Energy Research & Social Science: Kimberly S. Wolske, Paul C. Stern, Thomas Dietz

Energy Research & Social Science 25 (2017) 134–151
Contents lists available at ScienceDirect
Energy Research & Social Science

journal homepage: www.elsevier.com/locate/erss
Original research article
Explaining interest in adopting residential solar photovoltaic systems

in the United States: Toward an integration of behavioral theories
Kimberly S. Wolske a,∗ , Paul C. Stern b,c,e , Thomas Dietz d,e
a
Erb Institute for Global Sustainable Enterprise, University of Michigan, Ann Arbor, MI, 48109-1234, USA
b
National Academies of Sciences, Engineering, and Medicine, Washington DC 20001, USA
c
Norwegian University of Science and Technology, Trondheim, Norway
d
Department of Sociology and Environmental Science and Policy Program, Michigan State University, East Lansing, MI, 48864, USA
e
Social and Environmental Research Institute, P.O. Box 1487, Northampton, MA 01060, USA
a r t i c l e i n f o a b s t r a c t
Article history: Increased household adoption of solar photovoltaic systems has the potential to reduce greenhouse
Received 22 June 2016 gas emissions associated with providing electricity. Although residential solar has recently become more
Received in revised form affordable, market penetration in the U.S. remains relatively low. This study proposes a theoretical frame-
30 December 2016
work for investigating the psychological and social determinants of interest in residential solar drawn
Accepted 30 December 2016
from three theories that may explain the decision to pursue it: diffusion of innovations theory, theory
Available online 25 January 2017
of planned behavior, and value-belief-norm theory. We test this framework using survey data from 904
non- adopter homeowners, with the aim of identifying potential levers for intervention. Overall, we find
Keywords:
Value-belief- norm theory
that consumers see solar electricity in multiple ways: as an environmental benefit, a consumer good, and
Theory of planned behavior an innovative technology. Notably, individuals who trust installers and believe solar will be personally
Diffusion of innovations beneficial are more likely to consider contacting an installer, as are individuals drawn to novel products.
Residential solar photovoltaics Proenvironmental personal norms indirectly increase interest through perceived personal benefits, sug-
Renewable energy gesting that marketing efforts aimed at environmentally-concerned individuals may need to emphasize
Environmental decision making non-environmental benefits. The results also support leveraging trusted social networks to convey the
benefits of solar. We conclude by discussing the value of the integrated framework along with implications
for policymakers and marketers.
© 2017 Elsevier Ltd. All rights reserved.
1. Introduction in the U.S. has grown from 380 megawatts (MW) to 5644 MW in
2015 [5]. Several technological, economic, and policy factors have
In many parts of the world, including substantial portions made possible the recent increase in adoption rates. These include
of the U.S.A., solar photovoltaics have great potential to reduce more reliable and efficient technologies, lower production costs,
greenhouse gas emissions and other adverse impacts from current and more favorable laws and financial incentives. Notably, the fed-
methods of generating electricity [1–3]. Achieving this potential eral solar investment tax credit (ITC), state and utility rebates, and
will require behavioral changes on the part of producers, suppli- net metering policies have significantly lowered the installed cost
ers, and potential adopters of this technology. This study considers of RPV in many states [6]. In addition, new third-party ownership
the factors that may lead potential adopters of residential photo- models have made RPV affordable to a larger portion of U.S. house-
voltaics (RPV) to take a key first step toward adoption: showing holds [7,8]. Under such arrangements, RPV systems are installed on
interest in contacting an RPV installer. a homeowner’s property but owned and maintained by a third party
Although RPV have been available for household use in the U.S. company – often with no upfront costs. Homeowners can lease the
since the late 1970s, their deployment proceeded at a slow rate until system, paying a subsidized upfront cost or a monthly fee, or enter
recently [4]. Since 2010, the cumulative installed capacity of RPV into a power purchase agreement (PPA) to buy the electricity gen-
erated from the system at a set per-kilowatt-hour rate. Both lease
fees and PPAs are designed to be competitive with electricity rates
∗ Corresponding author. Present address: Harris School of Public Policy, University charged by the utility.
of Chicago, Chicago, IL, 60637, USA. Despite these advances in the U.S. RPV market, there remain
E-mail addresses: wolske@uchicago.edu (K.S. Wolske), pstern@nas.edu opportunities for improvement. The U.S. Department of Energy
(P.C. Stern), tdietz@msu.edu (T. Dietz).
http://dx.doi.org/10.1016/j.erss.2016.12.023
2214-6296/© 2017 Elsevier Ltd. All rights reserved.
K.S. Wolske et al. / Energy Research & Social Science 25 (2017) 134–151 135
projects that if the price of solar technologies were to fall by 75% of may be possible levers for enhancing adoption, as well as to con-
2010 prices by 2020, solar could meet 27% of U.S. electricity needs sider whether a full understanding must go beyond these individual
by 2050 [2]. In the context of residential solar, this means reducing theories. The parable of the blind men and the elephant reflects
the installed cost of RPV to $1.50 per peak watt. Currently 11.6% of our view of the adoption phenomenon. The parable is attributed
the installed cost of RPV (or $0.36 of $3.09 per watt) is attributed to to the Buddha as an admonition against sectarian division. Several
the “soft costs” of acquiring new customers [9]. For policymakers blind observers offer their assessment of an elephant. The observer
and PV companies looking to spur further adoption of RPV, under- who touches the leg declares the elephant to be rather like a tree
standing potential leverage points for increasing interest in RPV – trunk. The observer who touches the tail declares the elephant to
and reducing soft costs – is critical. be rather like a broom, the trunk leads to the conclusion of a snake
To date, research on U.S. RPV adoption has primarily involved and so on. Of course, the methods of science urge us to share obser-
statistically modeling adoption trends and future penetration rates vations, and after some conversation it is likely that the observers
[7,10–14]. Only recently has attention turned toward understand- would come to an accurate description of the elephant. So, too,
ing the characteristics of consumers who install RPV and the with the multiple theories that describe aspects of the decision to
psychological and social factors that affect their decision-making. go solar: each describes an aspect of reality. As Klöckner [27] noted,
Most work in this area has relied on retrospective surveys of RPV integrating variables from the most successful theories of behavior
adopters. Researchers have examined, for example, self- reported change may yield both theoretical and practical benefits: we may
motivations for “going solar,” the role of expected financial returns, better understand which variables are key determinants of behav-
the effects of seeing neighbors with RPV installations, and the influ- ior, which should be targeted for intervention, and which of the
ence of learning about RPV through different information sources more distal variables may be useful for influencing more proximal
such as installers or existing owners [8,15–18]. ones.
While this body of work sheds light on different aspects of RPV In this case, we are looking at three aspects of the “elephant” of
decision making, it lacks theoretical coherence [19]. Some studies RPV. Since it is an innovative technology, we have drawn on ele-
reference Rogers’ [20] diffusion of innovations theory to explain the ments of DOI theory. Since it is a consumer behavior, we have drawn
influence of information channels and peer effects on RPV adoption on elements of TPB. And since it is a proenvironmental behavior, we
[15,17], others evaluate the extent to which adopters engaged in have drawn on elements of VBN theory. We emphasize that given
rational economic evaluations of RPV [8] , while still others examine the practical limits of survey length, we have not conducted a com-
determinants of adoption in the absence of any theoretical frame- prehensive test of any of these theories. To continue the analogy,
work [16]. To our knowledge, only one study, based on qualitative there are likely aspects of the elephant we have not considered.
interviews, attempts to look across multiple behavioral theories to Those could be added to an integrated model in future research, as
explain RPV adoption [18]. This failure to build on known behav- could other aspects of our three theories that we have not included
ioral theories and concepts from the noneconomic social sciences in this study. At some point, it would be useful to have an over-
that might have predictive value is characteristic generally of the arching comprehensive theory, a full picture of the elephant, that
literature on household energy investments [21], and does not bode integrates across existing partial theories. We are taking a first step
well for the accumulation of knowledge. Furthermore, the existing in that direction but certainly do not offer such a full integration.
literature, with its emphasis on examining the decisions of adopters Rather than an integrative theory or a comprehensive compar-
post hoc, may overlook crucial information about the motivations ative test of the theories, the aim of this paper is to offer a better
or barriers affecting non- adopters. understanding of what factors influence interest in contacting an
In this study, we take a step toward building a theoretically RPV installer, a key precursor to actual adoption. We examine this
grounded account of RPV adoption by examining interest in RPV question using survey data collected from non-adopter homeown-
– among non-adopters – as a function of numerous potential influ- ers in four U.S. states that have relatively high rates of RPV adoption
ences, including not only aspects of the diffusion of innovations (AZ, CA, NJ and NY). Measures on the survey were informed by each
theory but also elements of two other well-known behavioral the- of the three theories as well as conversations with leaders in the
ories that can reasonably be expected to have predictive value. As RPV industry who suggested measures that, in their experience,
RPV is the purchase of a consumer good, albeit an unusual one, influenced the decision to move toward solar.
theories that have proven helpful in understanding consumer deci- In the remainder of this section, we discuss how each of the
sion making could be useful. In particular, the theory of planned three theories – DOI, TPB, and VBN – may help explain interest
behavior (TPB) [22] has been deployed in analyzing a wide vari- in RPV and provide evidence from relevant past literature. Section
ety of consumer behaviors. Because RPV has substantial benefits 2 describes our survey measures and methodology for testing the
in reducing greenhouse gas emissions, homeowners may also be framework. Section 3 presents results, and Section 4 discusses their
motivated in part by environmental concerns. Thus we consider implications.
measures drawn from a well-developed theory of individual envi-
ronmental decision making, the value-belief-norm theory (VBN) 1.1. Three theoretical perspectives on interest in RPV
[23]. Extensive literatures have used either TPB or VBN to under-
stand proenvironmental consumer behavior [21,24–26]. However, 1.1.1. Diffusion of innovations: RPV as an innovation
as adoption of RPV involves a larger financial commitment and a In most of the United States, RPV will be seen by homeowners
more visible change in household technology than perhaps any as an innovative and relatively unknown, new technology. Even in
other consumer behavior, it is not clear the extent to which results states such as California where climatic and policy conditions for
from studies of less expensive and less visible behaviors will gen- RPV are favorable, only 5.7% of homeowners have adopted them
eralize. Indeed, one contribution of this analysis is to examine [28,29]. Diffusion of innovations (DOI) theory describes the process
whether aspects of existing theories of proenvironmental con- by which an innovation diffuses through a social system as a result
sumer behavior are valuable in explaining such a substantial step. of information being communicated through media and person-
We view these theories – DOI, TPB, and VBN – as complementary to-person communication channels (see Fig. 1a) [20]. The theory
rather than competitive; each considers unique aspects of decision proposes that the process for adopting an innovation occurs in five
making. Accordingly, we propose and test an integrated model of stages: individuals become aware of an innovation (knowledge),
interest in RPV that includes and combines aspects of all three the- form attitudes about it (persuasion), decide whether to adopt it
ories to identify factors that are particularly influential and that (decision), implement the innovation (implementation), and then
136 K.S. Wolske et al. / Energy Research & Social Science 25 (2017) 134–151
Fig. 1. Simple schematic representations of DOI, TPB, and VBN theories.
decide whether to continue using it (confirmation). The speed at The rate at which an innovation is adopted is also influenced by
which this process occurs is thought to be influenced by several how it is perceived. During the persuasion stage of diffusion, indi-
factors. One concerns the innovativeness of potential adopters. viduals evaluate the advantages and disadvantages of an innovation
Individuals who by disposition are more innovative or novelty- relative to their personal situation. Five attributes are thought to
seeking have a higher propensity to adopt early on. Accordingly, lead to favorable attitudes and, thus, faster adoption. First, the inno-
Rogers describes individuals who are earliest to adopt as “innova- vation must be perceived to have a relative advantage over the
tors,” followed by “early adopters,” the “early majority,” the “late status quo. Potential adopters must believe RPV will offer advan-
majority,” and “laggards.” Later adopters are thought to adopt pri- tages (financial, environmental, social, or otherwise) over relying
marily as a result of normative influences whereas earlier adopters solely on electricity generated by their utilities. An innovation is
tend to actively seek information about the technology and weigh also more likely to be adopted to the extent that it is perceived
its benefits. to be compatible with existing values, needs and practices and,
similarly, not too complex. In other words, RPV may be more readily expectancy-value calculation. Attitudes form from beliefs about the
adopted if they do not require significant shifts in routines or, like- consequences of a behavior and the expected likelihood of those
wise, significant effort to learn. Innovations are also easier to adopt outcomes occurring. A number of beliefs might influence attitudes
if they are “trialable,” that is, if they can be tested for a limited toward RPV. These could include perceptions of benefits like pro-
time or in an incremental fashion, without requiring full commit- tection from rising electricity costs, improved home value, and
ment. The low trialability of solar panels may explain, in part, why reduced environmental pollution, as well as perceived drawbacks
they have been slow to gain wider popularity. Finally, people are such as concerns about maintenance costs, unreliability, and the
more likely to adopt innovations that have observable results. In risks of altering one’s house. The degree to which someone feels
this regard, the ability to see rooftop solar panels on others’ homes social pressure to perform a behavior depends on whether impor-
may enhance their adoption. tant referent groups are thought to approve or disapprove of the
A key aspect of DOI concerns the communication channels behavior. In the context of RPV, feeling that friends and neigh-
through which individuals learn about an innovation. Mass media bors support the adoption of solar may partially counterbalance
communications – such as advertising from RPV installers or solar concerns about the risks of adopting a relatively unknown tech-
advocacy groups – are hypothesized to be most effective at the nology. Conversely, feeling that peers would frown upon RPV may
knowledge stage of the diffusion process when individuals are first hinder individuals from pursuing it, even if they otherwise have
learning about an innovation [20]. During the persuasion stage favorable attitudes. Finally, even when individuals have positive
when individuals begin to evaluate the specific advantages and dis- attitudes toward a behavior and the support of their social net-
advantages of an innovation, communications with peers who have works, the decision to engage in a behavior may hinge on whether
already adopted are expected to have significant weight – espe- they feel able to perform it. Factors that might lead individuals to
cially if the innovation’s performance is perceived to be uncertain. believe they cannot adopt RPV include local climate, exposure to
Conversations with RPV installers about installation options and sunlight, and expectations about having to move residences in the
system economics may also be influential at this stage. near future.
Several studies have examined solar technologies through the Research by Korcaj et al. [36] provides evidence that TPB can be
lens of DOI. Early research by Labay and Kinnear [30] found that useful for explaining interest in RPV. Among German homeowners,
compared to the general population, early adopters of solar ther- intentions to purchase RPV were significantly predicted by posi-
mal systems were more likely to see solar thermal as less financially tive attitudes, subjective norms, and perceived behavioral control.
and socially risky, less complex, and more compatible with their Positive attitudes, in turn, were strongly influenced by beliefs that
personal values. More recent work in the U.K. suggests that com- RPV would lead to enhanced social status, energy independence,
pared to early adopters, potential “early majority” adopters of RPV and financial gain. Other research by Claudy et al. [37] confirms
need convincing that RPV are attractive, affordable, and hassle-free that perceptions of RPV’s advantages and disadvantages are impor-
[31]. In the U.S., a number of studies have investigated the influ- tant factors in the decision-making process. Though their study
ence of peer effects and observability on RPV adoption rates – but did not specifically test TPB, positive reasons for considering RPV,
not other attributes of the innovation such as relative advantage, such as its economic and environmental benefits were found to
compatibility, etc. Research in California found that the more com- lead to more positive attitudes toward RPV, which, in turn, influ-
mon RPV are in a zip code, the greater the likelihood of another enced intention to purchase. Negative attributes of RPV, such as its
household adopting [10], while work in Texas suggests that the perceived expense and maintenance requirements, were found to
decision-making process for considering solar is shortened when directly reduce intention to buy.
homeowners either see solar panels in their neighborhood or talk
to a neighbor who has solar [17]. Other studies have investigated 1.1.3. Value-belief-norm theory: RPV as green technology
the influence of different information channels. Evidence from Cal- RPV reduce greenhouse gas emissions and other types of pollu-
ifornia suggests that the perceived importance of an information tants by displacing fossil-fuel generated electricity. To the extent
source depends on how individuals first become interested in RPV that RPV are perceived as “green,” the motivation to pursue it
[16]. Among those who received direct mail marketing, interac- should be explained by theories of proenvironmental behavior. We
tions with the installation company were perceived to have the make use of the value-belief-norm theory to capture this aspect
greatest import; among those who became interested after talk- of decisions about RPV [23,25], since as Ajzen noted, values nicely
ing to a neighbor or seeing a neighbor installer a system, further complement TPB [38].
conversations with neighbors were perceived to be very important. VBN theory argues that the root cause of proenvironmental
behavior lies in values and emphasizes the importance of altruism
1.1.2. Theory of planned behavior: RPV as a consumer durable directed at other humans (social altruism) and altruism directed at
In addition to being novel, RPV are relatively expensive con- other species and the biosphere (biospheric altruism). In the case of
sumer durable goods that require making rather significant RPV, both of these factors might matter, since reduced stress on the
alterations to customers’ homes. In this regard, the decision to environment benefits the biosphere as well as other humans. The
pursue RPV may be explained, in part, by how one weighs the other values referenced in VBN could also influence interest in RPV.
costs and benefits and the degree to which RPV seem advantageous Self-interest may play a role either by reducing environmental con-
given one’s home and finances. Here the theory of planned behav- cern and thus interest in RPV, or increasing interest because of the
ior (TPB) offers a useful analytic framework. TPB is one of the best perceived economic benefits of RPV. Openness to change and tra-
developed approaches within social psychology, has been applied ditional values could also matter. As a new technology, RPV might
extensively to proenvironmental behaviors [e.g., Refs. [32–35]], and appeal to those open to change but be viewed with skepticism by
easily accommodates a variety of factors that may matter for adop- those who value the status quo.
tion of RPV. Congruent with past research on VBN, we expect the effects of
TPB posits that the intention to perform a behavior is the values to be mainly indirect, mediated by other variables more
outcome of a rational decision making process that involves con- causally proximate to the decision to pursue RPV. The full VBN
sidering (1) one’s attitudes toward the behavior, (2) perceived model (Fig. 1c) posits a causal chain in which values shape gen-
social pressure to do the behavior (subjective norms), and (3) an eral beliefs about human-environmental relationships (ecological
assessment of one’s ability to perform it (perceived behavioral worldview), which in turn influence beliefs about the impact of
control) (see Fig. 1b) [22] . Each of these is the outcome of an environmental problems on the things one values (awareness of
consequences). These beliefs then affect the sense of responsibility both leasing and purchasing options [9], yet the states are diverse
one feels to address those consequences (ascription of responsibil- in terms of climate, insolation, and available financial incentives.
ity) and one’s sense of moral obligation to act (personal norms). In Quotas were used for age and education level to match homeowner
the context of pursuing RPV, a renewable energy technology, we census distributions in each of the four states. A number of respon-
might expect the effects of values to be mediated by beliefs about dents were excluded from the dataset, either because they failed an
climate change, and a sense of obligation to address it. attention check question (N = 259), selected the same Likert scale
Although no past research, to our knowledge, has explicitly response (i.e., “straight-lined”) for six or more banks of items on
used VBN to explain RPV adoption, several studies provide evi- the survey (N = 67) or had missing data on one or more variables
dence that environmental motives and certain value orientations (N = 41). Another 144 respondents were removed because they had
may be an important driver of interest in RPV. RPV adopters in already spoken to a solar company about the possibility of installing
Austria [39] and the Netherlands [40] self-reported that helping PV and therefore may be at a different stage of the decision process.
the environment was a top motivation for buying solar. Respon- In total, 1156 surveys were included in our initial analyses. Fol-
dents in Austria additionally indicated that they felt a responsibility lowing factor analysis, an additional 252 individuals were dropped
to future generations, suggesting that social altruism was a factor from the dataset because they had marked “Don’t Know” for all
in their decision-making. In both studies, participants also wanted items in one or more factors. The final dataset (N = 904) had 250
to take advantage of limited-time government incentives, so self- respondents in Arizona, 220 in California, 205 in New Jersey and
interest may be in play as well. In a small qualitative study of early 229 in New York.
adopters in Wisconsin, nearly two-thirds of participants mentioned
environmental concern among their reasons for getting RPV; none 2.2. Measures
of the participants, however, cited environmental considerations
as the exclusive reason for going solar [18] . In developing the survey, we were guided both by potentially
Other research has looked at the role of environmental motives relevant social science theories and by the practical experience
in pursuing sustainable eco-innovations such as electric cars and of members of the research team who have worked within the
renewable energy more broadly. Jansson et al. [26,41] used VBN to RPV industry in the USA. Thus, we included both theory-based and
examine willingness to buy an environmental friendly car. Consis- practice-based survey items. We used our understandings of the-
tent with past research, values, beliefs, and norms each predicted ory in an attempt to move toward an analysis that was coherent,
willingness to buy, with personal norms having the strongest direct reflective of theory to the extent possible, and that also spoke to
effect. In addition, existing owners of alternative fuel vehicles practical concerns.
exhibited higher levels of environmental values, beliefs and per- The survey instrument included questions intended to measure
sonal norms than non-adopters. Findings by Noppers et al. [42] variables from each of the three theories that might contribute
suggest that perceived environmental attributes are a key driver directly or indirectly to interest in talking to an RPV installer
of sustainable innovation adoption, above and beyond beliefs that (see Appendix A). As we sought to understand how DOI, TPB,
adoption will have instrumental value or increase social status. and VBN might complement each other in explaining this inter-
Similarly, environmental concern was found to be the strongest est, we focused on the variables in each theory that seemed most
predictor of interest in renewable micro-generation technologies relevant and practically useful. For some variables, validated mea-
among Dutch households [43]. sures already existed; for others, especially those related to specific
beliefs and concerns about RPV, such measures had to be developed.
Unless otherwise noted, survey items presented 5-point Likert
1.2. Research objectives
scales from 1 = Strongly disagree to 5 = Strongly agree, with an addi-
tional option to mark Don’t Know. We used principal components
The primary objective of the present paper is to elucidate pos-
analysis to organize large numbers of questions into coherent and
sible leverage points for increasing interest in RPV. We draw on
interpretable scales. Scale scores were computed by taking the
elements from each of the above theories – DOI, TPB, and VBN – to
average of all items that loaded on a single factor above 0.45. Indi-
explore the extent to which perceiving RPV as an innovative tech-
viduals who marked “Don’t know” for all items in a factor were
nology, a beneficial consumer good, or as a green choice predicts
excluded from subsequent analyses (N = 252).
interest in pursuing it. Our intent is neither to set the theories in
We discuss the variables under the theories that invoke them,
competition with each other nor to arrive at definitive conclusions
ordering the theories starting with the most general, VBN, and end-
about each theory’s explanatory power. Indeed, practical limits to
ing with the most specific, TPB. However, several variables can be
the number of items we can include in a survey mean that we have
invoked by more than one theory. For example, while the value
neglected some elements of each theory. Following the parable of
openness to change is explicitly referenced by VBN theory, it can
the elephant, we propose that interest in RPV is best understood
also be interpreted within DOI theory as a measure of consumer
by integrating perspectives from all three theories. As research to
innovativeness. Likewise, some variables appear in the models for
date on U.S. RPV adoption has been fairly atheoretical and non-
both TPB and DOI. We discuss such overlaps when examining the
cumulative, another objective is to provide a framework for guiding
combined model.
future work.
2.2.1. VBN variables
2. Methods 2.2.1.1. Values. The survey used standard items from the literature
to measure biospheric altruism, social altruism, self-interest, tra-
2.1. Participants ditional values, and openness to change [25,44] (see Table A1 in
Appendix A). Respondents rated the importance of each value state-
Our data come from an online survey that was distributed in ment as a guiding principle in their lives on a scale from 1 = Not at all
June 2014 to 1611 members of a paid panel managed by Qualtrics. important to 5 = Extremely important, with the additional option of
com. Eligible participants were limited to homeowners living in one marking Opposed to my values (coded as −1). Four factors emerged
of our four study states – Arizona, California, New Jersey and New from the analysis. The first, Altruism, captures items from both the
York – who did not already have solar panels. We chose these states biospheric and social altruism scales (Cronbach’s ␣ = 0.87) [see Ref.
because they are among the largest RPV markets in U.S. that offer [44]. Other items scaled as expected for Self-interest (Cronbach’s
␣ = 0.63), Traditional values (Cronbach’s ␣ = 0.62), and Openness to to measure this construct in our pilot study with items such as
Change (Cronbach’s ␣ = 0.79). “Installing solar panels is a hassle,” “There is a lot of paperwork
involved in installing solar,” and “Installing solar takes a lot of time.”
2.2.1.2. Awareness of consequences (AC). We measured AC with a However, more than a third of respondents marked “Don’t Know”
single item that assessed the degree to which respondents felt cli- on each item, resulting in an inadequate number of responses for
mate change was a serious problem for society. Given the strong factor analyses. We subsequently excluded these measures from
views on climate change in the US at present, this single item may other analyses. Many people may have so little familiarity with RPV
be sufficient to capture much of the variation from individual to that they are simply unsure about complexity.
individual.1 Trialability
DOI theory posits that new technologies are more likely to be
2.2.1.3. Personal norm to act (PN). Three items were created to adopted if potential adopters can experiment with them or try
assess personal norms about energy issues and climate change and them out incrementally. This attribute is thought to be especially
were averaged to form a Personal Norms scale (Cronbach’s ␣ = 0.83). important among early adopters who do not have the benefit of
learning from others’ experiences [20]. Rather than measuring per-
2.2.2. DOI variables ceptions of RPV’s trialability – which we would expect to be low –
2.2.2.1. Consumer innovativeness. A central assumption of DOI is we assessed the degree to which respondents expressed a desire to
that early adopters of a technology are more innovative than later try out solar using four items (Cronbach’s ␣ = 0.70).
adopters; they actively seek new information and are more com- Observability
fortable coping with uncertainty [20]. To measure innovativeness, Diffusion of an innovation is more likely to occur the more con-
we adapted two scales previously used by Jansson [45] and origi- sumers are able to observe that technology in use. Five yes/no
nally developed by Manning et al. [46] (Table A2 in Appendix A). The questions gauged whether respondents had seen solar panels
first scale, Consumer Novelty Seeking (CNS; Cronbach’s ␣ = 0.87), first-hand through their neighbors or peers. “Yes” responses were
measures the extent to which individuals seek out innovative and summed to create a score ranging from 0 to 5 (Cronbach’s ␣ = 0.71).
new products. The second scale, Consumer Independent Judgment
Making (CIJM; Cronbach’s ␣ = 0.83), measures the extent to which 2.2.2.3. Communication channels. We also examined the extent to
individuals turn to others for opinions on products before making which respondents had been exposed to commercial marketing
purchasing decisions. Past evidence suggests that CNS is most pre- about RPV and the degree to which they would trust information
dictive at early stages of the adoption process while CIJM matters from different sources. While trust is not an explicit variable in DOI,
more at later stages [26,46]. it is an implied aspect of the diffusion process.
Recent exposure to solar marketing
2.2.2.2. Perceived characteristics of the innovation. We measured Two yes/no items captured whether respondents had recently
various perceptions and beliefs about RPV that might affect levels been exposed to residential solar marketing. These items were
of interest. Factor analysis grouped these items into eight scales, summed to form a score of Solar Marketing Exposure ranging from
three of which fall under the DOI framework (see Table A2 in the 0 to 2 (Cronbach’s ␣ = 0.55).
Appendix A for the full list of items). Trust in PV industry
Relative advantage The extent to which homeowners trust information from RPV
People are more likely to adopt an innovation if they perceive providers may influence not only how receptive they are to solar
it to have advantages over currently available technologies [20]. marketing materials but also how willing they are to consider pur-
Six items on the survey measured various perceived advantages suing RPV. To measure trust in the PV industry, we averaged two
of RPV, such as whether it would save respondents money, pro- items concerning trust in solar providers and trust in solar trade
tect against rising electricity costs and blackouts, or increase their organizations (Cronbach’s ␣ = 0.88).
homes’ values. These items formed a reliable factor with Cronbach’s Trust in social network
␣ = 0.90. If both RPV and the industry promoting it are unfamiliar, con-
(In)compatibility/riskiness sumers may prefer to get information from friends and neighbors.
Technologies are more readily adopted when they are com- Two items assessed the extent to which respondents would trust
patible with existing practices, values, and needs. In pilot testing neighbors and friends to provide accurate information about RPV.
our survey we found that measures of perceived compatibility fac- These items were averaged into a single scale, Cronbach’s ␣ = 0.81.
tored with measures of personal norms. In order to have distinct
variables, we decided in the final instrument to use perceived riski- 2.2.3. TPB variables
ness as a proxy for incompatibility. Technologies that are perceived 2.2.3.1. Behavioral beliefs. Personal benefits
as incompatible because they require radical shifts from the sta- Given the potential expenses and risks involved, people may be
tus quo or have uncertain outcomes may be perceived as riskier more likely to consider RPV if they believe they stand to person-
for that reason. Several researchers have treated perceived risk ally benefit. To measure these beliefs, we used the same items that
as a decision-relevant characteristic of innovations [30,45,47]. A comprised Relative Advantage in the DOI framework.
three-item factor (Cronbach’s ␣ = 0.63) that included items such Environmental benefits
as “Installing solar panels could damage my home” reflected per- Beliefs about the efficacy of solar panels in addressing environ-
ceived risks. mental problems may also affect interest in RPV. A three-item factor
Complexity (Cronbach’s ␣ = 0.87) captured the extent to which respondents
Innovations are more likely to be adopted when they are per- thought installing solar panels would help slow climate change and
ceived to be simple to adopt and easy to understand. We attempted improve environmental quality (Table A3). We note that unlike our
other behavioral belief scales, the benefits tapped here do not flow
directly to the respondent. Thus, the position of this scale within
1
a TPB account of interest in RPV is ambiguous. These items may
As a reviewer noted, using only a single item focused on climate change is a
limited operationalization of AC. We have also estimated models not including AC
also reflect the relative advantage concept in DOI, if environmental
(results in online supplemental materials) and find essentially no change in sub- benefits are perceived as a type of advantage.
stantive results for other variables.
Perceived risks six months. Based on bivariate correlations these three items were
As beliefs about the risks of adopting RPV may affect attitudes averaged to form a scale measuring Interest in Talking to an RPV
toward it, we included the three items used to measure Riskiness in Installer (IT) (Cronbach’s ␣ = 0.81).
DOI.
Waiting for improvements
3. Results
Since the quality and cost-effectiveness of RPV seem to be con-
tinually improving, some homeowners may decide to wait until
3.1. Analytic approach
further improvements in cost and performance are available. A
two-item factor tapped this argument (Cronbach’s ␣ = 0.65).
As a preliminary step in our analysis, we conducted separate
Concerns about the costs of solar
path analyses to assess the explanatory power of the variables
For most large consumer purchases, such as automobiles and
derived from each theory. Since we did not include a comprehen-
appliances, the purchase price and operating costs are reasonably
sive set of variables from each theory, our intention was not to “test”
well known. For RPV, consumers are likely to have substantial con-
each theory, but rather to confirm that each set of variables had a
cerns about the costs of this unfamiliar technology. We used three
relationship to SC and IT. Each path analysis involved regressing
items to measure the perceived costs associated with RPV and
each variable in the causal chain on all of the variables that pre-
whether respondents thought it was affordable for their household
ceded it, controlling for household constraints. Using these path
budget (Cronbach’s ␣ = 0.63)
models, we estimated both direct effects, net of other variables,
and total effects, which include both the direct effects and indirect
2.2.3.2. Normative beliefs. Three items measured the extent to
effects through other variables. The causal models we investigated
which respondents felt their family members and friends would
using elements of the three theories are represented in Fig. 2. While
be supportive if they decided to get solar (Cronbach’s ␣ = 0.72).
existing theory guided the order of variables for the VBN and TPB
path models, no causal ordering, to our knowledge, has previously
2.2.3.3. Perceived behavioral control. Several practical matters
been suggested for DOI. Based on the logic that broad dispositions
might make RPV unfeasible for some homeowners, even if they
may influence how an innovation is perceived, we tested a model
have positive beliefs about RPV and support from their social net-
in which the two measures of consumer innovativeness occur at
works.
the beginning of the causal chain, prior to relative advantage, trial-
Perceived unsuitability
ability, and riskiness, which in turn determine social curiosity and
Homeowners may quickly dismiss RPV if they believe their
interest. We expect perceptions of RPV to be further influenced
homes or locations unsuitable. Three items measured beliefs about
by marketing messages and information from trusted communi-
impediments to getting RPV such as inadequate sunlight and the
cation channels, such as the PV industry and one’s social network.
shadiness of their property (Cronbach’s ␣ = 0.89).
We additionally modelled observability as antecedent to the other
Expecting to move
perceptions of RPV. While this relationship is not explicit in DOI,
RPV is a relatively long-term investment. Unless homeowners
logically, the more prevalent RPV seems, the less likely consumers
perceive that RPV will substantially improve resale value, they are
are to question its risks and benefits or to feel the need to “try it
not likely to be interested if they anticipate moving in the near
out.”
future. We measured this concern with a single item, “I may not be
All models were estimated using the structural equation model-
in my home long enough to get the benefits of investing in solar.”
ing features of Stata 14 (sem and teffects). This provides maximum
likelihood estimates with standard errors based on the observed
2.2.4. Household characteristics and constraints
information matrix. For ease of interpretation, we estimated
We included a variety of socio-demographics and household
adjusted R2 values for each model using OLS multiple regression.
characteristics that may directly or indirectly affect a household’s
To determine how much unique variance the set of variables from
ability to pursue RPV. These include variables related to financial
each theory contributes to IT, we also examined each set of variables
status (income, perceived financial stress, retirement status), char-
using a nested sequence of OLS multiple regressions. Household
acteristics of household decision-makers (age, gender, education),
constraints were controlled in the first step, variables from the the-
and household characteristics (home size, number of inhabitants,
ory of interest were added in the second step, and social curiosity
past electricity bills, and experiences with power outages). The
(the penultimate DV) was added in the third step. Finally, we tested
online supplementary material provides more details about these
an integrated model that combined the variables from all three the-
measures. We additionally controlled for state of residence to
ories, again controlling for household constraints. Table 1 displays
account for differences in electricity rates, available incentives, and
the correlations of all included variables. The highest variance infla-
local policies.
tion factor in the integrated model was 3.68, indicating no overall
problematic collinearity among independent variables.
2.2.5. Dependent variables: interest in pursuing residential solar
We used two measures to assess interest in adopting solar. The
final variable in our causal models is Interest in Talking to an RPV 3.2. Sample profile
installer. The penultimate variable is Social Curiosity about Solar.
We are modeling not adoption per se, but interest in learning more Survey respondents were 44% male and 56% female with a mean
and in engaging with firms that could provide RPV. Theoretically, age of 56 (SD = 14.5, range = 18–94). The sample was moderately
these dependent variables are more akin to behavioral intentions educated: 21% of respondents reported having a high school degree
than to actual behavior. or less, 24% completed some college, 33% had an associate’s or
We measured interest in residential solar using five items. Two bachelor’s degree, and 21% had a graduate or professional degree.
items asked respondents how interested they would be in learning Median annual household income was between $75,000–$99,999,
about the costs and benefits of solar if a friend, family member, and 37% of the sample reported being retired. Familiarity with RPV
or neighbor installed a RPV system. These items were averaged to was generally high: 62% of respondents indicated that they had
form the Social Curiosity scale (SC) (Cronbach’s ␣ = 0.84). Two other thought about getting solar panels for their home, 73% had seen
items measured interest in talking to a solar installation company, solar panels installed on someone’s home, and 65% recalled see-
and one measured intention to talk to a solar company in the next ing PV advertising in the previous 6 months. The supplementary
Table 1
Correlation table of scales and constructs (N = 904).
K.S. Wolske et al. / Energy Research & Social Science 25 (2017) 134–151
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
1 Altruism
2 Traditionalism 0.44**
3 Self-interest 0.29** 0.28**
4 Openness 0.52** 0.33** 0.45**
5 AC 0.61** 0.07** 0.11** 0.30**
6 PN 0.68** 0.19** 0.18** 0.37** 0.71**
7 CNS 0.36** 0.27** 0.37** 0.49** 0.23** 0.36**
8 CIJM −0.26** −0.28** −0.17** −0.23** −0.13** −0.24** −0.42**
9 Observability 0.12** 0.08* 0.05 0.14** 0.07* 0.16** 0.21** −0.16**
10 Marketing 0.05 0.05 0.05 0.05 0.00 0.07* 0.08* −0.10** 0.33**
11 Trust PV Ind. 0.35** 0.17** 0.18** 0.27** 0.31** 0.38** 0.31** −0.21** 0.09** −0.02
12 Trust Social 0.25** 0.25** 0.17** 0.13** 0.09** 0.22** 0.24** −0.38** 0.14** 0.00 0.32**
13 Pers. Benefits 0.44** 0.16** 0.20** 0.26** 0.44** 0.56** 0.35** −0.25** 0.12** −0.02 0.48** 0.27**
14 Environ. Ben. 0.57** 0.13** 0.16** 0.28** 0.69** 0.74** 0.32** −0.20** 0.11** 0.02 0.45** 0.25** 0.74**
15 Riskiness −0.18** 0.02 0.09* −0.09** −0.20** −0.20** −0.10** 0.03 −0.12** −0.01 −0.28** −0.07** −0.31** −0.29**
16 Expense Concerns 0.00 0.13** 0.02 0.01 −0.07* −0.06 −0.02 −0.06 −0.10** 0.01 −0.15** 0.09 0.00 −0.01 0.22**
17 PV may improve −0.02 0.08* 0.17** 0.08* −0.04 −0.03 0.07* −0.06 0.05 0.07* −0.07* 0.05 −0.06* −0.05 0.32** 0.23**
18 Trialability 0.16** 0.21** 0.13** 0.14** 0.06 0.11** 0.15 −0.26** 0.05 0.04 −0.02 0.22** 0.09** 0.08** 0.24** 0.29** 0.34**
19 Soc. Support 0.27** 0.10** −0.03 0.14** 0.29** 0.36** 0.25** −0.13** 0.14** 0.01 0.32** 0.23** 0.46** 0.45** −0.32** −0.03 −0.07* 0.06
20 Home Unsuitable −0.04 −0.09** 0.03 −0.05 −0.01 −0.06 −0.07* 0.02 −0.23** −0.20** −0.04 −0.10** −0.17** −0.11** 0.23** −0.05 0.01 0.01 −0.25**
21 May move −0.05* −0.03 0.04 −0.05 −0.09* −0.08* −0.16** .08* −0.04 0.00 −0.19** −0.08* −0.24** −0.14** 0.21** 0.17** 0.21** 0.16** −0.22** 0.10**
22 SC 0.27** 0.16** 0.07* 0.18** 0.25** 0.32** 0.27** −0.26** 0.13** 0.05 0.29** 0.33** 0.42** 0.40** −0.10** 0.03 0.07* 0.40** 0.46** −0.14** −0.10**
23 IT 0.27** 0.09** 0.17** 0.19** 0.24** 0.36** 0.37** −0.17** 0.15** −0.01 0.42** 0.25** 0.55** 0.44** −0.18** −0.11** −0.02 0.18** 0.43** −0.03 −0.22** 0.54**
*
p < 0.05. ∗∗ p < 0.01.
141
Fig. 2. Proposed path models for elements of each of the three theories represented in the dataset.
material provides the means for all theoretical predictors used in controlling for household constraints – provided in Table 2. The
our analyses. added variance explained by each set of theoretical variables is
shown in Table 3. On their own, household constraints explained
3.2.1. Tests of single-theory variables 11% of variance in interest. Younger individuals and men were more
Significant paths for the tested elements of each theory are likely to express interest in solar, as were individuals with higher
shown in Fig. 3, with corresponding direct and total effects – after summer electricity bills, more power outages, and lower household
Fig. 3. Significant pathways for each theory of interest. Statistically significant and positive relationships are indicated with a solid arrow while significant, negative
relationships are marked with a dashed arrow. See Table 2 for standardized coefficients.
incomes. Only age and gender remained statistically significant in 3.2.2. Value-belief-norm theory
all three models once theoretical variables were added. After controlling for household constraints, the tested VBN
variables accounted for an additional 11% of variance in interest.
Consistent with the VBN framework, personal norm had a signifi-
cant direct effect on interest once social curiosity was included as a
Table 2
Causal chain analyses to test direct and total effects of select VBN, TPB, and DOI variables on interest in talking to an RPV provider (standardized coefficients), controlling for
household constraints (not shown).
IT SC PN AC
VBN Direct Total Direct Total Direct Total Direct Total
Social curiosity 0.44*** 0.44***

Personal norms 0.21*** 0.31*** 0.22*** 0.22***
Awareness of conseq. −0.05 0.11** 0.05 0.15*** 0.46*** 0.46***
Values
Altruism 0.05 0.28*** 0.07 0.27*** 0.40*** 0.72*** 0. 70*** 0.70***
Self-interest 0.08** 0.06 −0.05 −0.06 0.01 −0.01 −0.05 −0.05
Traditionalism −0.05 −0.04 0.09** 0.05 −0.02 −0.13*** −0.22*** −0.22***
Openness to change −0.03 −0.02 0.02 0.03 0.02 0.03 0.02 0.02
2
R Adj 0.39*** 0.16*** 0.61*** 0.42***
IT SC Relative Advantage Riskiness Trialability
DOI Direct Total Direct Total Direct Total Direct Total Direct Total

Characteristics of Innovation
Relative advantage† 0.29*** 0.39*** 0.31*** 0.31***
Riskiness −0.02 −0.04 −0.05 −0.05
Trialability 0.03 0.15*** 0.36*** 0.36***
Observability 0.05 0.07* 0.00 0.02 0.05* 0.05* −0.11** −0.11** −0.02 −0.02
Exposure to marketing −0.04 −0.03 0.02 0.02 −0.01 −0.01 0.01 0.01 0.01 0.01
Trust in PV Industry 0.15*** 0.31*** 0.10*** 0.17*** 0.34*** 0.34*** −0.28*** −0.28*** −0.14*** −0.14***
Trust in Social Network −0.02 0.07* 0.12*** 0.20*** 0.06* 0.06* 0.01 0.01 0.19*** 0.19***
Innovativeness
CNS 0.13*** 0.21*** 0.03 0.11*** 0.16*** 0.16*** −0.02 −0.02 0.09* 0.09*
CIJM 0.09** 0.04 −0.03 −0.10*** −0.04 −0.04 −0.05 −0.05 −0.18*** −0.18***
R2 Adj 0.48*** 0.38*** 0.35*** 0.08*** 0.11***
IT SC
TPB Direct Total Direct Total

Behavioral Beliefs
Personal benefits 0.35*** 0.43*** 0.26*** 0.26***
Environ. Benefits −0.01 0.02 0.11* 0.11*
Riskiness −0.02 0.01 0.09** 0.09**
PV will improve 0.01 0.03 0.06 0.06
Expense concerns −0.11*** −0.09*** 0.04 0.04
Normative beliefs 0.11*** 0.22*** 0.32*** 0.32***
Perceived behavioral control

Home unsuitable 0.14*** 0.14*** −0.01 −0.01
Might move −0.03 −0.03 −0.00 −0.00
R2 Adj 0.48*** 0.31***
N = 904. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001; †Relative advantage in DOI is the same as Personal benefits in TPB.
Table 3 predictor. Altruism and awareness of consequences had significant

Series of nested regressions to predict interest in RPV.
and positive indirect effects on both dependent variables through
Step Predictor set R2 F personal norms. Self-interest had a small and positive direct effect
VBN on IT while traditionalism had a positive, direct effect on social
1 Household constraints 0.13 8.21*** curiosity.
2 VBN variables 0.12 23.18***
3 Social Curiosity 0.16 237.39***
3.2.3. Diffusion of innovations
Total R2 Adj 0.39
Variables from DOI explained an additional 31% of variance
DOI after controlling for household constraints. As predicted by DOI,
1 Household constraints 0.13 8.21*** innovativeness was positively related to interest in talking to
2 DOI variables 0.31 53.06***
an RPV installer. Consumer novelty seeking had strong positive
direct effects on interest and positive indirect effects through
Total R2 Adj 0.48
perceived relative advantage. Consumer independent judgment
TPB making (CIJM) had a positive direct effect on interest in talking to an
1 Household constraints 0.13 8.21*** installer and an indirect, negative relationship with social curiosity
2 TPB variables 0.29 55.38***
through trialability. That is, individuals who depend on others to
inform their purchasing decisions have a greater need to “try out”
Total R2 Adj 0.48 RPV and learn about it from current adopters.
N = 904. *** p < 0.001. Among the perceived characteristics of the innovation, rela-
tive advantage had the strongest effect on interest in talking to an
installer, while the need for trialability was the strongest predictor
of social curiosity. Seeing others with RPV systems (observability) VBN model, the change in direction may indicate the presence of a
indirectly influenced interest by increasing relative advantage and suppression effect. Novelty seeking, consumer independent judg-
reducing the perceived riskiness of RPV. The results also highlight ment making, observability via a respondent’s social network, and
the importance of trustworthy communication channels. Trust in trust in the industry all increase interest. Personal norms, expo-
the PV industry significantly reduced perceived riskiness and need sure to marketing and trust in one’s social network have no direct
for trialability, while increasing relative advantage. Trust in one’s effect. Values have no net direct effect, except openness to change
social network was positively correlated with trialability and inter- which had a small negative effect. Among the demographics, older
est in learning about others’ systems (SC). Exposure to PV marketing respondents and women are less likely to express interest, while
did not influence how respondents weigh the advantages and dis- no other demographic variable has a significant direct effect. Most
advantages of going solar. This result is in line with Rogers’ [20] variables with a significant direct effect also have a significant total
proposition that mass media communications are less useful at the effect. Awareness of consequences, consumer independent judg-
persuasion stage of diffusion. ment making, and openness to change are exceptions in that each
has a direct effect, but paths with opposite signs lead to no overall
3.2.4. Theory of planned behavior total effect.
TPB variables accounted for 29% of the variance in interest after Several variables have no direct effect but do have signifi-
accounting for household constraints. With social curiosity in the cant total effects, thus indirectly influencing interest in talking
model, the belief that solar panels would be personally beneficial to a PV company. These include personal norms, altruism, and
had the strongest direct and total effects on interest, followed by higher summer electric bills. Notably, our causal chain analysis (see
the belief that one’s peers would be supportive of the decision to Supplementary material) suggests that the effect of personal envi-
go solar. As might be expected, believing that RPV were too expen- ronmental norms may be partially mediated by perceived personal
sive was negatively associated with interest. Surprisingly, the more benefits of solar and perceived social support. In other words, those
individuals thought their properties was unsuitable for solar, the who feel morally obligated to act on energy and climate change
more likely they were to be interested in talking to a PV company. issues are more likely to believe that RPV would benefit them per-
This finding may reflect an interest in confirming whether one’s sonally and would be supported by their peers.
property is truly unsuitable. In the model predicting social curiosity, we find a similar pat-
Respondents were more likely to express interest in learning tern of results; however, variables concerning one’s social network
about peers’ RPV systems (social curiosity) if they believed solar have a stronger effect. As might be expected, the need to “try out”
panels would have personal benefits, help the environment, and RPV (trialability) and perceived social support both have strong
be supported by their peers. Perceiving that solar panels are risky effects on the desire to learn about RPV from friends and neigh-
also increased interest in learning from others’ experiences. Neither bors. Believing that RPV could be personally beneficial and trusting
of the perceived behavioral control measures had direct effects on one’s social network are also significant predictors. Similar to the
social curiosity. model for IT, females are less curious about others’ RPV systems
than their male counterparts. Again, personal norms and altruistic
3.3. Test of integrated model values have significant indirect effects on social curiosity, but not
direct effects.
Our proposed framework, which incorporates variables from all
three theories (shown in Fig. 4), hypothesizes that interest in RPV is
4. Discussion
most directly affected by specific beliefs about RPV. These include
beliefs about the benefits and risks of solar, perceived social support
The aim of this study was to understand what drives or hinders
for getting RPV, and perceived behavioral control. We propose that
interest in residential photovoltaics, and particularly, interest in
these three sets of variables (drawn from TPB and DOI) are, in turn,
contacting an RPV provider. We developed an explanatory model
influenced by two sets of factors – external influences and personal
based on data from a survey of non-adopter homeowners in Ari-
dispositions. External influences encompass the communication
zona, California, New Jersey, and New York. In doing so, we drew
channels through which individuals learn about RPV as well as
on a selection of variables from three behavior change theories:
trust in those information sources. Personal dispositions combine
VBN, TPB, and DOI. As we conjectured, variables from each the-
elements of VBN and DOI, including personal norms about address-
ory explained some variance in interest, suggesting that consumers
ing climate change and energy issues and consumers’ dispositions
can be understood to see RPV in multiple ways: as an environmen-
about innovation. Finally, the model supposes that personal dispo-
tal benefit, as a consumer durable, and as a novel technology. Our
sitions are shaped by broad value orientations.
explanatory model combines variables from the three theories and
Table 4 provides the direct and total effects for the integrated
provides a more holistic picture of how these different perspec-
model. The full causal chain analysis is available in the supplemen-
tives interrelate. The following sections discuss the value of the
tary material. Looking first at direct effects on interest in talking
integrated model, offer implications for facilitating RPV adoption,
to a solar provider, we find that, not surprisingly, social curiosity
and propose directions for future research.
plays a very strong role. However, even when social curiosity is
controlled, a number of other variables still have significant direct
effects. Among the behavior-specific beliefs and attitudes, which 4.1. Contributions of individual theories
are assumed to be closest to interest and curiosity in causal order,
perceived personal benefits, trialability, and social support all have VBN, TPB and DOI each illuminate aspects of the process that
positive effects, while concerns about costs have negative effects. leads homeowners to consider RPV. In our integrated framework,
Perceived home unsuitability again had an unexpected positive variables in TPB and DOI have some of the strongest explanatory
effect. Several variables suggested by DOI and TPB have no net power, no doubt because many reflect behavior-specific beliefs
effect: perceived environmental benefits, riskiness, beliefs that RPV that we expect to be more proximate to intention [38]. These
will improve, and an expectation that the respondent might move. include perceived personal benefits (relative advantage in DOI) and
Moving back through the causal chain, awareness of consequences, concerns about costs and home suitability. In this regard, both the-
contrary to theoretical expectations, reduces interest, net of all ories account for a process of weighing the pros and cons of the
other variables. As AC had a positive relationship to IT in the partial behavior. TPB and DOI also each account for the significant role
Fig. 4. Proposed integrated framework for predicting interest in pursuing RPV.
of social influence, albeit in different ways. DOI captures the per- RPV for non-environmental reasons as well as specifying the types
ceived descriptive norm through observability while TPB considers of beliefs that may lead to favorable attitudes about the technol-
more explicitly whether individuals believe their peers would be ogy. The integrated framework illuminates the relationships among
supportive of the behavior. these aspects (and implicitly among the theories) and highlights the
DOI differs from TPB in that it also factors in broad personal ways in which each is important.
dispositions of the individual, particularly a disposition of innova- The path models that are fairly explicit in the literature for VBN
tiveness. In the full model, consumer novelty seeking, a measure of and TPB, as well as the diffusion process outlined by DOI, imply that
innovativeness, had strong direct effects on interest as well as sig- individuals go through steps in reaching the decision to adopt solar.
nificant indirect effects through more specific behavioral beliefs. As we note in the next section, for some variables, such as values,
By contrast, the integrated model to predict social curiosity seems we cannot expect interventions to change individuals; rather influ-
to illustrate the importance of social influence among individuals ential variables suggest groups to whom communication might be
who are less innovative. For this outcome, CNS has no direct or total directed. For other variables, there may be ways to intervene.
effects. Rather social curiosity is most strongly associated with the
need to try out RPV (trialability), perceived social support, trust in
4.2. Practical implications of integrated framework
the information provide by one’s social network, and believing that
RPV will be personally beneficial.
The results of our integrated framework have several impli-
VBN, like DOI, helps identify people who may be more inclined to
cations for policymakers and PV marketers looking to increase
consider RPV. Consistent with past research, altruistic values and
adoption of RPV. First, our results shed light on the types of peo-
awareness of consequences were found to have significant indi-
ple who may be most apt to consider solar at a relatively early
rect effects on interest in RPV through proenvironmental norms.
stage in its diffusion. In line with past research on adoption of eco-
While proenvironmental norms had a strong direct effect on inter-
innovations, our results show that people higher in novelty seeking
est in the VBN-only model, this relationship disappeared in the full
have greater interest in pursuing RPV [45,49]. For marketers, this
integrated model. These results suggest that the effect of personal
might suggest targeting consumers who are early adopters of
norms is mediated by various behavior-specific beliefs and atti-
other technologies. Owners of hybrid and electric vehicles are an
tudes. Notably, personal norm had the strongest direct effect on
obvious choice – and a group already heavily pursued by the PV
perceived personal benefits (␤ = 0.35) and was also a significant
industry. But, focusing on consumers of other non-environmental
predictor of perceived social support. These findings are consistent
innovations may be equally worthwhile. For example, individu-
with past research indicating that the explanatory power of per-
als interested in “smart” technologies (e.g., the Nest thermostat)
sonal norms for household energy actions decreases as other more
and home automation may find appeal in the online platforms that
proximate factors, such as cost and behavioral difficulty, become
allow RPV adopters to monitor how much energy their system is
more important [48]. In the context of understanding what leads
generating. Past work suggests that early adopters of RPV may be
people to adopt particular eco-innovations such as RPV, we see VBN
especially drawn to monitoring their energy generation and usage
as a useful addition for identifying consumer segments that have a
[18]. Our results also support targeting environmentally-minded
higher propensity to adopt.
households. In the integrated model, the strong indirect effect
Our interpretation of the overall results is that, as with the para-
of proenvironmental environmental norms on interest (largely
ble of the elephant, each off-the-shelf theory captures a different
through perceived personal benefits) suggests that those who feel
aspect of the phenomenon: VBN explains what predisposes individ-
morally obligated to address climate change already believe that
uals to take proenvironmental action. TPB helps explain why the
solar could be beneficial. As perceived environmental benefits did
individual chooses particular behaviors. DOI adds further insight
not predict interest in RPV, marketers may wise to target this seg-
by characterizing individuals who may be motivated to consider
ment by emphasizing how RPV would meet households’ needs.
Table 4
Direct and total effects for integrated model on Social Curiosity and Interest in Talking to a PV Company (standardized coefficients).
IT SC
Independent Variables Direct Total Direct Total
RPV-specific beliefs & attitudes

Personal benefits (TPB)/Relative advantage (DOI) 0.29*** 0.34*** 0.18*** 0.18***
Environ. Benefits (TPB) −0.02 0.00 0.07 0.07
Riskiness (TPB/DOI) −0.02 −0.01 0.03 0.03
Expense concerns (TPB) −0.09** −0.10*** −0.04 −0.04
PV will improve (TPB) −0.02 −0.03 −0.03 −0.03
Trialability (DOI) 0.08** 0.19*** 0.37*** 0.37***
Normative beliefs (TPB) 0.10*** 0.18*** 0.26*** 0.26***
Home unsuitable (TPB) 0.13*** 0.13*** −0.01 −0.01
Might move (TPB) −0.04 −0.04 −0.02 −0.02
Personal dispositions
Personal norms (VBN) 0.05 0.20*** −0.02 0.13**
Awareness of consequences (VBN) −0.08* −0.03 0.02 0.08
CNS (DOI) 0.14*** 0.20*** 0.02 0.08*
CIJM (DOI) 0.08** 0.04 −0.04 −0.09**
External influences
Observability (DOI) 0.05 0.06* −0.02 0.00
PV Marketing (DOI) −0.02 −0.04 0.02 0.01
Trust PV Industry (DOI) 0.14*** 0.26*** 0.07* 0.12***
Trust Social Network (DOI) −0.02 0.05 0.10*** 0.20***
Values
Altruism (VBN) 0.02 0.20*** 0.01 0.20***
Traditionalism (VBN) −0.03 −0.05 −0.01 0.03
Self-interest (VBN) 0.05 0.05 −0.07* −0.08*
Openness to change (VBN) −0.07* −0.04 0.02 0.04
Household constraints
Age −0.11** −0.23*** 0.01 −0.10*
Gender (female = 1, male = 0) −0.11*** −0.17*** −0.14*** −0.16***
Education 0.05 0.07* 0.04 0.08*
Income −0.04 −0.05 0.08* 0.02
Financial stress −0.01 −0.02 0.04 0.01
Retired 0.00 0.01 −0.04 −0.03
Number of adults −0.02 −0.01 0.04 0.02
Number of children 0.00 0.00 −0.02 −0.03
Summer electric bills 0.05 0.10* −0.02 0.06
Winter electric bills 0.00 −0.02 −0.03 −0.06
Square footage 0.00 −0.01 −0.01 −0.01
Years expected to remain in house 0.03 0.04 −0.02 −0.01
Number of power outages in past yr. −0.01 0.05 0.02 0.05
NJ 0.04 0.00 −0.03 −0.05
AZ 0.03 −0.03 0.07 0.04
CA 0.01 −0.02 −0.03 −0.03
R2 Adj 0.51*** 0.44***
N = 904. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. CFI = 0.784. RMSEA = 0.17.
Secondly, our findings suggest that potential adopters need complex programs may obscure the intended benefits and deter
assurance – from trusted sources – of the personal benefits of homeowners from adopting [51].
RPV. This conclusion is reflected in the strong influence of per- Finally, consistent with past research, our results highlight
ceived personal benefits on both social curiosity and interest, the the value of leveraging existing social networks to promote RPV
strong effect of social curiosity on interest in talking to an installer, [10,16,17]. Seeing solar panels on others’ homes, having social
and the significant effects of both trust in the industry and trust support and being curious about peers’ RPV systems were all sig-
in one’s social network. For industry, these findings suggest that nificant predictors of interest in talking to an installer. Likewise,
consumers need trustworthy information on the performance and trialability, social support, and trust in social networks were strong
costs of RPV. Trusted sources may be personal associates such as predictors of social curiosity. Taken together, these findings speak
neighbors with solar, or they might be neutral, third-party sources to the importance of customer referral programs and word-of-
such as government agencies or nonprofit organizations. Indeed, mouth marketing where consumers can “try out” RPV vicariously
other work has demonstrated the importance of trusted communi- through others. Some installers, for example, have worked with
ties of information such as regional renewable energy associations, existing customers to host Tupperware-style parties where friends
Solarize programs, and solar cooperatives [18,50]. RPV installers and neighbors are invited to the customer’s home to learn about
may also increase confidence in the benefits of RPV through con- the installed system and to hear from satisfied customers in the
tractual guarantees of system performance. For policymakers, our neighborhood. For local municipalities looking to promote RPV, our
findings point not only to the importance of incentivizing RPV adop- results reinforce the value of group-buy programs, such as Solar-
tion, but also of creating policies that are transparent and easy to ize campaigns, where homeowners get discounted solar panels by
take advantage of. Even in states where incentives are generous, signing up with others in their community [50].
4.3. Limitations and future research directions solar other than those captured in our implementation of the three
theories. For example, a desire for national energy independence
Our analysis examines factors relatively early in the chain of has been shown as an important motivator for energy efficiency in
actions that would lead to solar adoption: interest in learning more some studies [e.g., Ref. [52]]. A desire for the household to be inde-
about solar and intent to do so. As our sample comprised house- pendent of the local power grid may also be an important motivator
holders who had not previously pursued solar, our conclusions [36]. Although we did not find recent experience with power out-
about the determinants of interest in RPV may only be accurate for ages a significant influence, further research could develop a more
individuals who have not yet given serious consideration to solar robust approach to exploring the desire for being “off the grid.”
panels. Different influences may be important to adoption deci- Field experiments and panel data would be very useful to pro-
sions after contact with providers has been made, as DOI theory vide stronger causal inference, and we emphasize that both are
suggests. Moreover, our data concern the determinants of these valuable in different ways [53]. The most visible policies focus
factors at a relatively early stage in the evolution of RPV adoption on providing financial incentives but our results also demonstrate
(in the terms of DOI theory, our respondents mainly are potential the importance of various forms of trust. Clearly experiments that
early adopters). Under these conditions, variability across individ- consider alternative methods of building trust among potential
uals in perceptions of risks and benefits may be more influential adopters would be of practical importance, and because trust is
than they are when the technology is more familiar. We anticipate an issue of ongoing research, could contribute to fundamental
that predictors will be different when and where adoption of RPV is understanding as well [54–56]. There may also be important nat-
more commonplace. For example, factors such as consumer novelty ural experiments. For example, while in many communities there
seeking will likely be less predictive of interest as novelty declines. are financial incentives to feed electricity from RPV into the grid,
Furthermore, as our survey only sampled households in four in some communities incentives are being reduced or removed
U.S. states, our results may not be generalizable to geographic [57,58]. Evaluating such changes could provide useful information
regions where market or policy conditions markedly differ or where on the importance of financial incentives relative to other factors.
diffusion is at an earlier or later stage. The RPV markets in Ari-
zona, California, New Jersey and New York are among the most 4.4. Conclusion
established in the U.S. and have experienced consistent growth
since 2010 [9]. They are also dominated by third-party owned sys- In this study we introduce an integrated framework that com-
tems and, in the past, have had favorable state-level policies and bines variables from VBN, TPB, and DOI to explain consumer
incentives. Consequently, residents in these states may be more interest in residential photovoltaics. We find that each of these
acquainted with RPV and have more positive perceptions of it three theories contributes some understanding of interest in RPV,
than residents in states where the diffusion of RPV is relatively but that combining elements of the three theories provides a more
unchanged from a decade ago. Additional testing of our integrated complete picture. In places where RPV is at an early stage in the dif-
framework is needed to determine whether the pattern of results fusion process, we find that those most interested in RPV tend to be
holds among people less familiar with RPV, or in places outside the innovative consumers who seek novel goods and technologies as
USA where familiarity is greater or the decision context is different. well as those who feel a moral obligation to address energy issues.
For instance, in locations where RPV is relatively rare, beliefs about We find that need for trustworthy information about the costs and
the personal economics of solar might have even greater weight, as performance of RPV is an important factor influencing interest in
might the desire to learn about RPV from trusted others. In juris- speaking with a provider, and that consumers may seek such infor-
dictions that offer attractive feed-in tariffs, the personal economics mation from the experiences of personal acquaintances in their
may be so self-evident as to greatly reduce variation, and therefore social networks. Our data highlight the importance of meeting this
predictive value, of perceived economic benefit. need for promoting interest in this relatively novel technology.
Lastly, concerns about survey length and participants’ time kept
us from assessing all of the variables in each of the three theo- Funding
ries. For example, we did not measure ecological worldviews or
ascription of responsibility in VBN, and we might have attempted This work was supported by the U.S. Department of Energy’s
to measure additional behavioral beliefs in TPB or to add other mea- Solar Energy Evolution and Diffusion Studies (SEEDS) program
sures of relative advantage, complexity and compatibility in DOI. under award no. 26154 and by Michigan AgBio Research.
Because we have not assessed the theories in their entirety and
because several variables could logically be attributed to more than Acknowledgements
one theory, we have not attempted to draw conclusions that one
theory is “better” than others for explaining interest in RPV. We The survey used in this study was collaboratively developed
have shown, however, that explanation of interest in RPV benefits with (in alphabetical order): Easan Drury, Debi Elliott, Adam Dou-
from drawing on insights from all three theories. We believe this glas Henry, Aaron Ingle, Loren Lutzenhiser, Mithra Moezzi, Ben
study also shows the value of a multi-stage causal model that links Sigrin, Seth Spielman, Annika Todd, James Tong, and Ryan Wiser.
VBN and theories that focus on attitudes and beliefs about spe- We would also like to thank the editor and our anonymous review-
cific proenvironmental actions and that considers predictors both ers for their insightful comments on an earlier version of the
intrinsic and extrinsic to the household. Future research to test such manuscript.
structural models can enhance our understanding of how variables
from these theories interact as part of a larger explanatory frame- Appendix A.
work. Further work might also consider motivations for installing
Table A1
VBN Items and Factor Loadings.
Factor loading Mean (SD)
Values
Altruism (␣ = 0.87) 3.84 (0.80)
Respecting the earth, harmony with other species 0.85
Protecting the environment, preserving nature 0.82
A world at peace, free of war and conflict 0.79
Unity with nature, fitting into nature 0.77
Equality, equal opportunity for all 0.66
Social justice, correcting injustice, care for the weak 0.65
Self-interest (␣ = 0.63) 2.91 (0.81)

Wealth, material possessions, money 0.82
Authority, the right to lead or command 0.75
Influential, having an impact on people and events 0.54
Traditionalism (␣ = 0.62) 4.24 (0.63)

Honoring parents and elders, showing respect 0.81
Family security, safety for loved ones 0.75
Self- discipline, self-restraint, resistance to temptation 0.58
Openness to Change (␣ = 0.79) 3.52 (0.82)

Curious, interested in everything, exploring −0.82
A varied life, filled with challenge, novelty, and change −0.81
An exciting life, stimulating experiences −0.76
Awareness of consequences
Climate change is a serious problem for society. 3.74 (1.20)
Personal norm (␣ = 0.83) 3.48 (0.86)

I feel a personal obligation to do my part to move the country to a renewable energy future country to a renewable energy future. –
I feel a personal obligation to do my part to prevent climate change. –
I feel guilty when I waste energy. –
Factor loadings and scale reliabilities based on initial sample, N = 1156. Individuals who responded “Don’t Know” to all items in a factor were dropped from subsequent
analyses (N = 252). Means are based on remaining sample, N = 904. Value items were on a 5-point scale from 1 = Not at all important to 5 = Extremely important, with an option
of marking Opposed to my values (coded as −1). AC and PN items were measured on a 5-point scale from 1 = Strongly disagree to 5 = Strongly agree.
Table A2
DOI Items and Factor Loadings.
Consumer innovativeness
Consumer Novelty Seeking (␣ = 0.87) 3.05 (0.90)
I continuously look for new experiences from new products. 0.95
I continuously look for new products and brands. 0.93
I like to visit places where I’m exposed to information about new products and brands. 0.79
Consumer Independent Judgment Making (␣ = 0.83) 2.38 (0.75)

Before I buy a new product or service, I often ask acquaintances about their experiences with that product or service. (reversed) 0.90
Before buying a new brand, I usually ask someone who has experience with the brand for advice. (reversed) 0.88
When considering a new product/service, I usually trust the opinions of friends who have used the product/service. (reversed) 0.81
External influences
Exposure to PV marketing (␣ = 0.55) 0.94 (0.77)
In the last six months, I have seen or heard advertisements from companies that install solar panels. –
My family has recently received advertising or a call from a company that installs solar panels. –
Trust in PV Industry (␣ = 0.88) 2.55 (0.82)

Companies that provide or install solar panels –
Solar industry trade organizations –
Trust in Social Network (␣ = 0.81) 3.68 (0.72)

Friends or family –
Neighbors –
Characteristics of the innovation

Observability (␣ = 0.71) 1.87 (1.54)
I have personally seen solar panels installed on someone’s home. –
A friend or neighbor has recently installed solar panels. –
I know more than one person with solar panels on their home. –
Several people in my neighborhood have installed solar panels. –
I have talked about solar panels with someone who has already installed them on their home. –
Relative Advantage (␣ = 0.90) 3.38 (0.77)

Using solar would save me money. 0.89
Installing solar provides a great return on a family’s investment. 0.82
Using solar will help protect my family from rising electricity prices in the future. 0.70
Table A2 (Continued)
Having solar panels on my home would help meet my family’s needs. 0.64
Using solar could protect my family from electricity blackouts. 0.60
Solar panels would increase my property value. 0.57
Riskiness (␣ = 0.63) 2.66 (0.80)

I would worry about having solar panels because it would be an unfamiliar experience. 0.78
Installing solar panels is a risky thing for a household to do. 0.75
Installing solar panels could damage my home. 0.75
Trialability (␣ = 0.70) 3.82 (0.63)

Before contacting a solar company, I would like to see solar panels up close on someone else’s house. 0.88
Before considering solar panels, I would want to talk to someone who has solar panels on their home. 0.86
If a solar company tells me how much I would save on my electric bills by installing solar, I would want a second opinion. 0.55
I would be more interested in solar if there were some way for me to try it out before installing it. 0.51
analyses (N = 252). Means are based on remaining sample, N = 904. Consumer Innovativeness items were rated on a 5-point scale from 1 = Not at all like me to 5 = Just like
me. † All items for CIJM were reversed so that higher scores on the scale indicate greater independence in judgment making. Except for Observability, Characteristics of the
Innovation were rated on a 5-point scale from 1 = Strongly disagree to 5 = Strongly agree. ‡ The Observability and Exposure to PV Marketing scales were created by summing
“yes” responses on the listed binary items. Trust items were measured on a 5-point scale from 1 = Do not trust at all to 5 = Trust completely.
Table A3
TPB Items and Factor Loadings.
Behavioral beliefs
Personal Benefits (␣ = 0.90) 3.38 (0.77)
(same items as Relative Advantage in DOI)
Environmental Benefits (␣ = 0.87) 3.49 (0.88)

Solar panels help slow down climate change. −0.87
If more households get solar panels, environmental quality will improve. −0.73
Having solar panels would be a good way to reduce my environmental impact. −0.52
Perceived Risks (␣ = 0.63) 2.66 (0.80)

(same items as Riskiness in DOI)
Expense Concerns (␣ = 0.63) 3.61 (0.72)

I can’t afford solar on my family budget. 0.86
Solar panels are still very expensive, even with government incentives. 0.74
Maintaining solar panels is expensive. 0.54
Waiting for Improvements (␣ = 0.65) 3.27 (0.69)

The prices of solar power keep going down, so it is wise to wait before deciding whether to install it. 0.85
Solar panel technology will only get better, so it doesn’t make sense to get them now. 0.78
Subjective norms
Normative Beliefs (␣ = 0.72) 3.56 (0.74)
Most people who are important to me would support me if I decided to go solar. 0.83
People who are important to me would be in favor of installing solar panels. 0.75
My family members would be opposed to getting solar panels. (reversed) −0.59
Perceived behavioral control

Unsuitable home (␣ = 0.89) 2.18 (1.00)
It’s not sunny enough in my area for solar panels to work well. 0.96
It’s too cloudy where I live for solar panels to be effective. 0.92
At my home, there’s no place to put solar panels that would get enough sunlight. 0.84
Might Move 3.35 (1.17)

I may not be in my home long enough to get the benefits of investing in solar. –
analyses (N = 252). Means are based on remaining sample, N = 904. All items were rated on a 5-point scale from 1 = Strongly disagree to 5 = Strongly agree.
Appendix B. Supplementary data [4] D. Arent, The role of renewable energy technologies in limiting climate
change, Bridge 40 (2010) 31–39.
[5] GTM Research, Solar Market Insight 2015 Q4, SEIA, 2016, Accessed 7 June
Supplementary data associated with this article can be found, in 2016 http://www.seia.org/research-resources/solar-market- insight-2015-q4.
the online version, at http://dx.doi.org/10.1016/j.erss.2016.12.023. [6] G. Barbose, N. Darghouth, Tracking the Sun VIII: The Installed Price of
Residential and Non-Residential Photovoltaic Systems in the United States,
Lawrence Berkeley National Laboratory, Berkeley, CA, 2015, Accessed 28
References January 2015 https://emp.lbl.gov/sites/all/files/lbnl-188238 1. pdf.
[7] E. Drury, M. Miller, C.M. Macal, D.J. Graziano, D. Heimiller, J. Ozik, T.D. Perry
IV, The transformation of southern California’s residential photovoltaics
[1] E. Drury, P. Denholm, R.M. Margolis, The solar photovoltaics wedge: pathways
market through third-party ownership, Energy Policy 42 (2012) 681–690,
for growth and potential carbon mitigation in the US, Environ. Res. Lett. 4
http://dx.doi.org/10.1016/j.enpol.2011.12.047.
(2009) 034010, http://dx.doi.org/10.1088/1748- 9326/4/3/034010.
[8] V. Rai, B. Sigrin, Diffusion of environmentally-friendly energy technologies:
[2] U.S. Department of Energy, SunShot Vision Study, U.S. Department of Energy,
buy versus lease differences in residential PV markets, Environ. Res. Lett. 8
2012.
(2013) 014022, http://dx.doi.org/10.1088/1748- 932/8/1/014022.
[3] P. Zhai, P. Larsen, D. Millstein, S. Menon, E. Masanet, The potential for avoided
[9] D. Chung, C. Davidson, R. Fu, K. Ardani, R. Margolis, U.S. Photovoltaic Prices
emissions from photovoltaic electricity in the United States, Energy 47 (2012)
and Cost Breakdowns: Q1 2015 Benchmarks for Residential, Commercial, and
443–450, http://dx.doi.org/10.1016/j.energy.2012.08.025.
Utility-Scale Systems, National Renewable Energy Laboratory, Golden, CO, [36] L. Korcaj, U.J.J. Hahnel, H. Spada, Intentions to adopt photovoltaic systems
2015, Accessed 28 January 2016 http://www.nrel.gov/docs/fy15osti/64746. depend on homeowners’ expected personal gains and behavior of peers,
pdf. Renew. Energy 75 (2015) 407–415, http://dx.doi.org/10.1016/j.renene.2014.
[10] B. Bollinger, K. Gillingham, Peer effects in the diffusion of solar photovoltaic 10.007.
panels, Mark. Sci. 31 (2012) 900–912, http://dx.doi.org/10.1287/mksc.1120. [37] M.C. Claudy, M. Peterson, A. O’Driscoll, Understanding the attitude-behavior
0727. gap for renewable energy systems using behavioral reasoning theory, J.
[11] P. Denholm, E. Drury, R. Margolis, Solar Deployment System (SolarDS) Model: Macromark. 33 (2013) 273–287, http://dx.doi.org/10.1177/
Documentation and Sample Results, National Renewable Energy Laboratory, 0276146713481605.
Golden, CO, 2009. [38] I. Ajzen, Values, attitudes, and behavior, in: S. Salzborn, E. Davidov, J. Reinecke
[12] M. Graziano, K. Gillingham, Spatial patterns of solar photovoltaic system (Eds.), Methods Theor. Empir. Appl. Soc. Sci. Festschr. Peter Schmidt, Springer
adoption: the influence of neighbors and the built environment, J. Econ. VS, 2012, pp. 33–38.
Geogr. 15 (2015) 815–839, http://dx.doi.org/10.1093/jeg/lbu036. [39] R. Haas, M. Ornetzeder, K. Hametner, A. Wroblewski, M. Hübner,
[13] C.L. Kwan, Influence of local environmental, social, economic and political Socio-economic aspects of the Austrian 200 kwp-photovoltaic-rooftop
variables on the spatial distribution of residential solar PV arrays across the programme, Sol. Energy 66 (1999) 183–191, http://dx.doi.org/10.1016/S0038-
United States, Energy Policy 47 (2012) 332–344, http://dx.doi.org/10.1016/j. 092X(99)00019- 5.
enpol.2012.04.074. [40] W. Jager, Stimulating the diffusion of photovoltaic systems: a behavioural
[14] J. Paidipati, L. Frantzis, H. Sawyer, A. Kurrasch, Rooftop Photovoltaics Market perspective, Energy Policy 34 (2006) 1935–1943, http://dx.doi.org/10.1016/j.
Penetration Scenarios, National Renewable Energy Laboratory, Golden, CO, enpol.2004.12.022.
2008. [41] J. Jansson, A. Marell, A. Nordlund, Green consumer behavior: determinants of
[15] D. Noll, C. Dawes, V. Rai, Solar Community Organizations and active peer curtailment and eco-innovation adoption, J. Consum. Mark. 27 (2010)
effects in the adoption of residential PV, Energy Policy 67 (2014) 330–343, 358–370, http://dx.doi.org/10.1108/07363761011052396.
http://dx.doi.org/10.1016/j.enpol.2013.12.050. [42] E.H. Noppers, K. Keizer, J.W. Bolderdijk, L. Steg, The adoption of sustainable
[16] V. Rai, D.C. Reeves, R. Margolis, Overcoming barriers and uncertainties in the innovations: driven by symbolic and environmental motives, Glob. Environ.
adoption of residential solar PV, Renew. Energy 89 (2016) 498–505, http://dx. Change 25 (2014) 52–62, http://dx.doi.org/10.1016/j.gloenvcha.2014.01.012.
doi.org/10.1016/j.renene.2015.11.080. [43] J. Leenheer, M. de Nooij, O. Sheikh, Own power: motives of having electricity
[17] V. Rai, S.A. Robinson, Effective information channels for reducing costs of without the energy company, Energy Policy 39 (2011) 5621–5629, http://dx.
environmentally- friendly technologies: evidence from residential PV doi.org/10.1016/j.enpol.2011.04.037.
markets, Environ. Res. Lett. 8 (2013) 014044, http://dx.doi.org/10.1088/1748- [44] P.C. Stern, T. Dietz, G.A. Guagnano, A brief inventory of values, Educ. Psychol.
9326/8/1/014044. Meas. 58 (1998) 984–1001, http://dx.doi.org/10.1177/
[18] C. Schelly, Residential solar electricity adoption: what motivates, and what 0013164498058006008.
matters? A case study of early adopters, Energy Res. Soc. Sci. 2 (2014) [45] J. Jansson, Consumer eco-innovation adoption: assessing attitudinal factors
183–191, http://dx.doi.org/10.1016/j.erss.2014.01.001. and perceived product characteristics, Bus. Strategy Environ. 20 (2011)
[19] P.C. Stern, I. Wittenberg, K.S. Wolske, I. Kastner, Household production of 192–210, http://dx.doi.org/10.1002/bse.690.
photovoltaic energy: issues in economic behavior, in: A. Lewis (Ed.), Camb. [46] K.C. Manning, W.O. Bearden, T.J. Madden, Consumer innovativeness and the
Handb. Psychol. Econ. Behav., Cambridge University Press, 2017, in press. adoption process, J. Consum. Psychol. 4 (1995) 329–345, http://dx.doi.org/10.
[20] E.M. Rogers, Diffusion of Innovations, 5th edition, Simon and Schuster, 2003. 1207/s15327663jcp0404 02.
[21] I. Kastner, P.C. Stern, Examining the decision-making processes behind [47] L.E. Ostlund, Perceived innovation attributes as predictors of innovativeness,
household energy investments: a review, Energy Res. Soc. Sci. 10 (2015) J. Consum. Res. 1 (1974) 23–29, http://dx.doi.org/10.1086/208587.
72–89, http://dx.doi.org/10.1016/j.erss.2015.07.008. [48] J.S. Black, P.C. Stern, J.T. Elworth, Personal and contextual influences on
[22] I. Ajzen, The theory of planned behavior, Organ. Behav. Hum. Decis. Process. househould energy adaptations, J. Appl. Psychol. 70 (1985) 3–21, http://dx.
50 (1991) 179–211. doi.org/10.1037/0021-9010.70.1.3.
[23] P.C. Stern, T. Dietz, T. Abel, G.A. Guagnano, L. Kalof, A value-belief-norm [49] K.K. Chen, Assessing the effects of customer innovativeness, environmental
theory of support for social movements: the case of environmentalism, Hum. value and ecological lifestyles on residential solar power systems install
Ecol. Rev. 6 (1999) 81–97. intention, Energy Policy 67 (2014) 951–961, http://dx.doi.org/10.1016/j.
[24] L. Steg, Values, norms, and intrinsic motivation to act proenvironmentally, enpol.2013.12.005.
Annu. Rev. Environ. Resour. 41 (2016) 277–292, http://dx.doi.org/10.1146/ [50] D. Noll, C. Dawes, V. Rai, Solar Community Organizations and active peer
annurev-environ-110615- 085947. effects in the adoption of residential PV, Energy Policy 67 (2014) 330–343,
[25] T. Dietz, Environmental values, in: T. Bosch, D. Sander (Eds.), Oxf. Handb. http://dx.doi.org/10.1016/j.enpol.2013.12.050.
Values, Oxford University Press, Oxford, 2015, pp. 329–349. [51] M.P. Vandenbergh, P.C. Stern, G.T. Gardner, T. Dietz, J.M. Gilligan,
[26] J. Jansson, A. Marell, A. Nordlund, Exploring consumer adoption of a high Implementing the Behavioral Wedge: Designing and Adopting Effective
involvement eco-innovation using value-belief-norm theory, J. Consum. Carbon Emissions Reduction Programs, Social Science Research Network,
Behav. 10 (2011) 51–60, http://dx.doi.org/10.1002/cb.346. Rochester, NY, 2010, Accessed 25 February 2016 http://papers.ssrn.com/
[27] C.A. Klöckner, A comprehensive model of the psychology of environmental abstract=1617426.
behaviour—a meta-analysis, Glob. Environ. Change 23 (2013) 1028–1038, [52] D.M. Gromet, H. Kunreuther, R.P. Larrick, Political ideology affects
http://dx.doi.org/10.1016/j.gloenvcha.2013.05.014. energy-efficiency attitudes and choices, Proc. Natl. Acad. Sci. 110 (2013)
[28] State of California, California Solar Statistics. https://www. 9314–9319, http://dx.doi.org/10.1073/pnas.1218453110.
californiasolarstatistics.ca.gov/archives/nem currently interco nnected/, 2016 [53] K.A. Frank, S.J. Maroulis, M.Q. Duong, B.M. Kelcey, What would it take to
(Accessed 10 March 2016). change an inference? Using Rubin’s causal model to interpret the robustness
[29] U.S. Census Bureau, QuickFacts: California. www.census.gov/quickfacts/table/ of causal inferences, Educ. Eval. Policy Anal. 35 (2013) 437–460, http://dx.doi.
PST045215/06, 2015 (Accessed 22 March 2016). org/10.3102/0162373713493129.
[30] D.G. Labay, T.C. Kinnear, Exploring the consumer decision process in the [54] A.D. Henry, T. Dietz, Information, networks, and the complexity of trust in
adoption of solar energy systems, J. Consum. Res. 8 (1981) 271–278. commons governance, Int. J. Commons 5 (2011) 188, http://dx.doi.org/10.
[31] A. Faiers, C. Neame, Consumer attitudes towards domestic solar power 18352/ijc.312.
systems, Energy Policy 34 (2006) 1797–1806, http://dx.doi.org/10.1016/j. [55] D. Sleeth-Keppler, R. Perkowitz, M. Speiser, It’s a matter of trust: American
enpol.2005.01.001. judgments of the credibility of informal communicators on solutions to
[32] Y. Heath, R. Gifford, Extending the theory of planned behavior: predicting the climate change, Environ. Commun. 0 (2015) 1–24, http://dx.doi.org/10.1080/
use of public transportation, J. Appl. Soc. Psychol. 32 (2002) 2154–2189, 17524032.2015.1062790.
http://dx.doi.org/10.1111/j.1559-1816.2002.tb02068.x. [56] K. Nuttavuthisit, J. Thøgersen, The importance of consumer trust for the
[33] P. Harland, H. Staats, H.A.M. Wilke, Explaining proenvironmental intention emergence of a market for green products: the case of organic food, J. Bus.
and behavior by personal norms and the theory of planned behavior, J. Appl. Ethics (2015) 1–15, http://dx.doi.org/10.1007/s10551-015-2690-5.
Soc. Psychol. 29 (1999) 2505–2528, http://dx.doi.org/10.1111/j.1559-1816. [57] S. Weber, Debate Over Solar Rates Simmers in the Nevada Desert, PBS
1999.tb00123.x. NewsHour, 2016, Accessed 23 October 2016 http://www.pbs.org/newshour/
[34] H. Han, L.-T. (Jane) Hsu, C. Sheu, Application of the Theory of Planned bb/debate-over-solar-rates-simmers-in-the- nevada-desert/.
Behavior to green hotel choice: testing the effect of environmental friendly [58] J. Applegate, Santa Clara Passes Feed-in Tariff Policy for Solar Power, St George
activities, Tour. Manag. 31 (2010) 325–334, http://dx.doi.org/10.1016/j. News, 2015, Accessed 23 October 2016 https://www.stgeorgeutah.com/news/
tourman.2009.03.013. archive/2015/07/09/jla-santa-clara- passes-feed-tariff-solar-power/.
[35] M. Greaves, L.D. Zibarras, C. Stride, Using the theory of planned behavior to
explore environmental behavioral intentions in the workplace, J. Environ.
Psychol. 34 (2013) 109–120, http://dx.doi.org/10.1016/j.jenvp.2013.02.003.

Energy Research & Social Science: Kimberly S. Wolske, Paul C. Stern, Thomas Dietz

Uploaded by

Copyright:

Available Formats

Energy Research & Social Science: Kimberly S. Wolske, Paul C. Stern, Thomas Dietz

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Energy Research & Social Science: Kimberly S. Wolske, Paul C. Stern, Thomas Dietz

Uploaded by

Copyright:

Available Formats

Energy Research & Social Science 25 (2017) 134–151

Contents lists available at ScienceDirect

Energy Research & Social Science

Original research article

Explaining interest in adopting residential solar photovoltaic systems

Fig. 1. Simple schematic representations of DOI, TPB, and VBN theories.

VBN Direct Total Direct Total Direct Total Direct Total

Social curiosity 0.44*** 0.44***

IT SC Relative Advantage Riskiness Trialability

Social curiosity 0.32*** 0.32***

TPB Direct Total Direct Total

Social curiosity 0.33*** 0.33***

Perceived behavioral control

Table 3 predictor. Altruism and awareness of consequences had signiﬁcant

Fig. 4. Proposed integrated framework for predicting interest in pursuing RPV.

Independent Variables Direct Total Direct Total

Social curiosity 0.29*** 0.29***

RPV-speciﬁc beliefs & attitudes

Factor loading Mean (SD)

Self-interest (␣ = 0.63) 2.91 (0.81)

Traditionalism (␣ = 0.62) 4.24 (0.63)

Openness to Change (␣ = 0.79) 3.52 (0.82)

Personal norm (␣ = 0.83) 3.48 (0.86)

Factor loading Mean (SD)

Consumer Independent Judgment Making (␣ = 0.83) 2.38 (0.75)

Trust in PV Industry (␣ = 0.88) 2.55 (0.82)

Trust in Social Network (␣ = 0.81) 3.68 (0.72)

Characteristics of the innovation

Relative Advantage (␣ = 0.90) 3.38 (0.77)

Factor loading Mean (SD)

Riskiness (␣ = 0.63) 2.66 (0.80)

Trialability (␣ = 0.70) 3.82 (0.63)

Factor loading Mean (SD)

Environmental Beneﬁts (␣ = 0.87) 3.49 (0.88)

Perceived Risks (␣ = 0.63) 2.66 (0.80)

Expense Concerns (␣ = 0.63) 3.61 (0.72)

Waiting for Improvements (␣ = 0.65) 3.27 (0.69)

Perceived behavioral control

Might Move 3.35 (1.17)

You might also like

Social curiosity 0.44* 0.44*

Social curiosity 0.32* 0.32*

Social curiosity 0.33* 0.33*

Social curiosity 0.29* 0.29*