1

Adopting green innovations: A consumer practice-perspective

Abstract

Firms are increasingly innovating to offer greener products and services. Existing literature has

identified barriers and drivers of such consumer choice within a traditional understanding of

how innovations are adopted. This paper argues that the adoption of green innovations should

be understood as the gradual adoption of interrelated green consumer practices. These practices

involve more than the individual decision maker, encompass a change in habits and routines,

and take place over time. We investigate the adoption of green consumer practices by shedding

light on the case of zero waste (ZW) shopping practices. In three empirical investigations – two

qualitative studies and a field survey of consumers – we reveal barriers and drivers of the

adoption of green innovation, and the clustering of consumer practices. Our findings further the

understanding of green consumer practices in consumer goods markets, and the barriers and

drivers associated with such practices.

KEY WORDS:

Consumer behavior; Innovation; Practices; Sustainability

2

INTRODUCTION

Due to growing awareness about environmental problems, firms increasingly address

sustainability concerns in their strategies and business models. Sustainability innovations can

improve the social and environmental footprints in the way companies design, produce,

distribute, and sell products and services. This is not least the case for fast-moving consumer

goods companies, which produce vast amounts of single-use plastics and other packaging.

Greenpeace (2018) reports that “[f]orty percent of all plastics made in 2015 were used in

packaging, the largest of all markets for plastics” Consequently, consumer goods companies

increasingly innovate product and service design and distribution to lower their footprint.

Importantly, such changes in the production and distribution of products often require

comprehensive changes in consumer behavior and practices. According to practice theory

(Warde, 2005), practices are characterized as a routinized type of behavior. When companies

introduce new solutions (e.g. products, services, or distribution systems), they need to facilitate

consumers’ adoption of such practices. Scholars have previously addressed the inability of

existing adoption models to consistently predict green consumption practices (Perera, Auger, and

Klein, 2018). In this paper, we study the adoption of consumer practices related to zero waste

(ZW) shopping. Such shopping implies packaging-free solutions that require consumers to use

reusable bags and containers for refilling and carrying items. The adoption of such practices

usually requires considerable changes in the shopping habits of individuals and households, both

in the store and in the home.

ZW has been defined as “a goal [….] to guide people in changing their lifestyles and

practices to emulate sustainable natural cycles [….]” (Zaman 2015, p. 2). ZW shopping and

lifestyles represent a radical alternative to mainstream shopping and consumer behavior. This

3

means that adapting to a ZW lifestyle requires changes at many different levels. First, it requires

considerable behavioral changes at an everyday basis. Changing habits is challenging, and it

often requires substantial self-control and will power (e.g., Verplanken and Wood 2006). Even if

consumers are motivated and positive to a lifestyle change, they generally exhibit considerable

status quo bias and psychological inertia during decision-making.

Moreover, changes of everyday habits and routines influence and involve other actors in

the consumer’s micro-ecosystems. Adopting ZW consumer practices will typically involve

multiple members of a household – it goes beyond an individual decision maker conducting a

consumption decision. For example, waste management in a household may involve all members

in establishing new routines for recycling or composting kitchen waste (Grønhøj, 2006). This

implies that the subject of adoption is often not limited to an individual consumer. In an

ecosystem perspective: “(…) consumer motivations are often embedded in a variety of ordinary,

routine and habitualised behaviours which are themselves heavily influenced by social norms

and practices and constrained by institutional contexts” (Jackson, 2005, p.18).

Since ZW shopping is a matter of lifestyle, the object of adoption takes the form of an

abstract idea (i.e. a sustainable lifestyle), rather than a physical object (a store or a product). ZW

shopping involves new activities and routines, inside and outside the home, which at an

accumulated level constitute the practice of ZW shopping. Consumers adopt such practices

gradually over time, and as such, adoption should arguably be understood as a temporal process

rather than a discrete decision at a given point in time. One might, however, question whether

existing models of adoption are suitable for explaining the adoption of new practices (Perera et

al., 2018). In established adoption models (e.g. theory of planned behavior and theory of

technology acceptance), the object of adoption is a product, service, or technology, whereas

4

consumer practices, such as ZW, cannot be delimited in a similar way, but needs to be

understood as more complex adoption processes. Hence, we argue that existing perspectives on

adoption of innovations must be updated, taking into account the complex nature of the object of

adoption of new green practices.

In this article, we adhere to the idea that consumption is comprised of routines and habits

that are embedded in networks of everyday practices (Phipps and Ozanne, 2017). Consequently,

it is important to study how consumption practices change in response to companies’ innovative

solutions to environmental problems. Hence, the company’s offering (e.g. a ZW store or a

packaging-free product) is not regarded as the object of adoption, but as a property that initiates a

series of practices. When studying the adoption of green practices for ZW shopping solutions, we

need to understand the whole set of practices consumers need to change or adopt. Based on two

qualitative studies and three field-surveys, we identify four distinct domains of ZW practices: 1)

shopping practices, 2) practices in the household 3) social practices, and 4) general

environmental practices. This research investigates these practices among current users (i.e. early

adopters), potential users (i.e. transitional adopters), and non-users of ZW solutions.

This paper is structured as follows: First, we outline relevant insights from existing

research, including traditional adoption research and literature on the adoption of green products,

services and practices. Second, we present the methodology and results of our three studies – two

qualitative studies and a field survey. Finally, we discuss our findings and outline their

theoretical and practical implications.

LITERATURE REVIEW

Lessons from Adoption Research

5

In the academic fields of information systems research (IS), human-computer interaction (HCI),

and consumer behavior (CB), there are large volumes of research on the determinants of

individual consumers’ adoption of various technologies, products and services. Through the

theoretical lenses of the technology acceptance model (Davis et al. 1989), the unified theory of

acceptance and use of technology (Venkatesh et al., 2012), theory of planned behavior (Ajzen

1991), and theory of trying (Bagozzi et al., 1992), a large body of individual studies and meta-

studies (cf. Schepers and Wetzels, 2007; Blut et al., 2016) inform researchers about the factors

that drive adoption intentions and behaviors.

Also, through various applications and extensions of these adoption models, a wide array

of new adoption antecedents have been identified, including various network factors (Steiner et

al., 2016; Thorbjørnsen et al., 2009), self-confidence (Chaouali et al., 2017), social- and identity-

related factors (Thorbjørnsen et al., 2007), and the role of complementary products (Cenamor et

al., 2013). These new antecedents are important also for understanding adoption behavior in

other settings, because they tap into important factors pertaining to the consumers’ social

network, the role of identity expressiveness, and the role of the extended network of

complementary products, services and platforms available to consumers. For the adoption of

green products and practices, these factors are obviously relevant.

However, what traditional adoption research within IS, HCI and CB have in common is a

narrow focus on the single individual consumer as the subject of adoption, and the single

technology or products/services as the object of study. With the exception of the effects of social

norms, traditional technology adoption models treat both the consumer and the adopted

technology as islands, free of influence from their environment and the complex set of networks

and services surrounding them. Noticeable exceptions do exist though, and more recent

6

investigations of technology adoption explicitly deal with the complex dependencies between

networks, platforms and complementary products (cf. Steiner et al., 2016). The complex markets

and settings described by Steiner et al. (2016) and others seem to have much in common with the

heterogeneous markets, -platforms and value chains that characterize green consumption.

Still, as the object of study in consumer adoption studies is limited to the technology,

product or service itself and the subject to the individual consumer, one has to look outside the

IS, HCI and CB literatures to get a more nuanced and realistic perspective on the adoption

process. A natural place to look is the organizational innovation adoption literature (cf.

Frambach and Schillewaert, 2002; Mirvis et al., 1991; Martin et al., 2016). Organizational

contexts are richer, and in some aspects perhaps also more similar to, the complex network of

actors to which a ZW consumer has to relate. For instance, in deciding to adopt ZW shopping,

the consumer may have to convince and onboard the rest of the family or household, much in the

same manner as a manager needs to onboard his/her co-workers in adopting a new technology.

As described in institutional theory, the behaviors of individuals within organizations

(such as within a household) are significantly influenced by prevailing norms, values and culture

(e.g. Scott 2008). Adopting ZW shopping may therefore be far more difficult for the consumer if

the remaining parts of the family goes ‘on strike’. Also, by the same token, the ZW consumer has

to change habits and ‘work’-processes, in the same manner as do new technology adopters in

organizations. Hence, key theories of adoption at the firm level, including the

technology-organization-environment (TOE) framework and institutional theory might be

fruitful to apply in understanding new green consumption practices. Trying to map out and

understand the organizational and environmental contexts of the adoption seems particularly

7

relevant for adoption contexts wherein the individual adopter is dependent on and influenced by

others (Tornatzky and Fleischer, 1990; Martin et al., 2016).

Adoption of green practices

Object of adoption

There is an extensive literature on barriers and drivers of individuals’ adoption of sustainability-

oriented goods and services. The object of adoption in these studies is sustainable offerings, such

as low-carbon transport vehicles (Stryja and Satzger, 2018; Ozaki and Sevastyanova, 2011),

green electricity (Ozaki, 2011), smart grid technology (Toft and Thøgersen, 2015), green food

and household goods (Vermeir and Verbeke, 2006), and sustainable travel (Carasuk et al., 2016).

Other studies investigate barriers and drivers of more general categories of sustainable behaviors,

regardless of specific market offerings. These behaviors include waste behavior (Taylor and

Todd, 1997; Echegaray and Hansstein, 2017), energy use (e.g., Wang et al., 2014), sustainable

mobility behavior (Schoenau and Muller, 2017), and diets (e.g., Cadario et al., 2018; Roberto

and Kawachi, 2014). Some of the latter studies are relevant to our inquiry, because they deal

with the adoption of practices rather than offerings.

Although most studies do not take into consideration the complexity of interrelated

practices that needs to be changed or adopted, some scholars have argued that green

consumption should be studied more holistically, taking a practice perspective. For example,

Lim (2017, p. 70) discusses the need for a more holistic approach to the study of sustainable

consumption: “Although academics and practitioners from various disciplines (…) have

explored ways to encourage consumers to choose more sustainable products, scholarship still

lacks understanding of how to encourage more sustainable patterns of consumption, especially

8

for the society at large”. Shove (2003, p. 395) argues that green consumption “is bound up with

routine and habit and with the use as much as the acquisition of tools, appliances, and household

infrastructures”. According to this perspective, the object of adoption is physical offerings or

infrastructures, but the consumption patterns reflected in everyday routines and habits. Since

sustainable consumption is manifested in an interrelated set of consumer choices (e.g. electrical

vehicles, sustainable food, etc.), consumption habits (e.g. reusable shopping bags, refill

solutions, asset sharing etc.), post-consumption and lifestyle habits (e.g. recycling, reuse, etc.),

green consumption is arguably a complex set of practices and lifestyle habits.

Subject of adoption

Most studies on adoption of green offerings investigate the likelihood of a single consumer’s use

of a product or service at a discrete point in time. However, some scholars focus on the network

of actors that are involved in sustainable consumption. For example, Grønhøj (2006, p. 491)

criticizes the individualistic approach in research on sustainable consumer practices, as “many

green consumer practices involve more than one member of a household, who may suggest,

support, question, oppose, or in other ways influence household participation in these practices.”

Therefore, to carry out the lifestyle changes, “several family members need to agree on changing

established consumption habits” (Grønhøj 2006, p. 492).

In a sharing/collaboration economy perspective, adoption of new behavior occurs in a

network of peer-to-peer interactions. An important motivational factor for consumers to engage

in sharing practices is to maintain a more sustainable lifestyle (Hamari et al., 2016). Examples

include food-sharing practices (Morone and Navia, 2018), car sharing (Bardhi and Eckhardt,

2012) and asset sharing and access-based consumption more broadly (Belk, 2014). Whereas

9

studies on adoption of sharing practices per definition account for practices involving more than

one consumer, they generally do not investigate the broader range of practices involved in using

a service. Put differently, the literature on the sharing economy accounts for the network of

actors, but not for the network of practices. Like for sharing services (such as car-sharing or

other asset sharing), the adoption of ZW-consumption involves behavioral change and new

practices also for other actors of the consumers’ micro eco-system (i.e. household, friends,

family). Adoption models should thus also incorporate these factors when trying to explain and

predict adoption decisions.

Adopting social practices

Social practices are facilitated by three elements: 1) material things (e.g. technology or products),

2) motivations and emotions (embedded in social meanings, values and norms), and 3) know-

how and competence (Jaeger-Erben et al. 2015). When companies innovate to facilitate

sustainable consumption, then, they aim to facilitate “alternative practices or new variations of

practices which differ substantially from established or mainstream routines” (Jaeger-Erben et

al., 2015, p. 785). Hence, an evolutionary process underlies the adoption of new green practices,

and in order to better understand the adoption of such practices, these factors need to be taken

into account.

Jackson (2005) furthermore argues that understanding the adoption of practices implies

going beyond models of individual agency and instead understanding the collective nature of

such practices. As pointed out by Jackson (ibid.), such an understanding also emphasizes the

existence and the importance of social norms in human behavior. Crucially, adoption then needs

10

to be understood in a social context, embedded in an ecosystem of actors in and outside the

household.

Taken together, then, the adoption of green consumer practices involves decision-making

and behavior change that goes beyond the individual and is instead interpersonal and social; it

encompasses a behavior more complex and multifaceted than the mere choice of using a single

product or service; and it involves an intertemporal process of gradual and increasing change in

consumption and post-consumption habits in the store and in the home. As outlined above,

existing models of adoption only partially captures such aspects of the adoption process of green

consumer practices.

In the current paper, we aim at developing an adoption model that a) integrates new

consumer practices as drivers of behavioral intentions, b) takes into account practices that also

are social and involve other members of the household, c) can be applied at different stages of

adoption (pre, during and post) and d) builds on established theoretical relationships from

adoption research (i.e. the theory of planned behavior). In doing this, we first report on two

qualitative studies to identify and investigate green consumption practices. The first study

investigates the perceptions of the green practices associated with ZW shopping from the

perspective of focus groups of consumers in four different segments. The second study

investigates the perceptions of such green practices associated with ZW shopping from the

perspective of ZW store-owners. Using insight from these two qualitative studies, we develop an

adoption model of green practices and empirically test this through three related field studies

(Study 3). These field studies are conducted among non-users, transitional (potential) users and

early adopters (current users) respectively, to try to capture the intertemporal processes of green

consumption practices adoption.

11

STUDY 1

The purpose of study 1 was to investigate consumer perceptions of ZW solutions and the drivers

and barriers associated with adopting such shopping practices. We conducted focus group

interviews with 20 consumers in four different segments on their perceptions of various refill-

based solutions when shopping for consumer goods. Focus groups allow for data collection

through group interaction on a topic determined by the researcher (Morgan, 1996; Kitzinger,

1995), and for identifying experiences and perspectives that can be explored in more depth using

additional methods (Stewart and Shamdasani, 1990). We conducted four focus groups in Bergen,

Norway, in the spring of 2018. Four segments were included in the study: young female adults

(aged 22-23), young male adults (aged 25-27), adults with children (aged 30-45) and middle-

aged and elderly female adults (aged 57-75).

We selected soap and detergent products in single-use packaging as the case for the focus

groups. As part of the conversation, focus group participants were exposed to four different

prospective distribution solutions that can reduce the plastic footprint of the consumption and use

of such products. These solutions are prospective solutions developed by a corporate partner in

the research project from which this data is derived – a large Norwegian FMCG company. The

four solutions are illustrated in Figure 1, and are, respectively, (1) a big-bag product for home

refill, (2) a refill station in the store, (3) home delivery of refill via e.g. Amazon Key, and (4)

home delivery of refill via online grocery shopping. We structured the conversation around the

characteristics and practicalities of the selected product category, which made it easier to

compare the findings across focus groups (cf. Morgan, 1996).

[ADD FIGURE 1 HERE]

12

Results

Generally, the participants did not spend much time and effort on purchase decisions for liquid

soaps and detergents, and their shopping behavior was habit-driven. A “hassle-free” shopping

experience was generally regarded as the main concern, although there was some awareness

related to “the plastic problem” of using single-use containers.

Participants were exposed to the various refill solutions outlined in Figure 1, and

emphasized price, environmental friendliness, effort, and the functionality of the solutions.

Current refill packages were judged as being too expensive, when taking into account the effort

required by consumers to refill the containers at home. Many participants experienced making a

mess when trying to refill the containers, and saw them as requiring different behavior when

preparing for, and carrying out, shopping activities.

Several participants argued that bringing containers to the store for refill would be very

difficult to do in practice. One parent explained that “it would be like the reusable grocery bags

that they buy, but keep forgetting at home”. Female students and the senior citizens were less

concerned with this, i.e. there was variation in the assessment of difficulty and effort.

Many participants across segments expressed an unwillingness to change current

shopping practices, and there was little interest for refill solutions in the store or comparable

solutions. Participants were moreover skeptical their functionality: “They are not going to be

easy enough to use”, argued a 30-year old female. Several adult consumers feared making a mess

and the danger of mixing products. Such functional risks in the household were emphasized.

In addition, various types of physical risks were pointed out. Parents disliked storing

large quantities of soap products in the home, due to the risk for their children. Young female

students similarly felt that there were risks associated with home delivery. Such privacy concerns

13

were shared by many, except the young male students, who were more favorable. They however

emphasized that for such solutions to be attractive, it would need to be part of a larger

transformation of the way they shopped all types of product categories.

The insights from the focus groups revealed that consumers were skeptical to refill and

ZW shopping due to the lack of a coherent system for such shopping, i.e. the barriers were not

related to products (as objects), but to the broader ecosystem. The challenges related to

transportation, storage and waste handling from the store and in the house was also challenging.

Furthermore, the adoption of such solutions required coordination in the household, and

functional or privacy risks for household members beyond the “shopper” were emphasized.

Finally, since these products were habit-driven and alternative solutions required a larger

transformation of how shopping was organized, consumers found it burdensome. Thus, the

consumers expressed concerns that are relevant to our understanding of what is the object of

innovation, who is the adopter, and to what is the broader process of adoption of a ZW lifestyle.

STUDY 2

The purpose of study 2 was to investigate the perceptions of the ZW store owners of the nature

and characteristics of ZW shopping and barriers and drivers thereof. We conducted semi-

structured interviews with the owners of the three biggest ZW stores in Norway, Mølleren Sylvia

in Oslo, Råvarene in Bergen, and Unwrapped in Arendal. All of them are relatively small stores

with a limited scope of products – that is, customers cannot carry out their entire grocery

shopping in these stores alone. The interviews were conducted individually with each of the

three owners, henceforth referred to as Subject 1, 2 and 3. The interviews were conducted in the

fall of 2018 by two members of the research team – one interviewer and one note-taker.

14

Results

The store owners argued that the typical ZW customer was a relatively young female, who cares

about the environment and lives in the city. The early adopters shop almost all groceries in the

ZW store, are driven by environmental concerns and altogether avoid plastic. Subject 3 stated

that “the customers bring their own containers (...) and often stay for quite some time.”

Subject 3 emphasized that “...identity is absolutely important, it becomes more and more

important for people to communicate who you are [through shopping practices]”. Subject 2

described how “many people use this visit almost as a kind of ‘meditation’”. According to the

shop owners, ZW customers accept that groceries cost a little more, both in monetary terms and

in terms of time used. Subject 1 elaborated: “It requires a little more planning, and it is kind of a

barrier for customers to make all the food from scratch – it takes more time and is more

cumbersome in a hectic everyday life.” All subjects described early adopters as people who are

willing to make substantial changes in their habits and planning. Subject 2 claimed that “the

disadvantage [of ZW] is that you have to plan better. (...) There may also be a mental barrier,

because you have to bring more things”. In sum, she argued, “There are many steps you need to

take to get there, which might prevent people from trying in the first place.”

Subject 3 pointed out that ZW shopping is considered an “extreme” behavior, and her

goal was to make it “a common and widely accepted way of shopping.” Counterintuitively, she

argued that ZW can actually help consumers save both time and money, as long as you “get into

a good routine for how to shop and consume food”. For instance, Subject 3 argued, you can

reduce the waste generated in your household “when you only buy the amount you need”.

Subject 1 argued that ZW could become “mainstream”, but that change is happening

quite slowly. She emphasized the peer effects involved in new adopters being influenced by

15

early adopters, who are “proud to live like this.” She moreover pointed out that ZW stores are

also almost like “lifestyle hubs” or “information centers” for consumers who want to live a more

sustainable life.

The shop owners believed that there were technological barriers for ZW, which made it

more difficult to adopt. Subject 2 explained: “You must first weigh the empty jar so you don’t

have to pay for the weight of the container, then you must fill it up with the goods that you want,

and then you have to weigh it again and subtract the weight of the empty jar.” She predicted

scanning technology in the dispensers that identified the exact amount the customer had taken,

which would make it easier for consumers.

Overall, the store-owners described shopping practices that are still a niche phenomenon

of highly committed shoppers. However, they were seeing a steady expansion of the scope, i.e.

the type of customers who tried ZW solutions. They also believed that integrating ZW solutions

into “traditional” grocery stores could make it easier for consumers to adopt such practices, since

it would be in the context of their familiar shopping routines. The perceptions of these key

informants problematized whether the adoption of ZW necessarily becomes more burdensome,

but had a keen understanding of how such barriers led consumers to avoid ZW. This was due

both to the necessity of adopting planning and coordination practices that spanned from the store

to the home, and because adopting ZW requires a shift in mindset in order to be effective.

According to the informants, this relates to shopping, storing and making food in the house, and

managing waste, as well as the more symbolic and identity-related dimensions of adopting such

practices. Thus, they believed that rather than being a straightforward adoption of solutions in the

store, ZW implied a comprehensive transition into a new set of lifestyle practices.

16

STUDY 3

The purpose of study 3 was to develop and test a new adoption model of green consumption

practices. Based on existing research and the qualitative insights from study 1 and study 2, we

conceptualized five broad categories of consumer practices involved in ZW-shopping adoption;

1) shopping practices, 2) household practices for preparing and storing food, 3) household

practices for managing waste, 4) social practices and 5) general environmental practices. Both

actual and perceived changes in these five categories of practices likely influence consumers’

beliefs and adoption intentions. Moreover, we integrated these practices as drivers/antecedents in

a well-established adoption model, namely the theory of planned behavior (Ajzen, 1991).

Consumer practices will potentially influence key constructs in this model; attitudinal beliefs,

behavioral control and behavioral intentions. Our suggested adoption model is depicted in

Figure 2.

[ADD FIGURE 2 HERE]

Moreover, to try to capture the intertemporal and dynamic process of green practices

adoption, we empirically tested this model across three different samples: Non-adopters (i.e.

consumers with no experience with ZW shopping), transitional adopters (i.e. consumers who had

displayed an interest in ZW and environmentally sustainable shopping, but not adopted the

practice) and early adopters (i.e. current users of ZW shopping). The purpose of this was

twofold: First, to see if our model and measures fit both non-users, potential users, and actual

users. Second, to explore the temporal dimension of practices adoption by investigating

differences in model paths (driver pattern) across different stages of the adoption process. As

also argued by scholars in green adoption; non-users and actual users of an innovation likely

17

have different perceptions of the practices and behaviors involved in adoption, and these

perceptions will influence behavioral intentions. For instance, and as suggested by the two

qualitative studies; a non-user may envision the ZW-shopping process to be cumbersome and

unpractical, whereas an established users may perceive it as a pleasant experience. Consequently,

perceived changes in shopping practices may be a negative antecedent of behavioral intentions

for non-users, but a positive driver for existing users. By the same token, several of the perceived

barriers towards adoption revealed in study 1 above may apply to non-users and transitional

users, but not to actual users in which have integrated new practices to their everyday life.

We therefore conducted three related field surveys for non-adopters, transitional adopters

and early adopters. The surveys comprise nine overarching topics (see Table 1). First, we

developed measures of the five broad categories of consumer practices that could be expected to

be influenced by the adoption of ZW solutions. These were based in part on our review of prior

studies and in part on insights from the two qualitative studies. We asked participants to assess to

what extent they believed that ZW shopping would influence these practices. Second, we

included measures from the theory of planned behavior (Ajzen, 1991), capturing attitudinal

beliefs, social norm and behavioral intentions related to ZW shopping solutions. Finally, we

measured socio-demographic characteristics.

[ADD TABLE 1 HERE]

Sample characteristics

We conducted three field surveys in the fall of 2018. First, 171 current non-adopters of ZW

solutions were interviewed at two different shopping malls outside Norway’s capital Oslo. The

18

second sample consisted of 276 transitional adopters. We recruited these respondents online

through special interest forums for the Norwegian ZW community. Third, we surveyed 205 early

adopters of ZW solutions in one of the ZW stores mentioned above and through the same special

interests forums online. Respondents in the non-adopter study and the transitional adopter study

self-reported not to have adopted ZW shopping.

Note that the wording in the items in the surveys were slightly modified to account for

whether or not respondents had experience with ZW solutions or not (e.g. by changing the

present tense to the future tense in questions about adopting such practices; cf. Table 1). In total,

the sample consisted of 652 participants in three different groups. In the non-adopter group,

65.29% of respondents were female (mean age = 52, min = 18, max = 86), while the percentage

of female respondents was 85.71% for early adopters (mean age = 32, min = 17, max = 87), and

87.55% for transitional adopters (mean age = 34, min = 16, max = 74). Though the distribution

of age and gender seems to be different across our three groups, a robustness check shows that

age and gender have no significant effects on behavioral intentions and any other structural

relationships of our interest when they were included as control variables in our main model.

Results

Measurement model

As we are interested in substantive cross-group comparisons, we need to first test whether our

measurement model is invariant across groups (Kline, 2010). According to Steenkamp and

Baumgartner (1998), measurement invariance is tested at four different levels. The first level

involves the test for configural invariance to see whether we have the same factor patterns across

groups. In the second level, we test weak/metric invariance (i.e. invariance of factor loadings) to

19

see if the meaning of latent factors is similar across groups. In the third level, strong/scalar

invariance (i.e. invariance of factor loadings and item intercepts) is tested to check if the scores

on the latent constructs are comparable across groups. Finally, the test of strict measurement

invariance (i.e. invariance of factor loadings, item intercepts, and error variances) is implemented

to see if items are equally reliable across groups. Previous research suggests that if metric

invariance is met, it is valid and meaningful to further test structural invariance (e.g., comparing

structural relationships between groups) (e.g., Babin et al., 2016; Kline, 2010; Steenkamp and

Baumgartner, 1998).

Configural invariance

Following Shukla and Purani (2012), we first conduct confirmatory factor analysis (CFA) for

each group to check if our studied latent concepts are psychometrically valid within each group

of participants. Three different group-level CFA models are run for early adopters, transitional

adopters, and non-adopters using Lavaan package version 0.6-3 in R (Rosseel, 2012). The

maximum likelihood estimator robust to nonnormal and missing data (i.e., MLR estimator) are

used. Results show that the Satorra-Bentler’s (SB) scaled chi-square values are significant for all

groups, giving preliminary evidence against the model. However, given that the is often

overly sensitive to sample size, we focus on other approximate fit indices instead (Bagozzi and

Yi 2012). As shown in Table 2, the scores of CFI (Comparative Fit Index), TLI (Tucker Lewis

Index) are all above the recommended threshold value of .90 (Hair et al. 2010; Shukla and

Purani, 2012). In addition, RMSEA (Root Mean Square Error of Approximation) and SRMR

(standardized root mean square residual) values are all below .07, suggesting adequate fit

20

between the model and the observed data (Bagozzi and Yi 2012; Hu and Bentler 1998; 1999).

Overall, the results suggest that our measurement model is valid for all three groups.

[ADD TABLE 2 HERE]

Furthermore, all items significantly load on the expected latent variables and most of

their standardized factor loadings are all higher than .5 (see Table 2), providing high evidence for

convergent validity (Bagozzi and Yi, 1991; Hair et al., 2010). One exception is one item of

attitudinal beliefs in groups of transitional adopters (.47) and non-adopters (.49) that shows

moderately high and significant factor loadings. We keep this item to remain the full meaning of

the corresponding concept and avoid capitalization on chance (Xie et al., 2015). The values of

average variance extracted (AVE) range from .43 to .87, while those of composite reliability

(CR) range from .62 to .93, indicating high convergent validity. Following Bagozzi and Yi

(1991), we assess discriminant validity by checking whether any two latent constructs are

perfectly correlated, meaning that they are not discriminant. Previous research suggests that this

can be done by checking if the 95% confidence intervals of the correlation estimates for each

pair of latent constructs contain 1.0 (Sharma, 2010; Xie et al., 2015). An examination of the

correlation matrix between latent factors shows that all of our latent constructs meet this

requirement (see Appendix 1). In addition, we perform a series of Chi-square difference tests to

see if the unconstrained model is significantly better than a constrained model in which a specific

correlation coefficient between two latent constructs is set to 1. The results again confirm that

discriminant validity is achieved in our study.1

1 Details are available on request.

21

Next, we run a multi-group CFA constrained to test configural invariance. This M1

model provides adequate global fit measures: SB scaled 564 = 920.971, p < .001, CFI =

.935, TLI = .920, RMSEA = .054 (90% CI: [.048, .060]), and SRMR = .056. Thus, configural

invariance is supported.

Metric invariance

In this step, we constrain all factor loadings to be equal across groups in order to test whether we

can establish metric invariance with our data. This M2 model yields reasonably good global fit

measures: SB scaled 594 = 953.610, p < .001, CFI = .935, TLI = .924, RMSEA = .053

(90% CI: [.047, .059]), and SRMR = .060. To compare models M2 and M1, we conduct a scaled

chi-square difference test based on Satorra and Bentler (2001). Results show that the difference

between M2 and M1 is not significant: ∆ 30 = 33.25, p = .31 > .05, suggesting that the full

metric invariance model is not significantly worse than the configural invariance model (see

Appendix 2 for more details). In other words, metric invariance is supported, and it is valid to

compare structural relationships among latent constructs between groups.

Scalar invariance

We also attempt to test for scalar invariance by constraining the item intercepts to be equal

across groups (M3 model). However, the scaled chi-square difference test shows that the M3 is

significantly worse than M2: ∆ 30 = 145.65, p < .05 (see Appendix 2). In addition, the global

fit measures of this M3 model are also worse than M2: SB scaled 624 = 1094.451, p < .001,

CFI = .915, TLI = .905, RMSEA = .059 (90% CI: [.053, .064]), and SRMR = .066. Therefore,

22

we conclude that scalar invariance is untenable and it is not valid to compare means of latent

constructs across groups.

Structural model

As our measurement model is metric invariant across groups, we can examine if structural paths

are also invariant across groups. First, we perform a multi-group structural equation model

(SEM) in which all structural coefficients are freely estimated between groups. This “totally

free” structural model yields good global fit measures: SB scaled 708 = 1160.801, p < .001,

CFI = .921, TLI = .907, RMSEA = .054 (90% CI: [.049, .060]), and SRMR = .072. As shown in

Figure 3, we find significant positive effects of attitudinal beliefs on behavioral intentions in all

three groups. The positive effect of behavioral control on behavioral intentions is significant for

early adopters, and marginally significant for transitional adopters and non-adopters. The details

of model estimates are shown in Appendix 3.

[ADD FIGURE 3 HERE]

For early adopters, shopping practices, household practices (food), and general

environmental practices have significant direct effects on behavioral intentions. Moreover, social

practices and general environmental practices have significant indirect effects on behavioral

intentions through attitudinal beliefs, while only general environmental practices have significant

indirect effect on behavioral intentions through behavioral control. The fact that shopping

practices have a positive effect on behavioral intentions for those who already have adopted is

interesting, and in contrast to both transitional adopters and non-adopters (see below). Whereas

23

early adopters seemingly have integrated these new shopping habits into their daily life to the

extent that it positively drives behavioral intentions, for non-users the same model path is

negative. Hence, and as mentioned above, non-users likely perceive ZW shopping practices as

cumbersome and unfamiliar, and this is reflected in these practices as a negative antecedent of

behavioral intentions. For early adopters, household practices (food) and social practices have,

respectively, direct and indirect (via attitudinal beliefs) negative effects on behavioral intentions.

Although speculative at this point, this finding may suggest that ZW shoppers have a hard time

onboarding the rest of their household to the ZW practices pertaining to food storage, preparation

and cooking. Evidently, for early adopters, there is something pertaining to household practices

and social practices that leads to a negative influence on intentions of continued ZW

consumption.

For transitional adopters, no practices have significant direct effect on behavioral

intentions. However, both social practices and general environmental practices have significant

indirect effects on behavioral intentions through attitudinal beliefs and behavioral control.

Similar to the early adopters, the indirect effect of social practices on behavioral intentions is

negative, suggesting that perceptions related to changes in household social coordination induced

by ZW shopping is a show-stopper for their adoption intentions. General environmental practices

are on the other hand the single strong positive driver of behavioral intentions for these

transitional adopters.

For non-adopters, shopping practices have a marginally negative significant direct effect

on behavioral intentions, while the direct effect of general environmental practices is strongly

significant and positive. Similarly, only shopping practices and general environmental practices

have significant indirect effects on behavioral intentions through attitudinal beliefs. In addition,

24

general environmental practices have significant indirect effect on intentions through behavioral

control. Finally, the effect of social norm on intentions is only significant for transitional

adopters, while it is insignificant for early adopters and non-adopters.

Next, we perform a fully constrained structural model in which all structural parameters

(regression coefficients) are set to be equal across groups. This model yields worse global fit

measures: SB scaled 744 = 1225.034, p < .001, CFI = .916, TLI = .906, RMSEA = .055

(90% CI: [.049, .060]), and SRMR = .078. More importantly, the scaled chi-square difference

test shows that the fully constrained structural model is significantly worse than the “totally free”

one: ∆ 36 = 63.75, p < .01. In other words, the structural paths are not fully invariant

between groups, suggesting that some relationship(s) might be moderated by the grouping

variable (i.e., early adopters, transitional adopters, and non-adopters).

We further check for this by constraining structural parameters one by one and only for

two groups at a time. The constrained model is then compared against the totally free structural

model. We compare 18 relationships for each pair of groups and only five of them have

significant SB scaled chi-square difference test with 1 degree of freedom. In particular, as shown

in Table 3, results show that the effect of shopping practices on behavioral intentions is positive

and stronger for early adopters (b = .51, p <.01) than for non-adopters (b = -.24, p = .06 < .10).

The effect of shopping practices on attitudinal beliefs is however stronger and negative for non-

adopters (b = -.31, p < .05) than for both early adopters (b = .10, p > .10) and transitional

adopters (b = .31, p > .10). In addition, the effect of social practices on behavioral control is

negative and stronger for transitional adopters (b = -.40, p < .01) than for non-adopters (b = -.17,

p > .10). Finally, the effect of social norm on behavioral intentions is positive and larger for

transitional adopters (b = .07, p < .05) than for early adopters (b = -.03, p > .10).

25

[ADD TABLE 3 HERE]

GENERAL DISCUSSION

This study offers new perspectives on how consumers’ adoption of green innovations could be

studied. First, while established theories of adoption assumes that the object of adoption is a

technology, object, or service, the current research proposes that consumer practices should be

regarded as the focal object of adoption. Our suggested adoption model explicitly conceptualizes

changes in consumer practices as antecedents to the traditional TPB variables. Second,

traditional adoption models do not take into account the complex role of social networks in the

adoption processes. Obviously, many adoption decisions influence the lives of other members of

the household (or organization), and adoption models should explicitly take this into account.

The findings in study 3 on negative influences on adoption intentions from perceived

changes in household- and social practices clearly point to the importance of social networks in

adoption decisions. We have argued that adoption practices should be understood not in

isolation, but in relation to other individuals’ attitudes and behaviors. Thus, the subject of

adoption is not an individual consumer, but a network of actors. Finally, adoption of green

practices is likely to be a function of time. Therefore, instead of studying adoption in a discrete

point in time, adoption should be regarded as a process of implementing new habits and routines

(i.e. practices). In trying to accommodate this, we have surveyed the different phases of

adoption, from non-adoption, via transitional adoption, to early adoption – and tested model fit

and -interpretation across phases.

Our study demonstrates that a consumer practice-perspective is a useful approach to

understanding adoption of green innovations, like ZW shopping. The focus group interviews

26

(study 1) showed that barriers and drivers were related to various aspects of the adoption process.

The study showed that consumers had concerns related to the influence of new shopping

practices on the broader set of actors in the household. Moreover, they believed that for refill (or

ZW) solutions to be attractive, a broader set of solutions would have to be available and adopted

in concert. This led the participants to believe that ZW was something to potentially adopt in the

future, when such comprehensive solutions were available. The interview study (study 2)

suggested that ZW is still a relatively niche phenomenon, with a dedicated subculture of people

embracing the lifestyle. Study 2 moreover showed that the dedicated ZW shoppers had embraced

new practices, but also an overarching mindset that led them to change the nature of their

shopping as well as household practices, related to cooking, waste handling and beyond. Since

such lifestyle changes are less likely for less enthusiastic shoppers, the findings suggest that the

mainstream adoption of ZW would require innovation in the ZW solutions, technologies and the

ecosystem of products and services that could reduce the burden on the “ordinary” consumer.

Informed by the findings from the first two studies, the field surveys from Study 3 lend

empirical support to the proposed green practice adoption model. Multi-group invariance testing

shows that the parameters of our measurement model are equivalent across the three groups:

early adopters, transitional adopters, and non-adopters. Hence, we can conclude that our

measurement model is valid, and that the structural model is replicable across settings (Chin et

al., 2014). Results therefore confirm that consumers distinguish between a set of various

practices in relation to ZW shopping, and that these practices influence the intention to (continue

to) use ZW, either directly, or indirectly through attitudinal beliefs about ZW and/or behavioral

control.

27

The fact that there are differences between the three groups in terms of how the practice

categories influence attitudinal beliefs, behavioral control and intentions supports the idea that

adoption should be regarded as a process that occurs and changes over time. Although a

longitudinal design would be a proper test of the temporality function of adoption, the difference

between the three groups in our research resemble different stages in an adoption process: from

early to late adoption.

CONCLUSION

The nature of ZW shopping – the product and service offerings of companies aiming to facilitate

such shopping – is likely to change over time. Similarly, consumer preferences also gradually

evolve, and environmental concerns will perhaps shape consumer behavior and practices to a

larger extent in the future. Our study sheds light on three dimensions of green consumer practices

that are insufficiently captured in existing models of adoption. First, we have expanded on the

object of innovation, by conceptualizing green consumer practices along several distinct, but

related dimensions of consumer practice in the store, in the household and beyond. Second, we

go beyond an understanding of the subject of innovation, i.e. “the adopter”, as an individual.

Instead we suggest that adoption should be understood as taking place in a web of related actors.

Finally, we have investigated the temporal process of such adoption, albeit not longitudinally.

Instead, we have compared three groups of consumers: early adopters of ZW, transitional

adopters and non-adopters, in order to shed light on the varying influence of their beliefs about

ZW on their attitudes and behavioral intentions.

In doing so, we have furthered the understanding of an emergent phenomenon in

consumption, namely refill-based and ZW shopping practices. This rapidly developing

phenomenon has been scarcely studied, and our empirical investigation sheds light on its nature

28

and characteristics, and is suggestive of conditions and developments that will determine its

future growth and diffusion.

Our study contributes to the theoretical understanding of the adoption of sustainable

consumer behavior, and we extend the theory of planned behavior (TPB) by applying a

consumer practice perspective. Specifically, we demonstrate that consumer practices can be

integrated into TPB as antecedents of attitudinal beliefs about a green innovation (e.g. ZW),

perceived behavioral control, and intention to use the green innovation in the future. These

findings can inform future studies of such consumer behavior and practices, and such extensions

of this work are necessary to further the understanding of how sustainable consumer practices

can be facilitated. Moreover, the finding that practices, attitudes and behaviors of other actors

(here: members of the household) significantly influences the adoption process, suggests that

institutional theory and theories of adoption at the firm level, such as the technology-

organization-environment (TOE)-framework may provide valuable insight for understanding

adoption at the consumer level as well.

Future studies would benefit from longitudinal research designs through which the

temporal dimension can be investigated more properly. Moreover, experimental approaches that

investigate behavioral interventions or nudges that can reduce the identified barriers for green

consumer practices will be valuable.

Our study also has practical implications for companies that aim to introduce ZW or

refill-based solutions for consumer goods and other products, as well as for governments and

regulatory bodies that aim to facilitate such consumer practices. First, knowledge about the

various sets of practices associated with ZW shopping and the drivers and barriers associated

with them can inform the design of such solutions. Moreover, the varying degrees to which early

29

adopters, transitional adopters and non-adopters perceive changes in such practices to be

burdensome suggests that there is a need for differentiated product and service design to

stimulate such consumption. The design and communication of such solutions could for instance

be tailored to address the specific concerns of various segments. Moreover, our findings suggest

that complete solutions across product categories are likely to increase the inclination of

consumers to adopt ZW practices. Finally, from a government and regulatory perspective,

various incentives can be considered in order to lessen the burden on consumers who adopt ZW

shopping. Examples are tax breaks of the sort that are offered in some countries for services

related to the repair of products.

REFERENCES

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179-211. Aydinliyim, T., & Pangburn, M. S. (2012). Reducing packaging waste and cost via consumer price discounts. Decision Sciences, 43(6), 1063-1089. Babin, B. J., Borges, A., & James, K. (2016). The role of retail prcie image in a multi-country context: France and the USA. Journal of Business Research, 69(3), 1074-1081. Bagozzi, R. P., Davis, F. D., & Warshaw, P. R. (1992). Development and test of a theory of technological learning and usage. Human Relations, 45(7), 659-686. Bagozzi, R. P., & Yi, Y. (1991). Multi-trait-Multimethod Matrices in Consumer Research. Journal of Consumer Research, 17(4), 426-439. Bagozzi, R. P., & Yi, Y. (2012). Specification, evaluation, and interpretation of structural equation models. Journal of the Academy of Marketing Science, 40(1), 8-34. Bardhi, F., & Eckhardt, G. M. (2012). Access-based consumption: The case of car sharing. Journal of Consumer Research, 39(4), 881-898. Belk, R. (2014). You are what you can access: Sharing and collaborative consumption online. Journal of Business Research, 67(8), 1595-1600. Blut, M., Wang, C., & Schoefer, K. (2016). Factors influencing the acceptance of self-service technologies: A meta-analysis. Journal of Service Research, 19(4), 396-416. Bolton, L. E., & Alba, J. W. (2012). When less is more: Consumer aversion to unused utility. Journal of Consumer Psychology, 22(3), 369-383.

30

Cadario, R., & Chandon, P. (2018). Which Healthy Eating Nudges Work Best? A Meta-Analysis of Field Experiments. Marketing Science (forthcoming). Carasuk, R., Becken, S., & Hughey, K. F. (2016). Exploring values, drivers, and barriers as antecedents of implementing responsible tourism. Journal of Hospitality & Tourism Research, 40(1), 19-36. Cenamor, J., Usero, B., & Fernández, Z. (2013). The role of complementary products on platform adoption: Evidence from the video console market. Technovation, 33(12), 405 416. Chaouali, W., Souiden, N., & Ladhari, R. (2017). Explaining adoption of mobile banking with the theory of trying, general self-confidence, and cynicism. Journal of Retailing and Consumer Services, 35, 57-67. Chin, W. W., Mills, A. M., Steel, D. J., & Schwarz, A. (2014, May). Multi-group invariance testing: An illustrative comparison of PLS permutation and covariance-based SEM invariance analysis. In International Conference on Partial Least Squares and Related Methods. Cham: Springer, pp. 267-284. Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User acceptance of computer technology: a comparison of two theoretical models. Management Science, 35(8), 982-1003. Echegaray, F., & Hansstein, F. V. (2017). Assessing the intention-behavior gap in electronic waste recycling: the case of Brazil. Journal of Cleaner Production, 142, 180-190. Frambach, R. T., & Schillewaert, N. (2002). Organizational innovation adoption: A multi-level framework of determinants and opportunities for future research. Journal of Business Research, 55(2), 163-176. Greenpeace (2018). A Crisis of Convenience: The corporations behind the plastics pollution pandemic. Amsterdam, The Netherlands: Greenpeace. Available at: https://www.greenpeace.org/international/publication/19007/a-crisis-of-convenience-the corporations-behind-the-plastics-pollution-pandemic/ Grønhøj, A. (2006). Communication about consumption: a family process perspective on 'green' consumer practices. Journal of Consumer Behaviour, 5(6), 491-503. doi:10.1002/cb.198 Gustavo Jr, J. U., Pereira, G. M., Bond, A. J., Viegas, C. V., & Borchardt, M. (2018). Drivers, opportunities and barriers for a retailer in the pursuit of more sustainable packaging redesign. Journal of Cleaner Production, 187, 18-28. Hair, J. F., Black, W. C., Babin, B. J., & Anderson, R. E. (2010). Multivariate Data Analysis: A global perspective (Seventh ed.). Upper Saddle River, New Jersey: Pearson Education. Hamari, J., Sjöklint, M., & Ukkonen, A. (2016). The sharing economy: Why people participate in collaborative consumption. Journal of the Association for Information Science and Technology, 67(9), 2047-2059. Hu, L.-t., & Bentler, P. M. (1998). Fit Indices in Covariance Structural Equation Modeling: Sensitivity to Underparameterized Model Misspecification. Psychological Methods, 3, 424-453.

31

Hu, L.-t., & Bentler, P. M. (1999). Cutoff Criteria for Fit Indexes in Covariance Structure Analysis: Conventional Criteria versus New Alternatives. Structural Equation Modeling, 6, 1-55. Jackson, T. (2005). Live better by consuming less?: is there a “double dividend” in sustainable consumption? Journal of Industrial Ecology, 9(1‐2), 19-36. Jaeger-Erben, M., Rückert-John, J., & Schäfer, M. (2015). Sustainable consumption through social innovation: a typology of innovations for sustainable consumption practices. Journal of Cleaner Production, 108, 784-798. doi:10.1016/j.jclepro.2015.07.042 Kitzinger, J. (1995). Qualitative research: introducing focus groups. BMJ, 311(7000), 299-302. Kline, R. B. (2010). Principles and Practice of Structural Equation Modeling (3rd ed.). New York: The Guilford Press. Lim, W. M. (2017). Inside the sustainable consumption theoretical toolbox: Critical concepts for sustainability, consumption, and marketing. Journal of Business Research, 78, 69-80. Martin, D., Gustafsson, A., & Choi, S. (2016). Service innovation, renewal, and adoption/rejection in dynamic global contexts. Journal of Business Research, 69(7), 2397-2400. Mirvis, P. H., Sales, A. L., & Hackett, E. J. (1991). The implementation and adoption of new technology in organizations: the impact on work, people, and culture. Human Resource Management, 30(1), 113-139. Morgan, D. L. (1996). Focus groups. Annual Review of Sociology, 22(1), 129-152. Morone, P., & Navia, R. (2016). New consumption and production models for a circular economy. Waste Management & Research, 34(6), 489–490. Ozaki, R. (2011). Adopting sustainable innovation: what makes consumers sign up to green electricity? Business Strategy and the Environment, 20(1), 1-17. Ozaki, R., & Sevastyanova, K. (2011). Going hybrid: An analysis of consumer purchase motivations. Energy Policy, 39(5), 2217-2227. Perera, C., Auger, P., & Klein, J. (2018). Green Consumption Practices Among Young Environmentalists: A Practice Theory Perspective. Journal of Business Ethics, 152(3), 843-864. Phipps, M., & Ozanne, J. L. (2017). Routines disrupted: Reestablishing security through practice alignment. Journal of Consumer Research, 44(2), 361-380. Pietzsch, N., Ribeiro, J. L. D., & de Medeiros, J. F. (2017). Benefits, challenges and critical factors of success for Zero Waste: A systematic literature review. Waste Management, 67, 324-353. Roberto, C. A., & Kawachi, I. (2014). Use of psychology and behavioral economics to promote healthy eating. American Journal of Preventive Medicine, 47(6), 832-837. Rosseel, Y. (2012). lavaan: An R Package for Structural Equation Modeling. Journal of Statistical Software, 48(2), 1-36. Satorra, A., & Bentler, P. M. (2001). A Scaled Difference Chi-square Test Statistic for Moment Structure Analysis. Psychometrika, 66(4), 507-514.

32

Schepers, J., & Wetzels, M. (2007). A meta-analysis of the technology acceptance model: Investigating subjective norm and moderation effects. Information & Management, 44(1), 90-103. Schoenau, M., & Müller, M. (2017). What affects our urban travel behavior? A GPS-based evaluation of internal and external determinants of sustainable mobility in Stuttgart. Transportation Research Part F: Traffic Psychology and Behaviour, 48, 61-73. Scott, W. R. (2008). Institutions and organizations: Ideas and interests. London: Sage. Shama, A. (1985). The voluntary simplicity consumer. Journal of Consumer Marketing, 2(4), 57-63. Sharma, P. (2010). Measuring personal cultural orientations: scale development and validation. Journal of the Academy of Marketing Science, 38(6), 787-806. Shove, E. (2003). Converging Conventions of Comfort, Cleanliness and Convenience. Journal of Consumer Policy, 26(4), 395-418. Shukla, P., & Purani, K. (2012). Comparing the importance of luxury value perceptions in cross national contexts. Journal of Business Research, 65(10), 1417-1424. Steenkamp, J.-B. E. M., & Baumgartner, H. (1998). Assessing Measurement Invariance in Cross-National Consumer Research. Journal of Consumer Research, 25(1), 78-107. Steiner, M., Wiegand, N., Eggert, A., & Backhaus, K. (2016). Platform adoption in system markets: The roles of preference heterogeneity and consumer expectations. International Journal of Research in Marketing, 33(2), 276-296. Stewart, D. W., & Shamdasani, P. N. (2014). Focus groups: Theory and practice (Vol. 20). London: Sage Publications. Stryja, C., & Satzger, G. (2018). Try Before You Buy: How to Design Information Systems to Enhance Consumer Willingness to Test Sustainable Innovations. Journal of Technology Management & Innovation, 13(1), 19-26. Taylor, S., & Todd, P. (1997). Understanding the determinants of consumer composting behavior 1. Journal of Applied Social Psychology, 27(7), 602-628. Thorbjørnsen, H., Pedersen, P. E., & Nysveen, H. (2007). “This is who I am”: Identity expressiveness and the theory of planned behavior. Psychology & Marketing, 24(9), 763 785. Thorbjørnsen, H., Pedersen, P. E., & Nysveen, H. (2009). Categorizing networked services: The role of intrinsic-, user network-and complement network attributes. European Journal of Marketing, 43(3/4), 371-397. Toft, M. B., & Thøgersen, J. (2015). Exploring private consumers’ willingness to adopt Smart Grid technology. International Journal of Consumer Studies, 39(6), 648-660. Tornatzky, L.G., & Fleischer, M. (1990). The deployment of technology. In L.G. Tornatzky, J.D. Eveland & M. Fleischer (Eds.), The Process of Technological Innovation. Lexington, MA: Lexington Books.

33

Venkatesh, V., Thong, J. Y., & Xu, X. (2012). Consumer acceptance and use of information technology: extending the unified theory of acceptance and use of technology. MIS Quarterly, 157-178. Vermeir, I., & Verbeke, W. (2006). Sustainable food consumption: Exploring the consumer “attitude–behavioral intention” gap. Journal of Agricultural and Environmental Ethics, 19(2), 169-194. Verplanken, B., & Wood, W. (2006). Interventions to break and create consumer habits. Journal of Public Policy & Marketing, 25(1), 90-103. Wang, Z., Zhang, B., & Li, G. (2014). Determinants of energy-saving behavioral intention among residents in Beijing: Extending the theory of planned behavior. Journal of Renewable and Sustainable Energy, 6(5), 053127. Warde, A. (2005). Consumption and Theories of Practice. Journal of Consumer Culture, 5(2), 131-153. Wever, R., & Vogtländer, J. (2013). Eco‐efficient value creation: an alternative perspective on packaging and sustainability. Packaging Technology and Science, 26(4), 229-248. Wikström, F., Williams, H., Verghese, K., & Clune, S. (2014). The influence of packaging attributes on consumer behaviour in food-packaging life cycle assessment studies-a neglected topic. Journal of Cleaner Production, 73, 100-108. Zaman, A. U. (2015). A comprehensive review of the development of zero waste management: lessons learned and guidelines. Journal of Cleaner Production, 91, 12-25. Xie, C., Bagozzi, R. P., & Grønhaug, K. (2015). The role of moral emotions and individual differences in consumer responses to corporate green and non-green actions. Journal of the Academy of Marketing Science, 43(3), 333-356.

34

Table 1: Measures used in the study Variable Perceived changes in shopping practices (ShP) (1 = to a very small extent, 7 = to a very large extent)

ShP1 How I have to plan the shopping ShP2 What I buy in the store ShP3 How much I buy ShP4 How many different stores I shop in

Perceived changes in household practices (Food) (HPF) (1 = to a very small extent, 7 = to a very large extent)

HPF1 The way I store food products at home HPF2 The way I cook HPF3 What types of ingredients I use HPF4 How much food I make “from scratch”

Perceived changes in household practices (Waste) (HPW) (1 = to a very small extent, 7 = to a very large extent)

HPW1 The way I store waste HPW2 The way I get rid of waste

Perceived changes in social practices (SoP) (1 = to a very small extent, 7 = to a very large extent)

SoP1 Who in my household will do the shopping SoP2 What tasks the different members of my household have with regard to shopping, cooking, etc. SoP3 Who in the household will do the cooking

Perceived changes in general environmental practices (OEP) (1 = to a very small extent, 7 = to a very large extent)

GEP1 How I show my environmental commitment to others GEP2 How likely I am to recycle products and other waste GEP3 How environmentally friendly my method of transportation is GEP4 How eco-friendly (non-food) products I buy

Attitudinal beliefs (1 = completely disagree, 7 = completely agree)

AB1 Packaging-free shopping is beneficial for the environmentAB2 Packaging-free shopping saves me moneyAB3 I generally see many benefits with packaging-free shopping

Behavioral Control (BC) (1 = completely disagree, 7 = completely agree)

I would not have had any trouble understanding how packaging-free shopping works Social Norm (1 = completely disagree, 7 = completely agree)

I care about what others think about packaging-free shopping Behavioral Intentions (BI) (1 = completely disagree, 7 = completely agree)

BI1 I am going to shop packaging-free going forward BI2 I am more likely to shop packaging-free than most others

35

Table 2: Factor loadings, reliability, validity, and fit measures

Construct Early adopter Transitional adopter Non-adopter FLa CRb AVEc FLa CRb AVEc FLa CRb AVEc

Shopping practices (ShP) .75 .43 .74 .43 .80 .50 ShP1 .599 .555 .558 ShP2 .729 .795 .816 ShP3 .663 .603 .713 ShP4 .632 .635 .728 Household practices (food) (HPF) .82 .54 .86 .60 .88 .65 HPF1 .570 .662 .753 HPF2 .771 .811 .792 HPF3 .811 .849 .838 HPF4 .758 .771 .847 Household practices (waste) (HPW) .91 .84 .89 .81 .93 .87 HPW1 .904 .937 .894 HPW2 .930 .861 .970 Social practices (SoP) .88 .72 .87 .70 .87 .69 SoP1 .829 .784 .725 SoP2 .876 .860 .894 SoP3 .839 .863 .869 General environmental practices (GEP) .83 .55 .84 .56 .83 .55 GEP1 .590 .612 .686 GEP2 .821 .813 .801 GEP3 .727 .751 .665 GEP4 .798 .804 .793 Attitudinal beliefs (AB) .69 .43 .71 .47 .73 .49 AB1 .622 .654 .692 AB2 .532 .472 .488 AB3 .796 .868 .870 Behavioral intentions (BI) .62 .46 .71 .55 .84 .72 BI1 .755 .753 .872 BI2 .586 .727 .821 Fit Measures

289.029 (188), p < .001 337.239 (188), p < .001 292.607 (188), p < .001

/ 1.537 1.794 1.556 RMSEA .051; 90% CI : [.039, .062] .054; 90% CI : [.045, .062] .057; 90% CI : [.044, .069] CFI .938 .934 .935 TLI .923 .919 .920 SRMR .051 .054 .066 N 205 276 171 a FL: Standardized factor loading; b CR: Composite reliability; c AVE: Average variance extracted.

36

Table 3: Structural relationship comparisons between groups

Relationship Comparison ∆ p-value Note

ShP BI Early adopters vs. Non-adopters

8.68 (1) .003 Effect is larger and positive for early adopters

ShP Attitudinal beliefs

11.65 (1) .001 Effect is larger and negative for non-adopters

Social norm BI Early adopters vs. Transitional adopters

5.73 (1) .017 Effect is larger and positive for transitional adopters

SoP BC1 3.56 (1) .059 Effect is larger and negative for transitional adopters

ShP Attitudinal beliefs

Non-adopters vs. Transitional adopters

3.52 (1) .061 Effect is larger and negative for non-adopters

1 BC: Behavioral control.

37

Figure 1: Four distribution solutions for reduced plastic packaging

38

Figure 2: Conceptual model

Behavioral Intentions

Attitudinal Beliefs

Behavioral Control

Shopping practices

Household practices (Food)

Household practices (Waste)

General environmental

practices

Social practices

Social Norm

39

Figure 3: Estimation Results for “Totally Free” Structural Model

Behavioral Intentions

Attitudinal Beliefs

Behavioral Control

Shopping practices

Household practices (Food)

Household practices (Waste)

General environmental

practices

Social practices

Social Norm

.74***.51***

-.25**

.17***.23**

.34***

-.14**

.35**

Early Adopter

Behavioral Intentions

Attitudinal Beliefs

Behavioral Control

Shopping practices

Household practices (Food)

Household practices (Waste)

General environmental

practices

Social practices

Social Norm

.95***

.09*

-.27***

.38**

Transitional Adopter

.07**

.29***

-.40***

40

Behavioral Intentions

Attitudinal Beliefs

Behavioral Control

Shopping practices

Household practices (Food)

Household practices (Waste)

General environmental

practices

Social practices

Social Norm

.75***

.10*

.63***

Non-Adopter

.46***

-.24*

.38***

-.31**

41

Appendix 1: Correlation matrix between latent constructs

Early Adopter ShP HPF SoP GEP AB BI HPW

ShP 1

HPF 0.686 (0.074) 1

SoP 0.543 (0.076) 0.663 (0.063) 1

GEP 0.299 (0.090) 0.520 (0.083) 0.432 (0.085) 1

AB 0.087 (0.104) 0.108 (0.096) -0.044 (0.095) 0.355 (0.086) 1

BI 0.344 (0.111) 0.189 (0.093) 0.095 (0.101) 0.433 (0.092) 0.818 (0.083) 1

HPW 0.476 (0.077) 0.558 (0.073) 0.428 (0.087) 0.588 (0.067) 0.140 (0.087) 0.209 (0.086) 1

Note: Standard errors are in brackets

Transitional Adopter ShP HPF SoP GEP AB BI HPW

ShP 1

HPF 0.790 (0.045) 1

SoP 0.415 (0.082) 0.556 (0.072) 1

GEP 0.459 (0.067) 0.532 (0.062) 0.409 (0.077) 1

AB 0.144 (0.078) 0.043 (0.071) -0.179 (0.099) 0.295 (0.075) 1

BI 0.312 (0.089) 0.252 (0.079) 0.033 (0.100) 0.476 (0.072) 0.797 (0.054) 1

HPW 0.345 (0.075) 0.554 (0.059) 0.333 (0.075) 0.593 (0.063) 0.129 (0.068) 0.278 (0.072) 1

Note: Standard errors are in brackets

Non-Adopter ShP HPF SoP GEP AB BI HPW

ShP 1

HPF 0.618 (0.062) 1

SoP 0.425 (0.096) 0.609 (0.079) 1

GEP 0.382 (0.092) 0.479 (0.091) 0.514 (0.106) 1

AB -0.206 (0.110) -0.008 (0.102) 0.002 (0.111) 0.311 (0.085) 1

BI -0.210 (0.101) -0.055 (0.099) 0.064 (0.127) 0.411 (0.090) 0.737 (0.063) 1

HPW 0.179 (0.090) 0.467 (0.088) 0.417 (0.094) 0.428 (0.086) 0.193 (0.090) 0.093 (0.092) 1

Note: Standard errors are in brackets

42

Appendix 2: Fit measures for invariant models Model AIC BIC ∆ p-valueConfigural model (M1) 49625 50794 Metric model (M2) 49601 50636 M2 vs. M1 33.245 (30) .312 Scalar model (M3) 49687 50587 M3 vs. M2 145.647 (30) < .001

43

Appendix 3: Parameter Estimates on Structural Paths

Early Adopters Dependent Variable Independent Variable β SE t-value

Behavioral intentions

Attitudinal beliefs .736*** .154 4.779 Behavioral control .174*** .046 3.811 Shopping practices .511*** .145 3.527 Household practices (food) -.246** .116 -2.121 Household practices (waste) -.046 .060 -.760 Social practices -.020 .072 -.276 General environmental practices .225** .113 1.992 Social norms -.028 .031 -.911 R2 .850

Attitudinal beliefs

Shopping practices .099 .140 .710 Household practices (food) .000 .131 .001 Household practices (waste) -.043 .053 -.808 Social practices -.143** .069 -2.056General environmental practices .341*** .101 3.371 R2 .206

Behavioral control

Shopping practices -.039 .249 -.157 Household practices (food) .168 .255 .660 Household practices (waste) -.080 .102 -.781 Social practices -.144 .146 -.982 General environmental practices .354** .165 2.144 R2 .059

Transitional Adopters Dependent Variable Independent Variable β SE t-value

Behavioral intentions

Attitudinal beliefs .947*** .152 6.224 Behavioral control .089* .052 1.702 Shopping practices .058 .191 .305 Household practices (food) .051 .163 .311 Household practices (waste) .005 .054 .100 Social practices .014 .084 .166 General environmental practices .136 .101 1.345 Social norms .066** .032 2.055 R2 .690

Attitudinal beliefs

Shopping practices .310 .346 .896 Household practices (food) -.195 .284 -.685 Household practices (waste) .018 .069 .264 Social practices -.268*** .086 -3.109 General environmental practices .294*** .095 3.078R2 .332

Behavioral control

Shopping practices .586 .721 .813 Household practices (food) -.425 .591 -.719 Household practices (waste) .063 .159 .395 Social practices -.397*** .143 -2.773 General environmental practices .380** .187 2.034 R2 .164

44

Non-Adopters Dependent Variable Independent Variable β SE t-value

Behavioral intentions

Attitudinal beliefs .748*** .128 5.825 Behavioral control .103* .056 1.845 Shopping practices -.237* .127 -1.867 Household practices (food) -.046 .130 -.353 Household practices (waste) -.085 .063 -1.362 Social practices .068 .106 .643 General environmental practices .383*** .133 2.885 Social norms .054 .059 .906 R2 .619

Attitudinal beliefs

Shopping practices -.310** .139 -2.236 Household practices (food) .022 .144 .152 Household practices (waste) .062 .060 1.031 Social practices -.140 .128 -1.098 General environmental practices .463*** .117 3.977 R2 .251

Behavioral control

Shopping practices -.044 .209 -.213 Household practices (food) -.364 .253 -1.438 Household practices (waste) -.059 .090 -.659 Social practices -.171 .238 -.717 General environmental practices .627*** .209 2.997 R2 .132

Note: Table reports unstandardized coefficients. * p < .1; ** p < .05; *** p < .01.