Impact of communication sources for achieving campus sustainability

Resources, Conservation & Recycling 139 (2018) 366–376

Contents lists available at ScienceDirect

Resources, Conservation & Recycling journal homepage: www.elsevier.com/locate/resconrec

Full length article

Impact of communication sources for achieving campus sustainability a,⁎

b

a

Amy A. Kim , Hessam Sadatsafavi , Lysandra Medal , Marilyn J. Ostergren a b c

T

c

Department of Civil and Environmental Engineering, University of Washington, Seattle, WA 98195, United States Cornell Institute for Healthy Futures, Department of Design and Environmental Analysis, Cornell University, Ithaca, NY 14853, United States Renewable Energy Liaison, UW Environmental Stewardship Committee, University of Washington, Seattle, WA 98195, United States

A R T I C LE I N FO

A B S T R A C T

Keywords: Campus sustainability Sustainability communication Communication sources Occupant awareness Behavioral changes Effective communication

Using self-reported data from over 12,000 students, staff, and faculty members, this study aims to quantify the effectiveness of various sustainability information sources on the awareness and behavioral changes of participants. Exploratory factor analysis of 12 communication sources resulted in three clusters: news sources, blogs and social media sources, and local sources of information. All three clusters had positive impacts on the sustainability awareness of students, staff, and faculty members. Of the three clusters, local sources were the least effective strategy for enhancing sustainability awareness in all groups. Overall, for undergraduate students, blogs and social media sources were most effective, and for graduate or professional students and staff, news sources were most effective. While sustainability awareness increased from 2012 to 2014, the total effects on actions were small for all groups. The lack of impact on behavioral changes can be explained by the high initial level of action taken by participants, regardless of their low level of awareness about campus-specific sustainability initiatives. Recommendations for further behavioral changes include combining multiple communication sources and channels, providing engagement opportunities, increasing access to information, and using participatory methods to encourage further peer-to-peer support networks.

1. Introduction Achieving campus sustainability is a multifaceted and challenging endeavor. Many aspects of campus operation have sustainability implications, including building construction, building operation and maintenance, grounds maintenance, food service, purchasing, and travel for business and commuting (Wright, 2002). The campus environment is unique in that it can influence the behaviors of people through education, demonstration, and research (Krasny and Delia, 2015; Filho and Brandli, 2016). This complex environment provides an ideal context for continuous environmental learning and practice (König, 2013; Krasny and Delia, 2015; Norazah and Norbayah, 2016). Despite the important role that higher education has in the development of sustainability practices, studies find that universities still face challenges in transforming sustainability awareness into sustained action. Owens and Halfacre-Hitchcock (2006) showed that engagement with sustainability-related projects by faculty and students did not yield any behavioral changes in the individual-level assessment. Roorda (2004) found a possible explanation: poor communication between management and staff and between the university and students. De Vreede et al. (2014) recommend supplementing education with other support systems, such as including peer-to-peer support, demonstrating ⁎

real contribution, providing opportunities for leadership roles, and encouraging youth ownership to foster measurable change. In essence, measures being pursued on campuses have a behavioral component (e.g., following policies for paper purchasing, properly composting and recycling, choosing food that has been identified as more sustainable, choosing a lower-impact commute option, and accepting and supporting measures that may cost more), requiring temporary changes in access or asking people to change their habits to be successful. Additionally, studies have found that if campuses are to benefit from user engagement, effective communication is vital to improve sustainability outcomes (Sharp, 2002; Roorda, 2004; FranzBalsen and Heinrichs, 2007). Supporters of sustainable development and environmentalists acknowledge the importance of effective communication in making scientific findings more meaningful and impactful for a general audience (UNESCO, 1997). The literature indicates pressing needs for strategies to communicate sustainability in various contexts, such as in business (Rettie et al., 2012; Siano et al. 2016) and academic communities (Mazo and Macpherson, 2017). The eventual goal of communication is to change behavior through shared understanding. While there are studies on effective health communications (Zolnierek and DiMatteo, 2009) and environmental risk communication

Corresponding author. E-mail addresses: [email protected] (A.A. Kim), [email protected] (H. Sadatsafavi), [email protected] (L. Medal), [email protected] (M.J. Ostergren).

https://doi.org/10.1016/j.resconrec.2018.08.024 Received 1 August 2017; Received in revised form 2 August 2018; Accepted 26 August 2018 0921-3449/ Published by Elsevier B.V.

Resources, Conservation & Recycling 139 (2018) 366–376

A.A. Kim et al.

(UW). To address this important gap in the literature, this study investigated the following research questions:

(Newig and Fritsch, 2009), the university audience is quite different from the general public. Universities are distinctively characterized by three different population groups: students, staff, and faculty members. Generally, the student population rotates through the university more quickly than permanent staff and faculty groups. Franz-Balsen and Heinrichs (2007) discuss that the human factor is a critical component of sustainability on campus and that further studies investigating the higher-education-specific target audience and messaging would be valuable. Norazah and Norbayah (2016) had similar findings and found that campaign messages should be conveyed differently for different target populations. In studying staff members at higher-education institutions, Djordjevic and Cotton (2011) found that the difficulties of communicating about sustainability included lack of common definition, lack of a shared understanding of sustainability, and individual differences in values and attitudes. While many studies acknowledged the significant role of communication to encourage pro-environmental behavior, recent studies still indicate the lack of strategic communication use. In their experimental study, Godfrey and Feng (2017) found that the sustainability campaign led to a slight decrease in attitude toward pro-environmental actions being promoted, suggesting a disconnection between the scientific concepts in the environmental campaign and subjective ideas about sustainability. For example, an empirical study by Mazo and Macpherson (2017) suggests that universities are still failing to communicate their sustainability actions effectively, even if they incorporate environmentalism into their institutional identity statements or implement such sustainability initiatives. In summary, the authors are not aware of many studies that address the impact of specific sustainability information sources on awareness and on the intended behavioral actions, and that use repeated measures in a campus setting. The effectiveness of various university-initiated communication efforts and the university-specific opportunities for these groups to implement change in response has not been fully investigated.

• How do different information sources cluster together, or correlate with another, to better explain learning about sustainability? • How do groups (students versus staff versus faculty members) differ in learning from one type of cluster or another? • What is the relationship between awareness and intended behavioral action over time?

1.2. Literature review and theoretical background Developing effective communication strategies requires better understanding of communication sources, the institutional context, and the target audience to encourage desired behavior. In studying the various communication strategies to develop an effective campus sustainability campaign, the authors conducted a literature review that focused on three aspects and research questions: sources of communication (i.e., selecting and evaluating appropriate communication channels), content framing (i.e., understanding the target audience and controlling the message), and the link between sustainability awareness and behavioral action (i.e., measuring effective communication strategies). 1.2.1. Sources of communication 1.2.1.1. One-way and two-way communication. Various communication media can be grouped into one-way or two-way communication (Djordjevic and Cotton, 2011). One-way communication does not allow the direct involvement of the target audiences, and channels include blogs, newspapers, radio, television, websites, flyers, and research publications. Two-way communication allows direct involvement of the target audience where persuasion may occur, and channels include conference calls, emails, special events, exhibitions, classes, seminars, roundtable meetings, informal networking, and social media. There are advantages and disadvantages to both types of communication strategies. One-way communication offers an essential strategy with the purpose of disseminating information as objectively as possible (Morsing and Schultz, 2006). Information and news can be shared through a variety of media such as brochures, pamphlets, and magazines to inform the public. For example, a study by Mazo and Macpherson (2017) emphasized that printed materials on campus have a significant impact in sending clear and constant messages. Printed posters on campus are a core strategy for communicating sustainable initiatives to university stakeholders (Mazo and Macpherson, 2017). However, one-way communication is ineffective in fostering sustained behavioral engagement across a wide variety of study populations, from staff and students in higher education to general citizens (Ockwell et al., 2009; Corner and Randall, 2011; Wolf and Moser, 2011). Oneway communication is prone to misunderstanding of the intended message, which results in different views and interpretations and consequently tends to be insufficient in producing sustained behavioral engagement and expected actions. Therefore, it is critical that the communication media provide clear and efficient information about shared concerns, good intentions, and favorable decisions and actions to build and maintain positive stakeholder support (Morsing and Schultz, 2006). While one-way communication strategies are necessary for reaching broad audiences, recent studies have indicated that effective communication strategies require personal and face-to-face communication (Sharp, 2002; Barlett and Chase, 2004). Two-way communication can be explained by the theory of sense making, where two-way communication builds on ongoing iterative processes of not only sense making but also sense giving (Morsing and Schultz, 2006). Two-way communication channels, such as social media, can be successful in sharing information by identifying the channels that each university

1.1. Defining terms and objectives The terms communication source, mode, media, and channels are sometimes used interchangeably in the literature (Lee et al., 2002; DöUrso and Rains, 2008). The term communication source often refers to who or what the source of a piece of information is (Sundar and Nass, 2001). Sundar and Nass (2001) reported four different types of communication sources used in the online news environment: news editors, computers, other users, and the user himself or herself. Lee et al. (2002) described financial institutions, government/consumer agencies, and interpersonal (e.g., family and friends) as three different sources or channels of communication for electronic banking. Sundar and Nass (2001) further discussed three different conceptions of source:

• Concept 1: source as the visible gatekeeper-presenter of content, • Concept 2: source as the media technology that delivers the content, and • Concept 3: source as the receiver or audience choosing content for consumption.

In this paper, communication source refers to the second concept, where it represents the media technology that delivers the content. This concept proposes that the “medium or channel, not the sender, is the key that dictate the nature of content delivered through them” and “individuals respond to computers as a source in much the same way they respond to other human beings as sources” so that “the real source of messages is the technology qua medium itself.” The objective of this study was to quantify the effectiveness of various communication sources on the awareness of students, staff, and faculty members about campus-wide sustainability initiatives and their behaviors using longitudinal data from the University of Washington 367

Resources, Conservation & Recycling 139 (2018) 366–376

A.A. Kim et al.

stakeholder uses the most and ensuring that these channels are “regularly employed when sharing information” (Mazo and Macpherson, 2017). While two-way communication is effective, the challenges to incorporating two-way communication strategies are participants’ lack of time (Franz-Balsen and Heinrichs, 2007) and administrator burden. Therefore, balancing resources to maximize behavioral awareness and the impact on campus sustainability is important.

Correctly framing the message can motivate action even if the audience does not have a full understanding of the scientific evidence (Wolf and Moser, 2011). Some challenges to sustainability cannot be solved through rational methods alone. For example, Orr (2002) suggested that sustainability can also be perceived through spirituality. Additionally, McCright and Dunlap (2011) found that individuals’ political orientation is correlated with beliefs about scientific consensus and environmental concerns. While more education and understanding of climate warming may be positive for liberals, the same strategy was weaker or negative for conservatives, further emphasizing the need to better frame the message for different target audiences. Another message framing strategy uses emotion and communication: “In the pursuit of cognitive structure that predict conservation behavior, emotion has largely been ignored. However, there is strong potential for both positive and negative emotions to be both predictors of conservation behavior as well as mediators of predictor variables” (Vining and Ebreo, 2002). Emotion plays a role in communication and persuasion and is important for understanding conservation behavior. Lord (1994) studied the effects of message source and framing on recycling behavior and attitudes. While all types and sources of messages increased recycling behavior and improved attitudes, Lord found that positively framed messages tended to engender positive attitudes and belief in the message. However, negative fear-based appeals from personal acquaintances were the most effective means of increasing recycling behavior. Sustainability communication should avoid negative aspects such as the notion of altruism and sacrifice (Kaplan, 2002). Studies also suggest that the message should include information about the benefits of the initiatives and allocate responsibilities to individuals for sustainable actions to reduce skepticism (Richardson and Lynes, 2007; Horhota et al. 2014). The message should also be framed in such a way that it can reach a broad audience, meaning it not only attracts people that have interest in environmental benefits, such as the campus greening, but also targets people interested in the social and economic benefits, such as sharing the money-saving and the social impact (Dade and Hassenzahl, 2013). Specifically, in the context of higher education, Dade and Hassenzahl (2013) showed the need for additional scientific publications and best practices for communications. Excessive sustainability campaigns can be counterproductive, and a mechanism to control the messages should be in place. Horhota et al. (2014) found that uneven communication distribution is one of the barriers to sustainable actions, where some groups receive overwhelming information while other groups are unaware of some sustainability initiatives on campus. The frequency of information sharing needs to be considered, while maintaining clear, highly informative, and consistent messages (Djordjevic and Cotton, 2011).

1.2.1.2. Planned and incidental exposure. Sustainability information can also be shared through either planned or incidental exposure. Planned exposure in this study refers to exposure based on formal or non-formal education settings, such as sustainability courses, workshop, seminars, and other environmental trainings. In this setting, the audience must have the intention to obtain sustainability knowledge and have control of the exposure to such information. Research has extensively reported that the level of formal education correlates to sustainability knowledge, attitude, and actual behavior using the theoretical and empirical approach (Zsóka et al., 2013; Meyer, 2015; Chankrajang and Muttarak, 2017). In addition, sometimes incidental exposure is unavoidable. Incidental exposure may include in-person engagement with friends, family, international travel, or media (Swaim et al., 2014). Both types of exposure can explain the influence of social referents to intention. For example, studies promoted the important role of student peers in affecting sustainability intention (De Vreede et al. 2014; Swaim et al., 2014), in addition to other valid references for sustainability knowledge, such as professors, business leaders and political leaders (Swaim et al., 2014). 1.2.2. Content framing 1.2.2.1. Audience demographics. Understanding how different communication sources influence different target populations is beneficial. In order to increase user engagement, human factors play a significant role in shaping individuals’ commitment to sustainability issues (Franz-Balsen and Heinrichs, 2007; Corner and Randall, 2011). Most studies focusing on sustainability communication on campus have targeted single-audience instead of mix-audience groups (Djordjevic and Cotton, 2011; Zsóka et al., 2013; Meyer, 2015; Chankrajang and Muttarak, 2017; Lertpratchya et al., 2017). Recommendations include targeting different audiences by framing appropriate messages and considering political ideology or cultural world views (Marquart-Pyatt et al., 2011; McCright and Dunlap, 2011) in developing message content. However, literature about different target populations’ preferences for communication sources in a higher-education context is limited. The literature also emphasizes the significant effect of sociodemographic factors on people’s ability to behave more sustainably. Some of the sociodemographic factors of the target audience include age and level of education. Research has extensively investigated the correlation between education level and environmental behavior. For example, Zsóka et al. (2013) showed that the university sample had a significantly higher level of environmental knowledge than the high school cohort. Lertpratchya et al. (2017) suggested an effective communication strategy at higher-education institutions: considering the different groups of students based on their academic year because the study found that students in higher academic years reported somewhat more positive attitudes, beliefs, and sustainable actions. These behaviors support the theory of cultivation, where the longer people are exposed to media messages, the more they accept these messages as reflecting reality.

1.2.3. Sustainability awareness and behavioral action The connections between awareness as a result of a sustainability communication campaign and the actual behavior or action continue to be debated. One of the most influential theories for informing pro-environmental behavior change interventions is the theory of planned behavior (TPB) (Ajzen, 1991), which explains that behavior change can happen when individuals perceived that they have sufficient control over their actions and that certain norms suggest that they should act in a different way. Some studies reported that there is no direct relationship between an increase in knowledge of environmental concerns/ awareness and engagement (Geller, 1981; Kollmuss and Agyeman, 2002; Futerra, 2010; Wolf and Moser, 2011). Other studies found that to some extent, environmental education is one of the external factors that could impact students’ pro-environmental behavior (Lukman et al. 2013; Chankrajang and Muttarak, 2017). Moreover, other factors also influence the extent of an individual’s pro-environmental behavior, such as the visibility and accessibility of communication sources (Godfrey and Feng, 2017), the availability of options, and the degree of effort in performing the sustainability actions (Stern, 2000; Kagawa,

1.2.2.2. Message-framing strategies. The key to effective communication is matching the communication channel with the goal of the message (Barry and Fulmer, 2004). Effective communication strategies include strategically framing the message to enhance shared understanding about sustainability and move toward the intended behavior change. 368

Resources, Conservation & Recycling 139 (2018) 366–376

A.A. Kim et al.

composting and solar panels on roofs), setting policies (e.g., consolidating servers for efficiency), hiring staff to focus on sustainability (e.g., in Facilities, Housing, and Food Services; Transportation Services; and the Sustainability Office), and providing funds to support independent efforts (e.g., for student-led sustainability projects). These characteristics made UW a favorable choice to investigate further. The analysis in this study is based on a survey that was sent as part of the university’s larger sustainability goals of understanding the effectiveness of different communication methods.

2007; Boyes, Skamp et al., 2008). De Vreede et al. (2014) focused on ways to encourage high school students to take sustainability actions and found that a holistic educational approach is better than a conventional teaching model for high school students. The study was based on a peer-approach, youth-led educational program to increase knowledge and engagement about sustainability issues. The study reported that the value of peer support, contributions through tangible and meaningful activities, and opportunities to take on a leadership role and ownership can promote personal pro-environmental behavior and encourage further sustainability actions such as educating others. The findings show the dynamics and importance of positive relationships among peers with adult coaches. Lavelle et al. (2015) suggested that when designing sustainable intervention, the strategies should be tailored to not only particular population groups but also different behavior types, rather than using a one-size-fits-all approach. The study emphasized the need to distinguish habitual and occasional pro-environmental behavior. While occasional behavior is a non-routine action that may require a certain degree of sacrifice with psychological and financial impact, habitual behavior is a daily routine that is sustained as an individual’s lifestyle.

2.2. Data source and measurement approach This study used self-reported data from UW students, staff, and faculty members about communication sources they use for receiving information, their levels of awareness about campus sustainability initiatives, and actions they perform in their offices or dorms. The initial survey in 2012 was motivated to create an informed communications plan that could raise awareness and interest in the commitments, goals, and initiatives associated with the UW CAP (University of Washington 2009). The follow-up survey in 2014 was conducted to measure the impact of those communication efforts two years later. This study used existing data sets collected by UW but reassessed the impact using other analytical methods.

2. Materials and methods This study used exploratory factor analysis (EFA) with the maximum likelihood extraction method to identify groupings of different types of media used to receive sustainability information. Researchers analyzed feedback from two sets of surveys, one from 2012 and one from 2014, and compared the data sets. Researchers also identified the relationship between various groupings of communication sources and awareness, and between these groupings of communication sources and actions taken. Furthermore, these relationships were analyzed for different groups of people (i.e., undergraduate students, graduate students, staff, and faculty members).

2.3. Survey content and participants Graduate students under the leadership of a faculty advisor from UW’s Foster School of Business wrote and executed the 2012 survey. The survey content, deployment, and descriptive analysis were completed as part of the Master of Business Administration curriculum. This particular effort, the keystone project, was a two-quarter-long project conducted with a community partner to address contemporary problems in environmental management. The survey included 13 questions, which were sent via email to 58,825 individuals on the UW campus in February 2012. The 2012 keystone project team compiled and generated a list of UW email addresses and sent the survey and reminders directly to those people. The team did not follow or use the central bulk email process. Reminders were sent out the following day and also one week later. The survey closed approximately three weeks after first being deployed. Table 1 includes the topics addressed in the survey along with example questions and response choices. In 2014, an improved survey instrument was developed based on the earlier version but did include the questions shown in Table 1. Only shared questions between the 2012 and 2014 surveys are used in this analysis. In other words, 2012 and 2014 questions used in this study (Table 1) were identical. The 2014 survey was emailed to 93,000 recipients in February 2014 following the central bulk email process. The amount of time and resources available to conduct the 2014 survey was considerably less, and no reminders were sent to participants. Reminders were discouraged because there was pushback about sending emails to the entire campus. In terms of the data cleanup and screening approach, researchers excluded all cases with missing values rather than imputing new values given the large sample size. Because the survey was short and confidential and did not ask for sensitive information, the number of missing values was very low, and less than 1 percent of cases from each year were excluded.

2.1. University of Washington UW is one of the largest public universities located in a densely populated setting, creating a complex context. With over 75,000 students, faculty, and staff, UW is larger than many small cities. The actions of the university have substantial impact on neighboring communities, and building and infrastructure changes must be coordinated with the City of Seattle. UW consistently ranks high in its sustainability performance relative to other institutions of higher education. In the most recent reports, it was 10th in the Sierra Club’s Cool School ratings (Sierra, 2015), was one of only 21 schools on the Princeton Review’s Green College Honor Roll (Princeton, 2017), and has achieved a Gold rating in the Sustainability Tracking, Assessment, and Rating System (Eklund, 2015). UW was chosen and used in this study because of its long history of sustainable practices (see “A Century of Sustainability,” an interactive timeline, at https://green.uw.edu/timeline). For example, the U-Pass was introduced in 1991 to make it easier for UW commuters to use mass transit, and UW was an early adopter of a recycling program. In 2004, UW drafted an environmental stewardship policy and created the Environmental Stewardship Advisory Committee to oversee progress toward the policy goals. In 2007, UW signed the American College and University President’s Climate Commitment. The Environmental Stewardship and Sustainability Office (now the Sustainability Office) was created a year later to help fulfill this commitment, including support for the effort to create a Climate Action Plan (CAP). The plan was completed in 2009, also the year when students created the Campus Sustainability Fund to support student-led sustainability projects. UW also maintains its own Sustainability Dashboard as part of that commitment (see at http://green.uw.edu/dashboard). Efforts to improve performance in these areas are considerable and span the gamut of approaches including upgrading infrastructure (e.g., bins for

2.4. Analysis approach The IBM® Statistical Package for the Social Sciences® and the Analysis of Moment Structures® version 24 were used for data analysis. First, researchers performed EFA with the maximum likelihood extraction method to investigate the dimensionality of communication methods. In performing EFA, oblique solutions were chosen (Pett et al., 2003; Ho, 2006). EFA was chosen to reduce the number of independent variables (communication strategies) in the model from 12 variables to 369

Resources, Conservation & Recycling 139 (2018) 366–376

A.A. Kim et al.

population and sample characteristics in terms of role is included in the Appendix.

Table 1 Details of the communication survey. Topic Addressed in the Survey

Example (with Response Choices)

Frequency of media used to receive information from UW

People receive news and information in a variety of ways. Please indicate how frequently you use the following types of media to receive information about UWa: -Frequently -Sometimes -Rarely -Never UW has many initiatives to reduce its carbon emissions. How aware are you of the following ongoing initiatives at UWb: -Very aware -Somewhat aware -Not aware -I don’t know what this is Please indicate how frequently you do the following when you leave your office and/or dorm roomc: -Almost always -Sometimes -Never -I do not have control over this/not applicable to me

3.2. Average use of communication source

Awareness of ongoing sustainability initiatives at UW

Sustainability actions taken by respondents

Email, online news sources, classes and seminars, and radio news sources were the most common sources of information across all role groups in both rounds of data collection. Among all communication source groups, Twitter is the least commonly used information sources in both years and in all different audience groups although the frequency slightly increased in 2014 for all four audience groups. Table 2 shows the average use of different communication media by study participants in each year.

3.3. Average levels of awareness and action Table 3 shows the average level of awareness and action for each role group in 2012 and 2014, along with the five dimensions of the action. A box plot representation is included in the appendix, comparing 2012 and 2014 levels of awareness and action for each role group. Table 3 shows that for all roles, the 2014 mean awareness level is higher than 2012 mean awareness level, while the difference in mean action level between groups is marginal in 2012 and 2014. Staff and faculty members show the highest levels of awareness in both rounds of data collection, while in terms of action, staff members have the lowest level of action in both 2012 and 2014. When the five specific action items are compared, for all group, turning off computers and adjusting temperature have the lowest level of action, while turning off lights and fans has the highest levels of action in both 2012 and 2014. This could be explained by some staff and faculty members being located in buildings that have centrally controlled mechanical systems, hindering their ability to adjust the thermostat. Many of the staff and faculty members also have remote desktop access and would prefer not to turn off their computers.

a This question covered 13 types of media: email, Facebook, Twitter, blogs, online news, hardcopy news, televised news, radio news, billboards, bus shelter advertisements, flyers, classes and seminars, and booths on Red Square (UW’s central plaza). Since the survey was administrated via email, “using email” was excluded from the final communication strategies assessed in this study. b This questions covered nine sustainability initiatives: developing a regional smart grid network, offering “Zimride,” operating free shuttle services around campus, monitoring all campus buildings’ energy consumption, offering bike lockers/racks/secure indoor bike rooms throughout campus, purchasing only Energy Star–rated appliances, monitoring all campus buildings’ water consumption, offering telecommuting options for staff, and committing to a CAP to reduce campus emissions. c This question covered five actions: turning off computers, adjusting the temperature, making sure windows are closed, turning off lights, and turning off fans.

3.4. Dimensionality of communication methods

a smaller number of strategy groups and to simplify the interpretation of the model. More importantly, EFA helped identify the underlying pattern of how the study participants used the 12 communication strategies. Next, researchers used structural equation modeling (SEM) techniques and performed multivariable path analyses to simultaneously test and estimate the relationships between communication methods and the levels of awareness and actions among the study participants. The choice of SEM over regression analysis was made because the theoretical model included more than one dependent variable. More importantly, SEM allowed the analysis of both the measurement part (the factor structure identified via EFA) and the structural part (modeling dependent constructs as linear functions of the independent constructs) in one model. In both EFA and SEM, separate models were run for each role group for each year.

Eight different EFA models were run, one for each role group in 2012 and 2014. In all models, the measures of sampling adequacy described by Kaiser (1960) were higher than the minimum acceptable value (0.6) recommended by Hair et al. (1998). Bartlett Tests of Sphericity were also statistically significant in all eight models at p < 0.000, indicating that the variables are not completely uncorrelated. Email was removed from this analysis because of the high levels of use among all groups. The authors consulted scree plots (Cattell, 2010) and retained factors with eigenvalues of one and greater (Kaiser, 1960). All eight models produced identical structures, in terms of the number of factors and items that clustered together within each factor. In all models, EFA resulted in three factors, with news sources (radio news, televised news, hardcopy news, and online news) clustering together, blogs and social media sources (Facebook and Twitter) clustering together, and local sources of information (flyers, bus shelter advertisements, classes and seminars, booths, and billboards) clustering together. No cross loading was identified in any of the eight models. The three-factor structure explained between 39 and 44 percent of the variance in the data. In summary, the structuring of variables in all eight EFA models was simple, theoretically meaningful, and, more importantly, similar across groups. Table 4 shows the correlation among communication media when all responses are pulled together, and shows the overall factor structure indicated by EFA.

3. Results 3.1. Sample characteristics In 2012, 9050 responses were received, and in 2014, 1631 responses were received, resulting in a response rate of 15% and 1.75 percent, respectively. The lower response rate in 2014 was most likely due to administrative constraints. In both years, staff were overrepresented and undergraduates underrepresented in the pool of survey respondents. A breakdown of survey respondents relative to campus 370

Resources, Conservation & Recycling 139 (2018) 366–376

A.A. Kim et al.

Table 2 Average use of communication media by study participants. 2012

Email Facebook Twitter Blogs Online news Hardcopy news Televised news Radio news Billboards Bus shelter advertisements Flyers Classes and seminars Booths on Red Square

2014

Undergraduate

Graduate, Professional

Staff

Faculty

Undergraduate

Graduate, Professional

Staff

Faculty

2.83 (0.46) 1.93 (1.16) 0.44 (0.88) 0.56 (0.85) 1.58 (1.04) 1.19 (0.96) 1.05 (0.97) 0.85 (0.95) 0.75 (0.84) 0.76 (0.84) 1.13 (0.90) 1.64 (1.00) 0.93 (0.92)

2.85 (0.47) 1.41 (1.19) 0.30 (0.68) 0.60 (0.85) 1.74 (1.03) 1.05 (0.91) 0.96 (0.96) 1.26 (1.08) 0.70 (0.80) 0.77 (0.84) 0.86 (0.83) 1.57 (0.98) 0.42 (0.66)

2.87 (0.44) 0.86 (1.10) 0.24 (0.62) 0.54 (0.82) 2.04 (0.95) 1.37 (1.02) 1.48 (1.05) 1.62 (1.06) 0.81 (0.82) 0.80 (0.84) 0.95 (0.83) 0.92 (0.87) 0.34 (0.59)

2.87 (0.44)

2.78 (0.58) 1.81 (1.13) 0.45 (0.86) 0.50 (0.81) 1.36 (1.06) 1.04 (0.97) 0.75 (0.89) 0.61 (0.86) 0.62 (0.79) 0.79 (0.85) 1.20 (0.91) 1.66 (0.98) 0.89 (0.82)

2.87 (0.41) 1.43 (1.19) 0.36 (0.72) 0.54 (0.77) 1.57 (1.04) 0.94 (0.86) 0.64 (0.85) 1.01 (1.06) 0.65 (0.77) 0.87 (0.86) 0.93 (0.89) 1.47 (0.98) 0.50 (0.71)

2.88 (0.43) 0.83 (1.04) 0.33 (0.72) 0.58 (0.80) 1.95 (0.98) 1.33 (1.01) 1.29 (1.04) 1.57 (1.06) 0.82 (0.78) 0.86 (0.83) 0.99 (0.84) 0.87 (0.84) 0.36 (0.60)

2.83 (0.58) 0.73 (1.02) 0.17 (0.47) 0.46 (0.70) 1.90 (1.04) 1.57 (1.11) 1.19 (1.05) 1.73 (1.10) 0.68 (0.74) 0.73 (0.80) 0.89 (0.79) 1.00 (0.94) 0.36 (0.59)

0.62 (0.95) 0.16 (0.47) 0.49 (0.76) 1.91 (0.97) 1.56 (1.08) 1.22 (1.05) 1.75 (1.07) 0.73 (0.75) 0.62 (0.74) 0.88 (0.79) 1.00 (0.91) 0.30 (0.55)

Notes: Values in parentheses show standard deviation. Mean values are calculated by assigning a value of 0 to “never,” 1 to “rarely,” 2 to “sometimes” and 3 to “frequently.” Email is removed from the subsequent analysis because the high level of usage by all participants will introduce bias in the analysis. Table 3 Average level of awareness and action among study participants. 2012

Overall awareness Overall action Action—turn off my computer Action—adjust the temperature Action—make sure windows are closed Action—turn off lights Action—turn off my fan

2014

Undergraduate

Graduate, Professional

Staff

Faculty

Undergraduate

Graduate, Professional

Staff

Faculty

0.57 (0.44) 1.52 (0.42) 1.25 (0.74) 1.16 (0.82) 1.56 (0.68) 1.85 (0.39) 1.76 (0.53)

0.49 (0.40) 1.59 (0.42) 1.24 (0.78) 1.27 (1.78) 1.72 (0.55) 1.91 (0.31) 1.90 (0.35)

0.72 (0.45) 1.38 (0.60) 0.98 (0.86) 0.81 (0.82) 1.64 (0.67) 1.82 (0.46) 1.86 (0.46)

0.64 (0.42) 1.54 (0.47) 1.16 (0.82)

0.71 (0.41) 1.58 (0.37) 1.26 (0.77) 1.20 (0.83) 1.67 (0.62) 1.91 (0.31) 1.87 (0.40)

0.65 (0.39) 1.61 (0.42) 1.26 (0.80) 1.32 (0.80) 1.76 (0.53) 1.90 (0.32) 1.88 (0.41)

0.83 (0.43) 1.35 (0.61) 0.93 (0.87) 0.79 (0.82) 1.70 (0.61) 1.81 (0.48) 1.85 (0.45)

0.78 (0.42) 1.60 (0.42) 1.22 (0.83) 1.00 (0.83) 1.82 (0.47) 1.91 (0.33) 1.88 (0.46)

0.83 (0.83) 1.77 (0.52) 1.92 (0.30) 1.93 (0.34)

Notes: Mean awareness is calculated by assigning a value of 0 to “I don't know what this is” and “not aware,” 1 to “somewhat aware,” and 2 to “very aware”. Mean action is calculated by assigning a value of 0 to “never,” 1 to “sometimes,” and 2 to “almost always”. Since scales used to operationalize awareness and action had different number of points, within each year awareness and action values are not comparable. Nevertheless, 2014 awareness values are comparable to 2012 awareness values, and 2014 action values are compare to 2012 action values as they had identical scales.

freedom = 238, p < 0.001). Table 5 shows the results of the multigroup SEM analysis and shows the direct effect of the exogenous variables on the endogenous variables. Table 6 shows the total effects, expressed in terms of the percentage of change in the endogenous variables as a result of a 1 percent change in the exogenous variables. Table 6 shows that all three factors have positive impacts on awareness. For staff members, news sources were more effective in enhancing awareness than the other two groups of strategies, while for students, blogs and social media were more

3.5. Relationships between communication methods and awareness and actions Fig. 1 shows the structural equation model used for understanding the relationships between communication methods and the levels of awareness and actions among the study participants. Unlike the EFA models, the chi-squared difference test indicated that a multigroup model fits the data significantly better than a model with all responses pooled together (Δχ² = 6672.3, degree of 371

Resources, Conservation & Recycling 139 (2018) 366–376

A.A. Kim et al.

the CAP and to communicate those efforts to the campus. The Sustainability Office often worked with students in the Program on the Environment to provide projects for their keystone (graduate) or capstone (undergraduate) projects. The survey evolved from the need to provide direction for communications and an opportunity to get student help. Conversely, the 2014 follow-up was conducted by a student intern with the goal of determining the effectiveness of communication efforts. With just one student, the amount of time available to conduct the 2014 follow-up survey was considerably less, and no reminders were sent due to pushback from the university about sending out emails to the entire campus. Two other possible limitations in the study are nonresponse bias and common-method bias. This assessment shows that the risk of the two biases is low in this study. For the nonresponse bias, the low response rate in the 2014 survey could be concerning. The best approach to test the risk of non-response bias is to compare the demographic characteristics of early versus late responses (Lindner et al., 2001). Unfortunately, survey submission data were not available to the research team, and it was not possible to see whether respondents’ characteristics changed over time. However, nonresponse bias is not likely to be a source of systematic error in the results for two reasons. First, as studies have suggested (Olson, 2006; Tourangeau et al., 2010; Pope and Stanistreet, 2017) the survey was designed (in terms of the survey length, confidentiality, and sensitivity of information requested) and administered in a way that did not make it more likely for certain groups to refuse to participate or be absent during the surveying period. This is evident in the fact that, as the appendix shows, the percent of participants from each group stayed relatively the same in round 1 (in 2012) and round 2 (in 2014) of the data collection (i.e., sample characteristics are similar in round 1 and round 2). Second, the data collection was conducted in a homogenous population on a university campus, and it is safe to assume that despite differences in the sample size, the samples in both rounds of data collection were essentially random subsets of the full survey sample and good representations of the broader group of potential participants (Holbrook et al., 2007). Common-method bias might exist in the study because—due to the limitations in the measurement procedure and the lack of objective measures—the research team had to use a single method (a survey questionnaire) to measure all study variables. However, Harman’s single factor test indicated that when all survey items are loaded on a single factor (with no rotation), only 19 percent of the total variance is

Table 4 Rotated factor matrix indication factor structure and portion or variance explained by each factor.

Radio news Televised news Hardcopy news Online news Flyers Bus shelter advertisements Classes and seminars Booths on Red Square Billboards Twitter Blogs Facebook

Factor I

Factor II

Factor III

Variance Explained

0.77 0.65 0.51 0.46 0.17 0.30 0.03 −0.05 0.38 0.05 0.18 −0.01

0.01 0.08 0.18 0.08 0.71 0.60 0.53 0.53 0.50 0.10 0.12 0.23

0.00 0.03 0.05 0.28 0.05 0.11 0.17 0.23 0.10 0.67 0.67 0.48

24.06%

10.07%

6.94%

Notes: The factor correlation matrix showed that the correlations among factors were not high (between 1.52 and 3.4). As a result, Varimax rotation with Kaiser normalization was used (Nunnally and Bernstein, 1994; Hinkin, 1995). Rotation converged in five iterations. Items with a loading of greater than 0.40 on a specific factor were assigned to the corresponding factor (Kim and Mueller, 1978). No cross loading was identified.

effective in enhancing awareness. Regarding the impact of communication strategies on action, in 2014, the total effects of communication strategies on action are very small or, in some cases, negative. Generally speaking, the relationship between communication strategies and awareness/action seems to be smaller in 2014 than 2012, which may have been due to a much larger sample size in 2012. The impact of awareness on action is very small in both years.

3.6. Limitation of the analysis The difference in the 2012 and 2014 survey response rates can be explained by the availability of resources and the current institutional barrier to sending out emails to the entire campus on an ongoing basis. The initial (2012) survey was conducted to inform a communications plan to raise awareness and interest in the commitments, goals, and initiatives associated with UW’s CAP. In 2012, the Sustainability Office had a dedicated staff member whose job was to support efforts to follow

Fig. 1. SEM diagram showing the relationships between communication methods and awareness and action. The level of concern about the possibility of climate change and level of concern about the ability to stop it were included as two control variables in the model (predictors of awareness and action). 372

Resources, Conservation & Recycling 139 (2018) 366–376

A.A. Kim et al.

Table 5 Parameter estimates of the multigroup structural equation model. 2012 Endogenous Variable

Exogenous Variablea

b

2014

Undergraduate

Graduate, Professional

Staff

Faculty

Undergraduate

Graduate, Professional

Staff

Faculty

Path Model

Awareness Factor I Factor II Factor III Factor I Factor II Factor III

-> -> -> -> -> -> ->

Action Awareness Awareness Awareness Action Action Action

0.09 0.07 0.12 0.16 0.07 n.s. −0.09

0.08 n.s. 0.08 0.19 0.10 n.s. n.s.

0.09 0.20 0.07 0.14 n.s. n.s. n.s.

0.12 0.13 n.s. 0.12 n.s. n.s. n.s.

0.21 n.s. n.s. 0.16 n.s. n.s. −0.19

n.s. 0.21 n.s. n.s. n.s. n.s. −0.20

n.s. 0.15 n.s. 0.14 n.s. n.s. n.s.

0.29 n.s. n.s. n.s. n.s. n.s. n.s.

Measurement Model

Factor Factor Factor Factor Factor Factor Factor

I I I I II II II

-> -> -> -> -> -> ->

0.50 0.50 0.72 0.74 0.70 0.39 0.84

0.60 0.47 0.69 0.62 0.71 0.35 0.82

0.57 0.37 0.69 0.67 0.74 0.34 0.79

0.64 0.33 0.68 0.67 0.73 0.35 0.74

0.56 0.55 0.71 0.67 0.69 0.33 0.73

0.58 0.48 0.68 0.66 0.71 0.49 0.73

0.53 0.41 0.70 0.65 0.64 0.35 0.84

0.68 0.42 0.70 0.64 0.74 0.34 0.78

Factor Factor Factor Factor Factor

II II III III III

-> -> -> -> ->

Hard copy news Online news Radio news Televised news Billboards Booths on Red Square Bus shelter advertisements Classes and seminars Flyers Blogs Facebook Twitter

0.37 0.56 0.73 0.34 0.62

0.34 0.61 0.76 0.47 0.61

0.40 0.60 0.72 0.55 0.68

0.33 0.59 0.69 0.50 0.69

0.31 0.40 0.67 0.43 0.57

0.41 0.66 0.74 0.41 0.58

0.32 0.60 0.66 0.61 0.69

0.41 0.58 0.64 0.64 0.72

Note: Awareness and action are log transformed to deal with skewness in these two variables. Chi-square ( χ 2 ) = 2952.69, χ 2 /degree of freedom = 4.34, 90% confidence interval for root mean square error of approximation (RMSEA) = 0.018 and 0.020, standardized root mean square residual (SRMR) = 0.073, comparative fit index (CFI) = 0.92. Modification indices were used to improve model specifications. Cut-off values for fit indexes were minimum value of 5 for χ 2 /df (Schumacker and Lomax, 2004). RMSEA values of 0.05 as an indicator of a good fit (Browne and Cudeck, 1992), SRMR values less than 0.08 for acceptable fit (Hu and Bentler, 2009), and CFI value of 0.90 for the acceptable fit. n.s.: not significant at p < 0.05. a Latent variable in the measurement part of the model. b Measurement item in the measurement part of the model. Table 6 Total effect of communication factors on awareness and actions. 2012 Exogenous Variable Awareness Factor I Factor II Factor III Factor I Factor II Factor III

-> -> -> -> -> -> ->

2014

Endogenous Variable

Undergraduate

Graduate, Professional

Staff

Faculty

Undergraduate

Graduate, Professional

Staff

Faculty

Action Awareness Awareness Awareness Action Action Action

0.09 6.72 12.98 17.59 8.12 1.14 −7.59

0.08 0 8.55 20.68 9.97 0.69 1.59

0.09 22.14 7.36 14.57 1.76 0.62 1.19

0.11 13.43 0 13.20 1.46 0 1.44

0.20 0 0 17.23 0 0 −14.48

0 23.86 0 0 0 0 −18.05

0 16.65 0 15.49 0 0 0

0.29 0 0 0 0 0 0

Note: For the effect of awareness on action (both log transformed), the following equation is used to calculate the percentage change in action as a result of one percent change in awareness: [(1.01) β − 1]*100 , where β is the estimated effect size. For the effect of Factors I to III on awareness and action, the following equation is used to calculate the percentage change in awareness or action as a result of one percent change in each factor: (e β − 1)*100 , where β is the estimated effect size.

into three groups: news sources (radio news, televised news, hardcopy news, and online news), blogs and social media sources (Facebook and Twitter), and local sources of information (flyers, bus shelter advertisements, classes and seminars, booths, and billboards). In general, the top three most frequently used communication sources (email, online news, and classes and seminars) fall under different clusters of blogs and social media, news sources, and local sources, respectively. These findings show that all cluster of sources—either one-way (e.g., online news) or two-way communication (e.g., email, classes, and seminars), and planned (e.g., classes and seminars) or incidental exposure (e.g., online news and social media)—are important to increase awareness. These findings also concur with previous studies (Stephens et al., 2008; Leonardi et al., 2012) that found clustering between communication channels that allow for asynchronous communication (e.g., email) and for synchronous or instant

associated with that single component, which is lower than the 50 percent threshold suggested by Podsakoff et al. (2003). As a result, the risk that common-method bias could affect the results of the study is low. Although SEM results were satisfactory and the respondent sample was large, given that the research team was not able to match individual-level responses in 2012 and 2014, study relationships identified in the model are not necessarily causal relationships and should be verified using time-series data. 4. Discussion 4.1. All sources were effective in increasing awareness—continue to repeat information and combine channels In this analysis, the different communication strategies clustered 373

Resources, Conservation & Recycling 139 (2018) 366–376

A.A. Kim et al.

(Hope et al., 2014; Jiang et al., 2015). For undergraduate students, blogs and social media sources were most effective. This finding is similar to studies where students found mobile computing devices and social media to be effective in communicating and learning with appropriate institutional support and training (Barnes and Lescault, 2011; Gikas and Grant, 2013). Students’ preferred choice of communication sources included the internet and social media (Vrocharidou and Efthymiou, 2012; Kim and Kim, 2017) and students saw benefits to their well-being. It would therefore be advantageous to invite and engage on their social networks to spread proenvironmental behaviors. In fact, Olli et al. (2001) found that the environmental network is a strong predictor of engaging in pro-environmental activities, and relevant studies (Haythornthwaite, 1996; Fell et al., 2009) have indicated that targeting social networks as opposed to individuals can be more effective in enhancing pro-environmental behavior. At UW, the Green Building Standards Committee seeks to engage students as stakeholders to trigger new thinking and to keep focused on value-driven goals. As part of that effort, the committee has recruited students from all levels and disciplines (e.g., architecture, engineering, and environment) through a mix of purposive sampling and snowballing recruitment strategy. The university is aiming to break new territory by integrating students as stakeholders in selecting and discussing sustainability features for new campus buildings. The first target building is the new and proposed Population Health Building. This building carries significant meaning because the vision is to create a living laboratory that will improve the health and well-being of people in the region and throughout the world. Students are selectively being invited to meetings during the design process to inform decisions. Students will be able to gain real-world experience because the design firm has expressed a desire to work with students in further providing education and hands-on learning opportunities. The goal is that the experience creates direct engagement and ownership to foster informal conversations with peers on campus to enhance pro-environmental behaviors. Another example from the university is the use of instructional platforms and human resources (e.g., faculty and students) unique to educational organizations. Hogg and Shah (2010) found that learning about climate change issues in school or later in life reduced the sense of powerlessness in making a difference. Furthermore, UW faculty have been invited to explore ways to incorporate sustainability across the curriculum. In 2014, a Sustainability across the Curriculum Committee developed a list of 12 to 24 students, staff, and faculty members to represent the community and will help advance sustainability in curriculum efforts. The activity report suggests continuing to build sustainability in the community, establishing staff and faculty leadership positions, and linking efforts in all university campuses (Mackenzie et al., 2015).

communication channels (e.g., telephone or face-to-face communication). In addition, Rice (1993) found intimacy or closeness of technology such as new and traditional technology in clustering of communication methods. The findings included three different sources of communication. In those clusters, there was a mix of asynchronous channels (e.g., flyers because they may be sent out regardless of whether the receiver is ready) and synchronous channels (e.g., face-to-face and interactive seminars, which are time dependent). One explanation of why all three sources were effective in increasing awareness in all university role groups could be the use of mixed sources and channels. For example, Stephens et al. (2013) studied the reactions to redundant communication in order to strategize emergency messages and efficiently capture attention. (Stephens, Barrett et al.) found that receiving a redundant message with at least one synchronous channel led to a perceived higher sense of urgency. Similarly, Cover and Thomas (2012) indicated portraying information through multiple communication channels over time could be effective. Email is the most effective source of communication on campus used by all participants groups, which can be explained since all participants must use a campus email account as their main communication source and media in their daily campus activities. The literature also supports the overwhelming preference of students and younger adults to use two-way dialogues and internet-based communication sources, which in this study fall into the blogs and social media group, as well as online news. Interestingly, the survey reported that Twitter and blogs are the least frequently used by the study participants. On the other hand, studies have highlighted the personalized, interactive, and collaborative benefits of Twitter in higher education and that a more frequent use of Twitter was related to more positive beliefs about the educational relevance of Twitter (Carpenter and Krutka, 2014; Tur et al., 2017). Moreover, in a comprehensive and generalizable study of how U.S. higher-education institutions use Twitter, Kimmons et al. (2017) suggest that institutions explore ways to use Twitter more meaningfully, such as using it as an efficient two-way communication platform instead of one-way messages to disseminate information. Only in that way can social media platforms affect positive changes within higher-education institutions and communities at large. While classes and seminars can be effective, there were no increase of use in 2014 as compared 2012. This finding suggests the need to be more innovative in terms of classes and seminars to increase their effectiveness (De Vreede et al., 2014; Swaim et al., 2014). At UW, the Sustainability Office works to create consistent and redundant messages by coordinating with academic and operations partners across campus. These messages are communicated through newsletters, websites, social media, posters, and campus events related to implementing sustainability efforts. In terms of creating the messages, content is more than just the words or information conveyed. Images, pictures, and music frequently accompany messages. The message is used to create emotions through pictures, photographs, certain color schemes, and tone of voice. An example of this is a video message created by the UW Paper Reduction Committee, which inspires viewers to participate in minimizing paper waste with examples, explanations, and ideas communicated by interested, passionate individuals speaking from beautiful settings across campus, accompanied by an upbeat soundtrack (UW Sustainability 2017).

4.3. Increase in awareness does not guarantee actions—consider further removing institutional constraints The study results indicate that while communication in general led to more awareness of sustainability initiatives on campus, it did not necessarily lead to more actions. A similar phenomenon was also found in another study, where the sustainability campaign eventually decreased attitudes about pro-environmental actions (Godfrey and Feng, 2017). They indicated that time limitations, the power of routines, and different understandings of sustainability as the possible reasons. This could also be the case in this study since the five pro-environmental actions measured in this study were all considered habitual behavior that could have been embedded in the individual. The findings support previous studies, where while good communication via different communication sources raises awareness to new information, awareness itself does not always lead to action (Gill et al.,

4.2. Individual groups respond differently to common communication sources—Foster effective engagement opportunities within individual groups to take active roles Of the three clusters, local sources of information were the least effective strategy for enhancing sustainability awareness for all groups. Overall, for graduate and professional students and staff, news sources were most effective. This finding is supported by studies that have found that information sharing through the internet or other electronic media is not the most effective way to communicate with older adults 374

Resources, Conservation & Recycling 139 (2018) 366–376

A.A. Kim et al.

communication sources were successful in enhancing the awareness of participants about the sustainability initiative on campus, there was little or no impact on their behaviors. The lack of impact on behavioral changes can be explained by the high initial level of action taken by participants, regardless of their low level of awareness about campusspecific sustainability initiatives. This study demonstrates that there is considerable variability in the awareness and behavioral changes associated with the type of communication sources used and the nature of the subjects (i.e., students, staff, and faculty members) to whom these sources were applied. In order to increase behavioral changes, the university should continue to combine multiple communication sources and repeat information, provide engagement opportunities, and use participatory methods to encourage further peer-to-peer support.

2008; Wolf and Moser, 2011; Horhota et al., 2014). For example, Gill et al. (2008) investigated the sustainability disclosure of oil and gas firms across North America, Asia, and Europe using automated web content technology. While the firms show awareness of sustainability reporting, they did not necessarily practice it, such as disclosing sustainability issues or documenting that information in the reports. This finding is also consistent with what (Podsakoff et al., 2003; Corner and Randall, 2011) discuss: social marketing strategy is not enough to engage the public. Failure to cause behavioral changes can also be explained by other key aspects of the communication process, which have to be supported in order to lead to behavioral changes (Moswer, 2010). In this study, the lack of impact on behavioral changes can be explained by the high initial level of action taken by participants, regardless of their low level of awareness about campus-specific sustainability initiatives. Further institutional barriers, such as low opportunity to practice (Mezirow, 1991; Brinkerhoff and Montesino, 1995), low support or lack of reinforcement and coaching (Clarke, 2001), and the difficulty of obtaining information (Stern, 2000), should be removed for participants to actively engage in campus sustainability initiatives, and there must be an opportunity or possibility to act proenvironmentally. For further behavioral changes, successful solutions may include developing a coordinated solution among different university audiences and involving cross-collaboration among these groups. These solutions should require little effort from the participants, such as providing easy access by locating the action in high-traffic areas or offering information about the action. For example, UW weaves sustainability throughout its orientation programs. At UW’s Freshman Orientation, students are introduced to issues of sustainability through a vignette or skit in a portion of the program. The program demonstrates how recycling works at UW, introduces UW’s culture of thinking about sustainability, educates students on institutional efforts to reduce energy consumption and waste, and introduces student organizations dedicated to educating the community about these issues. There, students may learn that they actually have multiple options for tossing waste. All around campus are trios of waste receptacles that include trash, recyclables, and compost-ready waste. The UW Recycling Unit is also supported by a hands-on interactive map that can be used to find different resources, including places to recycle. Concurrently, in Parent Orientation, parents are given UW-imprinted water bottles that they can use throughout the day in place of plastic or paper cups throughout campus, further reducing waste. Locations of water-bottle-filling stations are provided but also strategically positioned around the entire campus for easy access. Parents are also informed and educated about UW’s sustainability efforts when they receive the water bottle. The bottles serve the larger purpose of reminding parents of the community commitment to the environment when they use the water bottle. All of this is supported by UW Academic Affairs’ students and staff, who receive training from the UW Recycling Unit (as well as the UW Moving and Surplus Unit) and are very conversant in issues of sustainability. When students and their parents ask questions about sustainability, staff can point to specific institutional efforts to reduce physical and energy waste. Therefore, some of the most important work that UW does about sustainability is making sustainability explicit.

Acknowledgments The authors would like to thank the original researchers and team members who contributed their expertise and time to develop and collect the survey data. We would like to thank UW’s Environmental Stewardship Advisory Committee for sharing the data with us. Finally, we would like to thank professor Nancy Lee at the University of Washington’s Dan Evans School of Public Affairs for providing insight about our study results and implications. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.resconrec.2018.08. 024. References Ajzen, I., 1991. The theory of planned behavior. Organ. Behav. Hum. Decis. Process. 50 179–121. Barlett, P.F., Chase, G.W., 2004. Sustainability on Campus: Stories and Strategies for Change. The MIT Press, Cambridge, MA. Barnes, N.G., Lescault, A.M., 2011. Social Media Adoption Soars as Higher-Ed Experiments and Reevaluates its use of New Communications Tools. Center for Marketing Research. Barry, B., Fulmer, I.S., 2004. The medium and the message: the adaptive use of communication media in dyadic influence. Acad. Manag. Rev. 29 (2), 272–292. Boyes, E., Skamp, K., Stanisstreet, M., 2008. Australian secondary students’ views about global warming: beliefs about actions, and willingness to act. Res. Sci. Educ. 39 (5), 661–680. Brinkerhoff, R.O., Montesino, M.U., 1995. Partnerships for training transfer: lessons from a corporate study. Hum. Resour. Dev. Q. 6 (3), 263–274. Browne, M.W., Cudeck, R., 1992. Alternative ways of assessing model fit. Sociological Methods Res. 21 (2), 230–258. Carpenter, J.P., Krutka, D.G., 2014. How and why educators use twitter: a survey of the field. J. Res. Technol. Educ. 46 (4). Cattell, R.B., 2010. The scree test for the number of factors. Multivariate Behav. Res. 1 (2), 245–276. Chankrajang, T., Muttarak, R., 2016. Green returns to education: does schooling contribute to pro-environmental behaviours? evidence from Thailand. Ecol. Econ. 131 (2017), 434–448. Clarke, N., 2001. The impact of in‐service training within social services. Br. J. Soc. Work 31 (5), 757–774. Corner, A., Randall, A., 2011. Selling climate change? The limitations of social marketing as a strategy for climate change public engagement. Glob. Environ. Chang. Part A 21 (3), 1005–1014. Cover, T.M., Thomas, J.A., 2012. Elements of Information Theory. John Wiley & Sons. D’Urso, S.C., Rains, S.A., 2008. Examining the scope of channel expansion: a test of channel expansion theory with new and traditional communication media. Manag. Commun. Q. 21 (4), 486–507. Dade, A., Hassenzahl, D.M., 2013. Cummunicating sustainability: a content analysis of website communications in the United States. Int. J. Sustain. High. Educ. 14 (3), 254–263. De Vreede, C., Warner, A., Pitter, R., 2014. Facilitating youth to take sustainability actions: the potential of peer education. J. Environ. Educ. 45, 37–56. Djordjevic, A., Cotton, D.R.E., 2011. Communicating the sustainability message in higher education institutions. Int. J. Sustain. High. Educ. 12 (4), 381–394. Eklund, D., 2015. University of Washington Receives Gold Sustainability Rating From STARS. From. http://www.washington.edu/news/2015/11/25/university-ofwashington-receives-gold-sustainability-rating-from-stars/. Fell, D., Austin, A., Kivinen, E., Wilkins, C., 2009. The Diffusion of Environmental Behaviours; The Role of Influential Individuals in Social Networks. London. . Filho, W.L., Brandli, L., 2016. Engaging Stakeholders in Education for Sustainable Development at University Level. Cham, Switzerland. .

5. Conclusion According to Worthington and Whittaker (2006), regardless of how effectively researchers believe they can determine the theorized latent variables (i.e., factor structure), an empirical (i.e., data-driven) appraisal of the underlying factor structure (i.e., EFA) should be performed initially, as was done in this study. In particular, this study of 12,000 students, staff, and faculty members indicated that three major types of sources cluster together: news sources, blogs and social media sources, and local sources of information. All sources were effective in increasing awareness. Findings also indicate that while all 375

Resources, Conservation & Recycling 139 (2018) 366–376

A.A. Kim et al.

Franz-Balsen, A., Heinrichs, H., 2007. Managing sustainability communication on campus: experiences from Lüneburg. Int. J. Sustain. High. Educ. 8, 431–445. Futerra, 2010. New Rules: New Game. Futerra Sustainability Communications. Geller, E.S., 1981. Evaluating energy conservation programs: is verbal report enough? J. Consum. Res. 8 (3), 331–335. Gikas, J., Grant, M.M., 2013. Mobile computing devices in higher education: student perspectives on learning with cellphones, smartphones & social media. Internet High. Educ. 19, 18–26. Gill, D.L., Dickinson, S.J., Scharl, A., 2008. Communicating sustainablity: a web content analysis of North American, Asian and European firms. J. Commun. Manag. 12, 243–262. Godfrey, D.M., Feng, P., 2017. Communicating sustainability: student perceptions of a behavior change campaign. Int. J. Sustain. High. Educ. 18 (1), 2–22. Hair, J.F., Black, W.C., Babin, B.J., Anderson, R.E., Tatham, R.L., 1998. Multivariate Data Analysis. Prentice hall, Upper Saddle River, NJ. Haythornthwaite, C., 1996. Social network analysis: an approach and technique for the study of information exchange. Libr. Inf. Sci. Res. 18 (4), 323–342. Hinkin, T.R., 1995. A review of scale development practices in the study of organizations. J. Manage. 21 (5), 967–988. Ho, R., 2006. Handbook of Univariate and Multivariate Data Analysis and Interpretation with SPSS. Chapman & Hall/CRC. Hogg, M., Shah, H., 2010. The Impact of Global Learning on Public Attitudes and Behaviours Towards International Development and Sustainability. DEA. Holbrook, A.L., Krosnick, J.A., Pfent, A., 2007. The causes and consequences of response rates in surveys by the news media and government contractor survey research firms. Adv. Telephone Surv. Methodol. 499–528. Hope, A., Schwaba, T., Piper, A.M., 2014. Understanding digital and material social communications for older adults. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems. Horhota, M., Asman, J., Stratton, J.P., Halfacre, A.C., 2014. Identifying behavioral barriers to campus sustainability. Int. J. Sustain. High. Educ. 15, 343–358. Hu, L.T., Bentler, P.M., 2009. Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Struct. Equation Model.: Multidiscip. J. 6 (1), 1–55. Jiang, Y., Dong, H., Yuan, S., 2015. Developing a Framework for Effective Communication with Older People. 9193. pp. 61–72. Kagawa, F., 2007. Dissonance in students’ perceptions of sustainable development and sustainability: implications for curriculum change. Int. J. Sustain. High. Educ. 8 (3), 317–338. Kaiser, H.F., 1960. The application of electronic computers to factor analysis. Educ. Psychol. Meas. 20 (1), 141–151. Kaplan, S., 2002. New ways to promote proenvironmental behavior: human nature and environmentally responsible behavior. J. Soc. Issues 56 (3), 491–508. Kim, J.O., Mueller, C.W., 1978. Factor Analysis: Statistical Methods and Practical Issues. Sage. Kim, B., Kim, Y., 2017. College students’ social media use and communication network heterogeneity: implications for social capital and subjective well-being. Comput. Hum. Behav. 73, 620–628. Kimmons, R., Veletsianos, G., Woodward, S., 2017. Institutional uses of twitter in U.S. higher education. Innov. High. Educ. 42, 97–111. Kollmuss, A., Agyeman, J., 2002. Mind the Gap : why do people act environmentally and what are the barriers to pro-environmental behavior? Environ. Educ. Res. 8 (3), 239–260. König, A., 2013. Regenerative Sustainable Development of Universities and Cities: The Role of Living Laboratories. Edward Elgar Pub. Krasny, M.E., Delia, J., 2015. Natural area stewardship as part of campus sustainability. J. Clean. Prod. 106, 87–96. Lavelle, M.J., Rau, H., Fahy, F., 2015. Different shades of green? Unpacking habitual and occasional pro-environmental behavior. Glob. Environ. Change 35 (2015), 368–378. Lee, E.-J., Lee, J., Schumann, D.W., 2002. The influence of communication source and mode on consumer adoption of technological innovations. J. Consum. Aff. 36 (1). Leonardi, P.M., Neeley, T.B., Gerber, E.M., 2012. How managers use multiple media: discrepant events, power, and timing in redundant communication. Organ. Sci. 23 (1), 98–117. Lertpratchya, A.P., Besley, J.C., Zwickle, A., Takahashi, B., Whitley, C.T., 2017. Assessing the role of college as a sustainability communication channel. Int. J. Sustain. High. Educ. 18 (7), 1060–1075. Lindner, J.R., Murphy, T.H., Briers, G.E., 2001. Handling nonresponse in social science research. J. Agric. Educ. 42 (4), 43–53. Lord, K.R., 1994. Motivating recycling behavior: a quasiexperimental investigation of message and source strategies. Psychol. Mark. 11 (4). Lukman, R., Lozano, R., Vamberger, T., Krajnc, M., 2013. Addressing the attitudinal gap towards improving the environment: a case study from a primary school in Slovenia. J. Clean. Prod. 48, 93–100. Mackenzie, S., Ryan, C., Litfin, K., 2015. Sustainability Across the Curriculum. University of Washington. Marquart-Pyatt, S.T., Shwom, R.L., Dietz, T., Dunlap, R.E., Kaplowitz, S.A., McCright, A.M., Zahran, S., 2011. Understanding public opinion on climate change: a call for research. Environ. Sci. Policy Sustain. Dev. 53 (4), 38–42. Mazo, L., Macpherson, I., 2017. A strategic communication model for sustainable initiatives in higher education institutions. Athens J. Mass Media Commun. 3 (4), 321–342. McCright, A.M., Dunlap, R.E., 2011. The politicization of climate change and polarization in the American public’s views of global warming, 2001–2010. J. Midwest Sociol. Soc. 52, 155–194. Meyer, A., 2015. Does education increase pro-environmental behavior? Evidence from Europe. Ecol. Econ. 116 (2015), 108–121.

Mezirow, J., 1991. Transformative Dimensions of Adult Learning. Jossey-Bass. Morsing, M., Schultz, M., 2006. Corporate social responsibility communication: stakeholder information, response and involvement strategies. Bus. Ethics 15 (4). Moswer, S.C., 2010. Communicating climate change: history, challenges, process and future directions. WIREs Clim. Change 1 (1), 31–53. Newig, J., Fritsch, O., 2009. Environmental governance: participatory, multi‐level–and effective? Environ. Policy Gov. 19 (3), 197–214. Norazah, M.S., Norbayah, M.S., 2016. In: Walter Leal Filho, L.B. (Ed.), Campus Sustainability: Does Student Engagement with Eco-Campus Environmental Activities and Green Initiatives Really Matter? Engaging Stakeholders in Education for Sustainable Development at University Level. Springer. Nunnally, J.C., Bernstein, I.H., 1994. Psychometric Theory. McGraw-Hill, New York, NY. Ockwell, D., Whitmarsh, L., O’Neill, S., 2009. Reorienting climate change communication for effective mitigation: forcing people to be green or fostering grass-Roots engagement? Sci. Commun. 30 (3). Olli, E., Grendstad, G., Wollebaek, D., 2001. Correlates of environmental behaviors. Bringing back social context. Environ. Behav. 33 (2), 181–208. Olson, K., 2006. Survey participation, nonresponse bias, measurement error bias, and total bias. Int. J. Public Opin. Q. 70 (5), 737–758. Orr, D.W., 2002. The Nature of Design: Ecology, Culture, and Human Intention. Oxford University Press, Oxford, England ; New York, New York. Owens, K.A., Halfacre-Hitchcock, A., 2006. As green as we think? The case of the College of Charleston green building initiative. Int. J. Sustain. High. Educ. 7 (2), 114–128. Pett, M.A., Lackey, N.R., Sullivan, J.J., 2003. Making Sense of Factor Analysis: The Use of Factor Analysis for Instrument Development in Health Care Research. SAGE Publications. Podsakoff, P.M., MacKenzie, S.B., Lee, J.-Y., Podsakoff, N.P., 2003. Common method biases in behavioral research: a critical review of the literature and recommended remedies. J. Appl. Psychol. 88 (5), 0. Pope, D., Stanistreet, D., 2017. Quantitative Methods for Health Research: A Practical Interactive Guide to Epidemiology and Statistics. John Wiley & Sons. Princeton, 2017. Green College Honor Roll. From. https://www.princetonreview.com/ college-rankings/green-guide/green-honor-roll. Rettie, R., Burchell, K., Riley, D., 2012. Normalising green behaviours: a new approach to sustainability marketing. J. Mark. Manag. 28 (3–4), 420–444. Rice, R.E., 1993. Media appropriateness. Hum. Commun. Res. 19 (4), 451–484. Richardson, G.R.A., Lynes, J.K., 2007. Institutional motivations and barriers to the construction of green buildings on campus. Int. J. Sustain. High. Educ. 8 (3), 339–354. Roorda, N., 2004. Developing Sustainability in Higher Education Using Aishe. Higher Education and the Challenge of Sustainability. Schumacker, R.E., Lomax, R.G., 2004. A Beginner’s Guide to Structural Equation Modeling. Lawrence Erlbaum Associates, New Jersey. Sharp, L., 2002. Green campuses: the road from little victories to systemic transformation. Int. J. Sustain. High. Educ. 3 (2), 128–145. Siano, A., Conte, F., Amabile, S., Vollero, A., Piciocchi, P., 2016. Communicating sustainability: an operational model for evaluating corporate websites. Sustainability 8 (950). Sierra, 2015. Sierra’s 10th Annual Cool Schools Rankings. Sierra Magazine. Stephens, K.K., Sørnes, J.O., Rice, R.E., Browning, L.D., Sætre, A.S., 2008. Discrete, sequential, and follow-up use of information and communication technology by experienced ICT users. Manag. Commun. Q. 22 (2), 197–231. Stephens, K.K., Barrett, A.K., Mahometa, M.J., 2013. Organizational communication in emergencies: using multiple channels and sources to combat noise and capture attention. Hum. Commun. Res. 39 (2), 230–251. Stern, P.C., 2000. New environmental theories: toward a coherent theory of environmentally significant behavior. J. Soc. Issues 56 (3), 407–424. Sundar, S.S., Nass, C., 2001. Conceptualizing sources in online news. J. Commun. 51 (1), 52–72. Swaim, J.A., Napshin, S.A., Maloni, M.J., Henle, A.B., 2014. Influences on student intention and behavior toward environmental sustainability. J. Bus. Ethics 124, 465–484. Tourangeau, R., Groves, R.M., C. D. R. P. O. Q, 2010. Sensitive topics and reluctant respondents: demonstrating a link between nonresponse bias and measurement error. Public Opin. Q. 74 (3), 413–432. Tur, G., Marín, V.I., Carpenter, J., 2017. Using twitter in higher education in Spain and the USA. Media Educ. Res. J. 25 (51), 19–27. UNESCO, 1997. Educating for a Sustainable Future: A Transdisciplinary Vision for Concerted Action. Vining, J., Ebreo, A., 2002. Emerging theoretical and methodological perspectives on conservation behavior. New Handbook of Environmental Psychology. Vrocharidou, A., Efthymiou, I., 2012. Computer mediated communication for social and academic purposes: profiles of use and university students’ gratifications. Comput. Educ. 58 (1), 609–616. Wolf, J., Moser, S.C., 2011. Individual understandings, perceptions, and engagement with climate change: insights from in-depth studies across the world. WIREs Clim. Change 2 (4), 547–569. Worthington, R.L., Whittaker, T.A., 2006. Scale development research: a content analysis and recommendations for best practices. Couns. Psychol. 34 (6), 806–838. Wright, T.S.A., 2002. Definitions and frameworks for environmental sustainability in higher education. Int. J. Sustain. High. Educ. 3 (3), 203–220. Zolnierek, K.B.H., DiMatteo, M.R., 2009. Physician communication and patient adherence to treatment: a meta-analysis. Med. Care 47 (8), 826. Zsóka, Á., Szerényi, Z.M., Széchy, A., Kocsis, T., 2012. Greening due to environmental education? Environmental knowledge, attitudes, consumer behavior and everyday pro-environmental activities of Hungarian high school and university students. J. Clean. Prod. 48 (2013), 126–138.

376