs with regard to the
environmental impact of e-bike use.
H6: Perceived persuasiveness positively impacts individual control beliefs regarding the environmental
impact of e-bike use.
Overall, we formulate six hypotheses with regard to the relation between anthropomorphic and persuasive
conversational agent design and individual behavioral, normative and control beliefs concerning the
environmental impact of e-bike use. Figure 1 visualizes the hypotheses in our research model.
Figure 1. Research Model
Research Design
To investigate our hypotheses regarding the relation between persuasive, anthropomorphic CA design and
sustainability beliefs, we conducted an online experiment in the context of organizational e-bike sharing.
The experiment took place in the first quarter of 2019 over a span of three months and is part of a larger
design science research project which aims to design an information system that facilitates and promotes
e-bike sharing within different organizations. In the following, we describe the experimental conditions,
the data collection procedure and sample, and the measures used in our post-experimental survey.
Data Collection Procedure and Sample
The participants of our experiment were asked to interact with a conversational agent to look up and
modify a reservation for an e-bike to ride from one company site to another. Every participant received
exactly the same information for the experiment (Dennis and Valacich 2001) in the form of a briefing
document in which we described the context (e-bike sharing in a company), the structure of the
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 7
experiment (interaction with a virtual reservation agent followed by a questionnaire) as well as
experimental tasks. The document further contained a link that randomly assigned the participant to one
of the experimental conditions. Every participant was asked to complete five tasks, independent from his
or her intention to use e-bikes: Look up an existing e-bike reservation (task 1) and authenticate with the
virtual reservation agent (task 2), find out whether a bike is available at a given company site tomorrow in
the afternoon (task 3), modify the existing reservation for tomorrow afternoon (task 4), and request a
confirmation of the change via e-mail (task 5).
Although the agent was able to process most of the participant’s input regardless of the wording used, we
chose a rather specific set of tasks to allow for a relatively structured dialogue and comparability across
the experimental conditions, similar to other experimental studies on current conversational agents (e.g.
Gnewuch et al. (2018)). After completing the final task, the agent responded with a link to the survey. The
experiment took around eight minutes per person. Participants were acquired using the panel company
Clickworker (2018) and received a small compensation for their participation. After removing six invalid
responses, identified by two attention checks where participants were asked to indicate a specific number,
we arrived at a sample size of n = 225 participants ranging from 18 to 49 years old (mean: 32.2 years) and
a share of 45.3 % female persons.
Experimental Conditions
We chose a 2x2 experimental design comprising the dimensions persuasive design (neutral/persuasive)
and anthropomorphic CA design (machine-like/human-like) as visualized in Table 2. Every participant
was randomly assigned to one experimental condition (between-subjects design), thus avoiding carryover
effects (Boudreau et al. 2001). To conduct the experiment, we designed and implemented a text-based CA
using the natural language processing platform Dialogflow by Google. We integrated the agent using a
custom-built web interface to provide access from different devices. The agent was able to understand and
process variations of sentences with the same meaning and was able to extract, validate and repeat
different parameters, such as reservation numbers or e-mail addresses, to indicate that it understood a
user’s request.
Anthropomorphic Design
Machine-Like
Human-Like
Persuasive Design
Neutral
Condition 1
Agent without anthropomorphic cues
and without persuasive design elements
Condition 2
Agent with anthropomorphic cues
and without persuasive design elements
Persuasive
Condition 3
Agent without anthropomorphic cues
and with persuasive design elements
Condition 4
Agent with anthropomorphic cues
and with persuasive design elements
Table 2. Conditions of the Experimental Design
After implementing the conversation model, we created four different instances corresponding to our
experimental conditions: Condition 1 consisted of an agent with a neutral and machine-like design, thus
the CA did not exhibit additional anthropomorphic cues or persuasive design elements, but provided
support for the user in the experimental task. The agent in condition 2 did not contain persuasive design
elements, but included anthropomorphic cues, such as a name, self-references and dynamic response
times. Condition 3 contained an agent without anthropomorphic cues and persuasive design elements.
Finally, condition 4 comprised an agent with human-like cues as well as the aforementioned persuasive
design elements. Concerning the dimension anthropomorphic CA design, the human-like design in
conditions 2 and 4 contained the anthropomorphic cues shown in Table 1. Regarding the human identity,
we equipped the CAs with a comic-like human avatar, a name (“Sarah”), and female gender. With regard
to verbal cues, it used self-references and self-disclosure, a personal introduction and greeting (“Hello!
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 8
My name is Sarah and I am here to help you reserving an e-bike”), different wording of sentences and
word choice as well as polite statements. Furthermore, the human-like CA was designed with non-verbal
cues by means of dynamic response times and blinking dots to indicate thinking and typing as well as the
use of emoticons to express emotions.
The CAs with persuasive design elements in condition 3 and 4 included self-monitoring for e-bike use,
praise, social cooperation and facilitation, as well as tailoring (Oinas-Kukkonen and Harjumaa 2009).
These elements were distributed among the five experimental tasks as follows: After looking up the
existing e-bike reservation (task 1) and authenticating with the agent (task 2), the persuasive CAs
provided self-monitoring information in combination with praise (“With this reservation, you already
took four e-bike rides this week and traveled more than 6.5 kilometers - without any traffic jam or CO2
emissions. Keep it up!”). After checking the availability of an e-bike, the CAs with persuasive designs
highlight that two colleagues will be travelling at the same time to the selected company site (social
facilitation). Finally, after requesting a confirmation of the modified reservation, the CAs with a
persuasive design thank the user for the contribution to the environment as follows: “Thank you for
contributing to the environmental goals of our company – only 29 rides to go until we reach our goal 300
commutes. With your e-bike use, you already reduced CO2 emissions by around 1,100g compared to
taking the car and helped to reduce traffic congestion in our city.” (praise, tailoring, social cooperation to
reach shared organizational environmental goal).
Figure 2. Agent with Anthropomorphic Cues and Persuasive Design Elements
(Condition 4, Material translated to English)
Measures
After the interaction with the conversational agent, every participant was asked to complete a survey
concerning their beliefs with regard the environmental impact of e-bike use and the perception of the
agent. For measuring behavioral, normative and control beliefs (Ajzen 1985, 1991), we followed the
guideline provided by Ajzen (2006) by first conducting interviews to elicit salient beliefs of participants as
a preparation for the survey. We carried out six qualitative interviews with free-form responses in which
we identified behavioral outcomes of e-bike use (reduced CO2 emissions, less traffic congestion, general
contribution to our environment), normative referents in the organizational context (supervisor,
colleagues), and control factors (positive environmental impact, worthwhile environmental impact). Next,
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 9
we used the results from the interviews to adapt and contextualize the instruments by Ajzen (2006) in our
survey. We extended these by established measures for perceived humanness (Holtgraves and Han 2007;
Holtgraves et al. 2007) and perceived persuasiveness (Lehto et al. 2012). As suggested in the respective
studies, perceived humanness was measured on a 9-point semantic differential scale, while all other items
were measured on a 7-point Likert scale. Two attention checks were added to the survey. Finally, we
collected demographic information (age, gender) and information on the frequency of digital assistant use
(e.g. Siri, Alexa, and chatbots) by the participants ranging from daily, weekly, monthly to never.
Construct
Items
Loadings
Behavioral beliefs
(Ajzen 1985, 1991)
Cronbach’s a = .887
CR = .914
AVE = .693
M = 5.910
SD = 1.004
Using a e-bike for commuting will contribute to reducing CO2 emissions.
Reducing CO2 emission through e-bike use is good.
Using a e-bike for commuting will result in less traffic congestion.
Reducing traffic congestion through e-bike use is good.
Using a e-bike for commuting will contribute to our environment
Contributing to our environment through e-bike use is good.
.831
.842
.785
.713
.803
.815
Normative beliefs
(Ajzen 1985, 1991)
Cronbach’s a = .835
CR = .889
AVE = .668
M = 4.307
SD = 1.262
My supervisor thinks that I should contribute to our environment by taking
the e-bike for commutes.
When it comes to the environmental sustainability of our organization, I
want to do what my supervisor thinks I should do.
My colleagues think that that I should contribute to our environment by
taking the e-bike for commutes.
When it comes to the impact of environmental sustainability of our
organization, I want to do what my colleagues think I should do.
.761
.789
.878
.837
Control beliefs
(Ajzen 1985, 1991)
Cronbach’s a = .871
CR = .912
AVE = .722
M = 5.429
SD = 1.252
I expect my e-bike use to positively influence the environmental impact of
our organization.
Positively influencing the environmental impact of our organization would
enable me to use the e-bike for commuting.
I expect my e-bike use to make a worthwhile contribution to our environment
Making a worthwhile contribution to our environment would enable me to
take the e-bike for commuting.
.856
.838
.852
.852
Humanness
(Holtgraves & Han 2007)
Cronbach’s a = .892
CR = .919
AVE = .656
M = 6.628
SD = 1.607
extremely inhuman-like – extremely human-like
extremely unskilled – extremely skilled
extremely unthoughtful – extremely thoughtful
extremely impolite – extremely polite
extremely unresponsive – extremely responsive
extremely unengaging – extremely engaging
.791
.891
.829
.634
.852
.836
Persuasiveness
(Lehto et al. 2012)
Cronbach’s a = .944
CR = .964, AVE = .899
M = 3.225, SD = 1.654
The agent has an influence on my thinking on environmental sustainability.
The agent is personally relevant for me.
The agent makes me reconsider my thinking about sustainable mobility.
.930
.954
.961
CR = Composite Reliability, AVE = Average Variance Extracted, M = Mean, SD = Standard Deviation
Table 3. Constructs, Items, and Factor Loadings
Table 3 summarizes the constructs, items, and factor loadings including Cronbach’s a, composite
reliability (CR) and average variance extracted (AVE). We included all items in the analysis as the factor
loadings were larger than .60 as suggested by Gefen and Straub (2005). All constructs showed sufficient
values for Cronbach’s a (larger than .80), composite reliability (larger than .80) and average variance
extracted (larger than .50) with respect to the levels proposed by Urbach and Ahlemann (2010).
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 10
Results
Partial least squares (PLS) was used to test our hypotheses regarding the relation between persuasive and
anthropomorphic design and individual sustainability beliefs due to our focus on prediction rather than
model fit and the use of latent variables with formative measures (Ringle et al. 2012). All analyses were
carried out using SmartPLS 3. As suggested by Chin (1998), the significance of path coefficients was
calculated with a bootstrapping resampling method with 5,000 samples. Figure 3 visualizes the path
coefficients, R2 values, and significance levels.
Figure 3. PLS Structural Model (n = 225)
The paths between the CA’s design dimensions (persuasive design, human-like, design) and user
perceptions regarding persuasiveness and humanness of the agent show significant relationships. Our
data indicates that persuasive design positively impacts perceived persuasiveness for environmental belief
change (Persuasive Design → Perceived Persuasiveness, β = 0.16, p = .007), thus providing support for
H1. Human-like design through anthropomorphic cues has a positive impact on perceived humanness
(Human-Like Design → Perceived Humanness, β = 0.15, p = .013), indicating support for H2. In addition,
our results show support for H3, stating that perceived humanness positively contributes to perceived
persuasiveness (Perceived Anthropomorphism → Perceived Persuasiveness, β = 0.30, p < .001). Finally,
two of the three paths between perceived persuasiveness and the three individual sustainability beliefs
indicate significant relationships. The data from our post-experimental survey shows no significant
relationship between perceived persuasiveness and individual behavioral beliefs (Perceived
Persuasiveness → Behavioral Beliefs, β = -0.03, p = .327), thus we do not find support for H4. With
regard to the relation between perceived persuasiveness and individual normative beliefs, we observe a
significant relationship (Perceived Persuasiveness → Normative Beliefs, β = 0.49, p < .001), indicating
support for H5. Finally, our survey data shows a significant path between perceived persuasiveness and
individual control beliefs (Perceived Persuasiveness → Control Beliefs, β = 0.19, p = .002). Consequently,
we find support for our sixth hypothesis, which states a positive influence of perceived persuasiveness on
control beliefs. Table 4 summarizes the hypothesized relationships and their significance.
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 11
Hypothesis
Relationship
β-value
t-value
p-value
Support
H1
Persuasive Design → Perceived Persuasiveness
0.16
2.49
.007
Yes
H2
Human-Like Design → Perceived Humanness
0.15
2.23
.013
Yes
H3
Perceived Humanness → Perceived Persuasiveness
0.30
5.16
.000
Yes
H4
Perceived Persuasiveness → Behavioral Beliefs
-0.03
0.45
.327
No
H5
Perceived Persuasiveness → Normative Beliefs
0.49
9.13
.000
Yes
H6
Perceived Persuasiveness → Control Beliefs
0.19
2.83
.002
Yes
Table 4. Results of Hypothesis Tests
We also analyzed the effect of the control variables (age, gender, frequency of digital assistant use) on the
latent variables and conducted an exploratory post-hoc analysis for moderator effects. Significant paths
include age to control beliefs (β = 0.10, p = .038), age to perceived persuasiveness (β = -0.21, p = .001),
digital assistant use on behavioral (β = 0.13, p = .014), normative (β = 0.12, p = .025), and control beliefs
(β = 0.14, p = .013) as well as on perceived persuasiveness (β = 0.17, p = .004), and gender to control
beliefs (β = 0.11, p = .041). In addition, two significant moderator effects were found in our dataset: First,
we found that the age of the participants moderated the relationship between human-like design and
perceived humanness (β = -0.11, p = .032), indicating that users with higher age did perceive an agent
with a human-like design as less anthropomorphic. Similarly, the frequency of digital assistant use
moderates the relationship between human-like design and perceived humanness (β = -0.15, p = .009).
Discussion
Our experiment aimed to explore the relation between persuasive and anthropomorphic design of
conversational agents and sustainability beliefs. The results have implications for promoting
environmental beliefs of individuals through the design of artifacts in general and conversational agents
in particular and provide empirical evidence for the potential of anthropomorphism in the context of
persuasive design.
Shaping Environmental Beliefs through Design
We find empirical support for our hypotheses that perceived persuasiveness has a positive impact on
individual normative and control beliefs regarding environmental sustainability of e-bike use and sharing,
however, our data does not indicate support for the relation between perceived persuasiveness and
behavioral beliefs. Concerning normative beliefs, showing the behavior and the implicit beliefs guiding
such action of normative referents, in this case colleagues and supervisors, seems to be efficient in
shaping individual normative beliefs of users. By means of social facilitation (“two of your colleagues will
be using the e-bike tomorrow at the same time”) and cooperation towards achieving a shared
organizational goal for e-bike use, the agent has the potential to facilitate social influence as also shown in
similar studies on environmental sustainability and design (e.g. Baeriswyl et al. 2011; Koo and Chung
2014; Römer et al. 2015). While we were able to show fictitious, positive behavior of the referents,
designers need to consider potentially unintended effects of such social support elements in persuasive
design, for example normative referents not showing the intended behavior and thus potentially
decreasing a user’s motivation, the influence of group characteristics (Yim 2011) or a decreased feeling of
autonomy for individual employees as e-bike use is made transparent across the organization (Flüchter
and Wortmann 2014).
With regard to control beliefs, the persuasive design including tailored information, praise for e-bike use,
and highlighting personal benefits positively influenced individual control beliefs. The provision of
personalized and tailored information in combination with praise (“you already took four e-bike rides this
week and traveled more than 6.5 kilometers” or “with this e-bike use, you contributed to reducing CO2
emissions by 1.100g – keep it up!”) indicated the participant that she or he made a positive and
worthwhile contribution, which is often challenging in this context as CO2 emissions can be a “mythical”
number without a tangible connection to individual lives (Corbett 2013b). In addition, highlighting the
individual contribution towards reaching a shared goal for e-bike use can increase the user’s feeling of
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 12
making a worthwhile contribution (Loock et al. 2013) and emphasizing personal benefits (e.g. no need for
parking space or waiting in traffic) (Marett et al. 2012) have the potential to further shape individual
control beliefs. Thus, designers have different persuasive design elements at their disposal to positively
influence perceived behavioral control as a key driver of sustainable behavior (De Groot and Steg 2007).
While our study underlines the potential of persuasive design elements to promote what we consider
desirable environmentally sustainable mobility beliefs and behavior, we emphasize the need for creating
transparency regarding the persuasion goals of such artifacts. From an ethical point of view, revealing the
designer’s intention to the artifact’s users is essential and persuasion should always be open and be based
on truthful information (Oinas-Kukkonen and Harjumaa 2009; Shevchuk and Oinas-Kukkonen 2016).
Thus, we consider it the designer’s responsibility to take into account intended ethical outcomes as well as
reasonably predictable unintended outcomes of persuasive technology in line with the arguments
provided by Berdichevsky and Neuenschwander (2002).
Human-Like Conversational Agents in the Context of Persuasive Design
We further find a positive impact of perceived humanness of the conversational agent on perceived
persuasiveness as formulated in our third hypothesis, indicating that the human-like design and human
perception of such agents does indeed contribute to persuasiveness. Similar studies on anthropomorphic
conversational agent design, drawing on the Social Response Theory, find a positive effect of perceived
humanness on further aspects, such as perceived competency, social influence, trustworthiness (Araujo
2018; Gong 2008), and service encounter satisfaction (Gnewuch et al. 2018). As for example perceived
trustworthiness of an artifact directly affects its potential for persuasion (Lehto et al. 2012), the human-
like design of such agents can enable reaching a higher degree of persuasiveness. Furthermore, praise
coming from a human-like CA, to which users show substantial social responses, might mean more to the
individual as praise expressed by a machine-like agent (Jucks et al. 2018).
The explorative post-hoc analysis for moderator effects revealed that a user’s age (β = -0.11, p = .032), and
frequency of digital assistant and chatbot use (β = -0.15, p = .009) moderate the relation between human-
like design and perceived humanness. Thus, the effect of a human-like design of a CA on perceived
humanness might be smaller for older users and users that frequently use comparable assistants.
Furthermore, different studies on anthropomorphic system design highlight the need to consider
unintended effects of increasing anthropomorphism of the agent, such as that human-likeness does not
linearly increase with trust, but that users might feel threatened at one point (Seeger et al. 2018).
Similarly, Wünderlich and Paluch (2017) describe that users can be irritated by human-like design as they
could feel unsure whether the interact with a machine or an actual human.
Limitations and Opportunities for Future Research
Our study is not free of limitations and offers different opportunities for future research. We conducted
the online experiment in a rather controlled setting with a set of specific tasks that every participant was
asked to complete and a single interaction with the conversational agent. Thus, we benefitted from control
yet lacked realism in our research design (Dennis and Valacich 2001). Future research could investigate
how CA design influences individual behavioral, normative, and control beliefs with regard to sustainable
mobility in the field where users interact with the CA in a variety of tasks.
In addition, observing individual environmental belief change, in particular with regard to behavioral
beliefs, over a longer period of time and over multiple interactions with an artifact can be beneficial as
successful persuasion takes place incrementally (Oinas-Kukkonen and Harjumaa 2009). A long-term
observation can further be useful to analyze whether unintended effects on individual beliefs might occur.
For example, Flüchter and Wortmann (2014) found a partially negative perception of social normative
feedback in the form of a ranking in an organizational e-bike competition by users after a longer period of
time. Furthermore, our work focused on individual sustainability beliefs in-depth but did not research
subsequent actions and outcomes (Melville 2010). We believe that investigating the interplay between
human-like CA design, individual sustainability beliefs, and their impact on actions and outcomes
represents a worthwhile research opportunity as pro-environmental behavior is guided by values and
norms (Lindenberg and Steg 2013; Lo et al. 2012). IS researchers can for example draw on the Theory of
Planned Behavior (Ajzen 1985, 1991) to gain a better understanding of the relation between persuasive
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 13
and anthropomorphic design, environmental behavioral, normative, and control beliefs, as well as
individual intentions to use and actual mobility behavior. From our point of view, a better understanding
of this relation can be useful to inform the design of artifacts that successfully promote environmentally
friendly individual behavior. In addition, our research design did not account for socially desirable
responding, which represents a further research opportunity in the context of anthropomorphic CA design
and might help to better understand if the elicited beliefs are substantial, persist over time and translate
in actual sustainable mobility behavior.
Furthermore, while our research makes a first step towards a better understanding between Information
System design and environmental belief-formation at the example of sustainable mobility, we did not
cover the full range of persuasive design elements (Oinas-Kukkonen and Harjumaa 2009) and our
research design did not allow for an analysis of the impact of single persuasive design elements on the
three types of individual beliefs (Ajzen 1985, 1991). Thus, future studies can explore the impact of single
persuasive design elements on belief types to gain a more precise understanding of the interplay between
design and individual (sustainability) beliefs. Additionally, and in line with Shevchuk and Oinas-
Kukkonen (2016), we suggest to explore the full range of persuasive design elements, such as tailored and
personalized presentation of environmental information, rewards and reminders, and social roles in the
context of individual belief and behavior change as a promising opportunity for future studies.
Furthermore, our study has shown that combining persuasive and anthropomorphic design can increase
the persuasiveness of CAs. However, our data also indicated that the human-like design of the agent used
in this experiment only explains a small part of the variance in perceived humanness (R2 = 0.063), thus
we propose future research to investigate which factors, related to the CA design or other aspects,
influence how human-like a CA is perceived by its users. In addition, while we focused on promoting
sustainable mobility behavior, we suggest to adapt the approach used in this study to use cases in domains
outside of Green IS, such as health or education. Finally, future research in the context of persuasive,
anthropomorphic CAs should explore and critically discuss the ethical questions that arise when
designing increasingly human-like, persuasive conversational agents.
Concluding Remarks
Positively influencing individual environmental beliefs and behavior through the design of technical
artifacts is a key concern in the area of green IS and of high practical relevance as it can contribute to
more sustainable attitudes and practices (vom Brocke et al. 2013; Melville 2010). In this study, we set out
to explore the relation of persuasive and human-like design and behavioral, normative, and control
beliefs. By means of an 2x2 online experiment with a conversational agent and 225 participants, we find
empirical evidence for the positive impact of persuasive and human-like design on individual
environmental control and normative beliefs at the example of organizational e-bike sharing. In addition,
our results underline the potential of anthropomorphic design to facilitate a stronger persuasiveness in
the context of artifacts that seek to promote sustainable individual beliefs and behavior.
Against this background, our study makes three contributions: First, we empirically show how an
artifact’s persuasive design can shape individual normative and control beliefs for sustainable mobility.
Second, we extend the focus of existing green IS designs on mobile apps and websites with a text-based
conversational agent as a new type of artifact that gains increasing interest in research and practice
(McTear 2017) and triggers substantial social responses (Seeger et al. 2018). Third, our results indicate
how the human-like design of such agents can contribute to the persuasiveness of the artifact, thus
increasing the potential to promote sustainable beliefs and behavior through IS design. We encourage
future design-oriented studies in this area to further explore the interplay between an agent’s design,
environmental beliefs and behavior over a longer period of time and in the field as well as to specifically
investigate the impact of single persuasive design elements on individual behavioral, normative, and
control beliefs in the context of environmental sustainability. Overall, a thorough understanding of
effective and efficient designs to positively influence beliefs and behavior on the individual level can allow
us to build artifacts that address prevalent challenges in our society and thus strengthen the positive
environmental impact of our discipline.
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 14
Acknowledgments
This research was supported by the German Federal Ministry for the Environment, Nature Conservation
and Nuclear Safety by resolution of the German Bundestag.
References
Adler, R. F., Iacobelli, F., and Gutstein, Y. 2016. “Are You Convinced? A Wizard of Oz Study to Test
Emotional vs. Rational Persuasion Strategies in Dialogues,” Computers in Human Behavior (57), pp.
75–81.
Ajzen, I. 1985. “From Intentions to Actions: A Theory of Planned Behavior,” in Action Control, J. Kuhl
and J. Beckmann (eds.), New York, USA: Springer Verlag, pp. 11–39.
Ajzen, I. 1991. “The Theory of Planned Behavior,” Organizational Behavior and Human Decision
Processes (50), pp. 179–211.
Ajzen, I. 2006. “Constructing A Theory of Planned Behavior Questionnaire,” Working Paper, Amherst,
Massachusetts.
Araujo, T. 2018. “Living up to the Chatbot Hype: The Influence of Anthropomorphic Design Cues and
Communicative Agency Framing on Conversational Agent and Company Perceptions,” Computers in
Human Behavior (85), pp. 183–189.
Baeriswyl, M., Staake, T., and Loock, C. 2011. “The Effects of User Identity and Sanctions in Online
Communities on Real-World Behavior,” in Proceedings of the International Conference on
Information Systems (ICIS), Shanghai, China, pp. 1–11.
Baier, D., Rese, A., and Röglinger, M. 2018. “Conversational User Interfaces for Online Shops? A
Categorization of Use Cases,” in Proceedings of the International Conference on Information
Systems (ICIS), San Francisco, USA, pp. 1–17.
Bandura, A. 1977. “Self-Efficacy: Toward a Unifying Theory of Behavioral Change,” Pyschological Review
(84:2), pp. 191–215.
Bandura, A. 1996. Self-Efficacy: The Exercise of Control, New York, USA: Freeman.
Berdichevsky, D., and Neuenschwander, E. 2002. “Toward an Ethics of Persuasive Technology,”
Communications of the ACM (42:5), pp. 51–58.
Boudreau, M.-C., Gefen, D., and Straub, D. 2001. “Validation in Information Systems Research: A State-
of-the-Art Assessment,” Management Information Systems Quarterly (25:1), pp. 1–16.
Brauer, B., Ebermann, C., and Kolbe, L. M. 2016. “An Acceptance Model for User-Centric Persuasive
Environmental Sustainable IS,” in Proceedings of the International Conference on Information
Systems (ICIS), Dublin, Ireland, pp. 1–22.
Brendel, A. B., Zapadka, P., and Kolbe, L. M. 2017. “Design Science Research in Green IS: Analyzing the
Past to Guide Future Research,” in Proceedings of the European Conference on Information Systems
(ECIS), Guimarães, Portugal, pp. 1–18.
vom Brocke, J., Watson, R. T., Dwyer, C., Elliot, S., and Melville, N. 2013. “Green Information Systems:
Directives for the IS Discipline,” Communications of the Association for Information Systems (33),
pp. 509–520.
Butler, T. 2011. “Compliance with Institutional Imperatives on Environmental Sustainability: Building
Theory on the Role of Green IS,” Journal of Strategic Information Systems (20:1), pp. 6–26.
Cafaro, A., Vilhjalmsson, H. H., and Bickmore, T. 2016. “First Impressions in Human-Agent Virtual
Encounters,” ACM Transactions on Computer-Human Interaction (24:4), pp. 1–40.
Chin, W. W. 1998. “The Partial Least Squares Approach for Structural Equation Modeling,” in Modern
Methods for Business Research, L. Erlbaum (ed.), Hillsdale, New Jersey, pp. 295–336.
Clickworker. 2018. “Clickworker.” (https://www.clickworker.com/, accessed November 26, 2018).
Corbett, J. 2013a. “Using Information Systems to Improve Energy Efficiency: Do Smart Meters Make a
Difference?,” Information Systems Frontiers (15:5), pp. 747–760.
Corbett, J. 2013b. “Designing and Using Carbon Management Systems to Promote Ecologically
Responsible Behaviors,” Journal of the Association for Information Systems (14:7), pp. 339–378.
Cowell, A. J., and Stanney, K. M. 2005. “Manipulation of Non-Verbal Interaction Style and Demographic
Embodiment to Increase Anthropomorphic Computer Character Credibility,” International Journal
of Human Computer Studies (62:2), pp. 281–306.
Dedrick, J. 2010. “Green IS: Concepts and Issues for Information Systems Research,” Communications of
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 15
the Association for Information Systems (27:1), pp. 173–184.
Dennis, A. R., and Valacich, J. S. 2001. “Conducting Experimental Research in Information Systems,”
Communications of the Association for Information Systems (7:5), pp. 1–41.
Diederich, S., Brendel, A. B., and Kolbe, L. M. 2019. “On Conversational Agents in Information Systems
Research: Analyzing the Past to Guide Future Work,” in Proceedings of the International Conference
on Wirtschaftsinformatik, pp. 1550–1564.
Diederich, S., Brendel, A. B., Lichtenberg, S., and Kolbe, L. M. 2019. “Design for Fast Request Fulfillment
or Natural Interaction? Insights from an Online Experiment with a Conversational Agent,” in
Proceedings of the European Conference on Information Systems (ECIS), Stockholm, Sweden, pp. 1–
17.
Donald, I. J., Cooper, S. R., and Conchie, S. M. 2014. “An Extended Theory of Planned Behaviour Model
of the Psychological Factors Affecting Commuters’ Transport Mode Use,” Journal of Environmental
Psychology (40), pp. 39–48.
Ekman, P., and Friesen, W. V. 1969. “The Repertoire of Nonverbal Behavior: Categories, Origins, Usage,
and Coding,” Semiotica (1:1), pp. 49–98.
Elliot, S. 2011. “Transdisciplinary Perspectives on Environmental Sustainability: A Resource Base and
Framework for It-Enabled Business Transformation,” Management Information Systems Quarterly
(35:1), pp. 197–236.
Feine, J., Gnewuch, U., Morana, S., and Mädche, A. 2019. “A Taxonomy of Social Cues for Conversational
Agents,” International Journal of Human-Computer Studies.
Fitzpatrick, G., and Smith, G. 2009. “Technology-Enabled Feedback on Domestic Energy Consumption:
Articulating a Set of Design Concerns,” Pervasive Computing (8:1), pp. 37–44.
Flüchter, K., and Wortmann, F. 2014. “Promoting Sustainable Travel Behavior through IS-Enabled
Feedback – Short-Term Success at the Cost of Long-Term Motivation?,” in Proceedings of the
International Conference on Information Systems (ICIS), Auckland, New Zealand, pp. 1–17.
Fogg, B. J. 2003. “Computers as Persuasive Social Actors,” in Persuasive Technology: Using Computers
to Change What We Think and Do, San Francisco, USA: Morgan Kaufmann Publishers.
Fogg, B. J., and Nass, C. 1997. “How Users Reciprocate to Computers,” in Proceedings of the ACM CHI
Conference on Human Factors in Computing Systems, Atlanta, USA, p. 331.
Froehlich, J., Dillahunt, T., Klasnja, P., Mankoff, J., Consolvo, S., Harrison, B., and Landay, J. A. 2009.
“UbiGreen: Investigating a Mobile Tool for Tracking and Supporting Green Transportation Habits,” in
Proceedings of the Conference on Human Factors in Computing Systems, Boston, USA, pp. 1043–
1052.
Gefen, D., and Straub, D. 2005. “A Practical Guide to Factorial Validity Using PLS-Graph: Tutorial and
Annotated Example,” Communications of the Association for Information Systems, pp. 91–109.
Gholami, R., Watson, R., Hasan, H., Molla, A., and Bjorn-Andersen, N. 2016. “Information Systems
Solutions for Environmental Sustainability: How Can We Do More?,” Journal of the Association for
Information Systems (17:8), pp. 521–536.
Gnewuch, U., Morana, S., Adam, M. T. P., and Maedche, A. 2018. “Faster Is Not Always Better:
Understanding the Effect of Dynamic Response Delays in Human-Chatbot Interaction,” in
Proceedings of the European Conference on Information Systems (ECIS), Portsmouth, United
Kingdom, pp. 1–17.
Gong, L. 2008. “How Social Is Social Responses to Computers? The Function of the Degree of
Anthropomorphism in Computer Representations,” Computers in Human Behavior (24:4), pp. 1494–
1509.
Graesser, A. C., Cai, Z., Morgan, B., and Wang, L. 2017. “Assessment with Computer Agents That Engage
in Conversational Dialogues and Trialogues with Learners,” Computers in Human Behavior (76), pp.
607–616.
De Groot, J., and Steg, L. 2007. “General Beliefs and the Theory of Planned Behavior: The Role of
Environmental Concerns in the TPB,” Journal of Applied Social Psychology (37:8), pp. 1817–1836.
Hanus, M. D., and Fox, J. 2015. “Persuasive Avatars: The Effects of Customizing a Virtual Salespersons
Appearance on Brand Liking and Purchase Intentions,” International Journal of Human Computer
Studies (84), pp. 33–40.
Harjunen, V. J., Spapé, M., Ahmed, I., Jacucci, G., and Ravaja, N. 2018. “Persuaded by the Machine: The
Effect of Virtual Nonverbal Cues and Individual Differences on Compliance in Economic Bargaining,”
Computers in Human Behavior (87), pp. 384–394.
Henkel, C., and Kranz, J. 2018. “Pro-Environmental Behavior and Green Information Systems Research –
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 16
Review, Synthesis and Directions for Future Research,” in Proceedings of International Conference
on Information Systems (ICIS), San Francisco, USA, pp. 1–17.
Holtgraves, T., and Han, T. L. 2007. “A Procedure for Studying Online Conversational Processing Using a
Chat Bot,” Behavior Research Methods (39:1), pp. 156–163.
Holtgraves, T. M., Ross, S. J., Weywadt, C. R., and Han, T. L. 2007. “Perceiving Artificial Social Agents,”
Computers in Human Behavior (23:5), pp. 2163–2174.
Jucks, R., Linnemann, G. A., and Brummernhenrich, B. 2018. “Student Evaluations of a (Rude) Spoken
Dialogue System Insights from an Experimental Study,” Advances in Human-Computer Interaction
(2018), pp. 1–10.
Koo, C., and Chung, N. 2014. “Examining the Eco-Technological Knowledge of Smart Green IT Adoption
Behavior: A Self-Determination Perspective,” in Technological Forecasting and Social Change (Vol.
88), Ho Chi Minh City, Vietnam, pp. 140–155.
Langrial, S., Lehto, T., and Oinas-Kukkonen, H. 2012. “Native Mobile Applications For Personal
Wellbeing: A Persuasive Systems Design Evaluation,” in Proceedings of the Pacific Asia Conference
on Information Systems (PACIS), Ho Chi Minh City, Vietnam, pp. 1–16.
Le, N. T., and Wartschinski, L. 2018. “A Cognitive Assistant for Improving Human Reasoning Skills,”
International Journal of Human Computer Studies (117), pp. 45–54.
Lehto, T., Oinas-Kukkonen, H., and Drozd, F. 2012. “Factors Affecting Perceived Persuasiveness of a
Behavior Change Support System,” in Proceedings of the International Conference on Information
Systems (ICIS), Orlando, Florida, pp. 1–15.
Lindenberg, S., and Steg, L. 2013. “Goal-Framing Theory and Norm-Guided Environmental Behavior,” in
Encouraging Sustainable Behavior: Psychology and the Environment, H. van Trijp (ed.), New York,
USA: Psychology Press, pp. 37–54.
Lo, S. H., Peters, G. J. Y., and Kok, G. 2012. “A Review of Determinants of and Interventions for
Proenvironmental Behaviors in Organizations,” Journal of Applied Social Psychology (42:12), pp.
2933–2967.
Loock, C.-M., Staake, T., and Thiesse, F. 2013. “Motivating Energy-Efficient Behavior with Green IS: An
Investigation of Goal Setting and the Role of Defaults,” Management Information Systems Quarterly
(37:4), pp. 1313–1332.
Malhotra, A., Melville, N. P., and Watson, R. T. 2013. “Spurring Impactful Research on Information
Systems for Envionmental Sustainability,” Management Information Systems Quarterly (37:4), pp.
1265–1274.
Marett, K., Otondo, R. F., and Taylor, S. G. 2012. “Assessing the Effects of Benefits and Institutional
Influences on the Continued Use of Environmentally Munificent Bypass Systems in Long-Haul
Trucking,” Management Information Systems Quarterly (I:4), pp. 1–9.
Mayer, R. E., Johnson, W. L., Shaw, E., and Sandhu, S. 2006. “Constructing Computer-Based Tutors That
Are Socially Sensitive: Politeness in Educational Software,” International Journal of Human
Computer Studies (64:1), pp. 36–42.
McTear, M. 2017. “The Rise of the Conversational Interface: A New Kid on the Block?,” Lecture Notes in
Computer Science.
McTear, M., Callejas, Z., and Griol, D. 2016. The Conversational Interface: Talking to Smart Devices,
Basel, Switzerland: Springer Publishing Company.
Melville, N. P. 2010. “Information Systems Innovation for Environmental Sustainability,” Management
Information Systems Quarterly (34:1), pp. 1–21.
Nass, C., and Moon, Y. 2000. “Machines and Mindlessness: Social Responses to Computers,” Journal of
Social Issues (56:1), pp. 81–103.
Nunamaker, J. F., Derrick, D. C., Elkins, A. C., Burgoon, J. K., and Patton, M. W. 2011. “Embodied
Conversational Agent-Based Kiosk for Automated Interviewing,” Journal of Management
Information Systems (28:1), pp. 17–48.
Oinas-Kukkonen, H., and Harjumaa, M. 2009. “Persuasive Systems Design: Key Issues, Process Model,
and System Features,” Communications of the Association for Information Systems (24:1), p. 96.
Oracle. 2016. Can Virtual Experiences Replace Reality? The Future Role for Humans in Delivering
Customer Experience, p. 19.
Pereira, A. T., Prada, R., and Paiva, A. 2014. “Improving Social Presence in Human-Agent Interaction,” in
Proceedings of the ACM CHI Conference on Human Factors in Computing Systems, Toronto,
Canada, pp. 1449–1458.
Pitt, L. F., Parent, M., Junglas, I., Chan, A., and Spyropoulou, S. 2011. “Integrating the Smartphone into a
Promoting Sustainable Mobility Beliefs through Design
Fortieth International Conference on Information Systems, Munich 2019 17
Sound Environmental Information Systems Strategy: Principles, Practices and a Research Agenda,”
Journal of Strategic Information Systems (20:1), pp. 27–37.
Reeves, B., and Nass, C. 1996. The Media Equation: How People Treat Computers, Television and New
Media Like Real People and Places, The Center for the Study of Language and Information
Publications.
Ringle, C. M., Sarstedt, M., and Straub, D. W. 2012. “A Critical Look at the USE of PLS-SEM in MIS
Quarterly,” Management Information Systems Quarterly (36:1), pp. 1–8.
Rodrigue, J. P., Comtois, C., and Slack, B. 2013. The Geography of Transport Systems, (4th editio.),
London, United Kingdom: Taylor & Francis.
Römer, B., Reichhart, P., and Picot, A. 2015. “Smart Energy for Robinson Crusoe: An Empirical Analysis
of the Adoption of IS-Enhanced Electricity Storage Systems,” Electronic Markets (25:1), pp. 47–60.
Sah, Y. J., and Peng, W. 2015. “Effects of Visual and Linguistic Anthropomorphic Cues on Social
Perception, Self-Awareness, and Information Disclosure in a Health Website,” Computers in Human
Behavior (45), pp. 392–401.
Schuetzler, R. M., Giboney, J. S., Grimes, G. M., and Nunamaker, J. F. 2018. “The Influence of
Conversational Agents on Socially Desirable Responding,” in Proceedings of the Hawaii
International Conference on System Sciences (HICSS) (Vol. 9), Waikoloa Village, Hawaii, USA, pp.
283–292.
Sebastian, J., and Richards, D. 2017. “Changing Stigmatizing Attitudes to Mental Health via Education
and Contact with Embodied Conversational Agents,” Computers in Human Behavior (73), pp. 479–
488.
Seeger, A.-M., Pfeiffer, J., and Heinzl, A. 2018. “Designing Anthropomorphic Conversational Agents:
Development and Empirical Evaluation of a Design Framework,” in Proceedings of the International
Conference on Information Systems (ICIS), San Francisco, USA, pp. 1–17.
Seidel, S., Chandra Kruse, L., Székely, N., Gau, M., and Stieger, D. 2017. “Design Principles for
Sensemaking Support Systems in Environmental Sustainability Transformations,” European Journal
of Information Systems (27:2), pp. 221–247.
Shevchuk, N., and Oinas-Kukkonen, H. 2016. “Exploring Green Information Systems and Technologies as
Persuasive Systems: A Systematic Review of Applications in Published Research,” in Proceedings of
the International Conference on Information Systems (ICIS), Dublin, Ireland, pp. 1–11.
Steg, L., Dreijerink, L., and Abrahamse, W. 2005. “Factors Influencing the Acceptability of Energy
Policies: A Test of VBN Theory,” Journal of Environmental Psychology (25:4), pp. 415–425.
Stock, R. M., and Merkle, M. 2018. “Customer Responses to Robotic Innovative Behavior Cues During the
Service Encounter,” in Proceedings of the International Conference on Information Systems (ICIS),
San Francisco, USA, pp. 1–17.
Urbach, N., and Ahlemann, F. 2010. “Structural Equation Modeling in Information Systems Research
Using Partial Least Squares,” Journal of Information Technology Theory and Application (JITTA)
(11:2), pp. 5–40.
Vergragt, P. J., and Brown, H. S. 2007. “Sustainable Mobility: From Technological Innovation to Societal
Learning,” Journal of Cleaner Production (15:11–12), pp. 1104–1115.
Verhagen, T., van Nes, J., Feldberg, F., and van Dolen, W. 2014. “Virtual Customer Service Agents: Using
Social Presence and Personalization to Shape Online Service Encounters,” Journal of Computer-
Mediated Communication (19:3), pp. 529–545.
de Visser, E. J., Monfort, S. S., McKendrick, R., Smith, M. A. B., McKnight, P. E., Krueger, F., and
Parasuraman, R. 2016. “Almost Human: Anthropomorphism Increases Trust Resilience in Cognitive
Agents,” Journal of Experimental Psychology: Applied (22:3), pp. 331–349.
Wang, N., Johnson, W. L., Mayer, R. E., Rizzo, P., Shaw, E., and Collins, H. 2008. “The Politeness Effect:
Pedagogical Agents and Learning Outcomes,” International Journal of Human Computer Studies
(66:2), pp. 98–112.
Willing, C., Brandt, T., and Neumann, D. 2017. “Intermodal Mobility,” Business & Information Systems
Engineering (59:3), pp. 173–179.
Wünderlich, N. V., and Paluch, S. 2017. “A Nice and Friendly Chat With a Bot: User Perceptions of AI-
Based Service Agents,” in Proceedings of the International Conference on Information Systems
(ICIS), Seoul, Korea, pp. 1–11.
Yim, D. 2011. “Tale of Two Green Communities: Energy Informatics and Social Competition on Energy
Conservation Behavior,” in Proceedings of Americas Conference on Information Systems (AMCIS),
Detroit, USA, pp. 1–13.
