Skip to main content
Intended for healthcare professionals
Free access
Research article
First published online November 10, 2017

Cognitive Accessibility as a New Factor in Proenvironmental Spillover: Results From a Field Study of Household Food Waste Management

Nicole Sintov, Sally Geislar [email protected], and Lee V. WhiteView all authors and affiliations
Volume 51, Issue 1
https://doi.org/10.1177/0013916517735638

Abstract

An emerging body of literature has contributed to understanding behavioral spillover; however, a limited range of behaviors and psychological pathways have been studied. The current study investigates whether starting to compost, a relatively difficult behavior receiving limited attention in the spillover literature, results in spillover to household waste prevention behaviors, including food, energy, and water waste prevention. It also tests cognitive accessibility as a new mediator in the spillover process, and advances an integrative process model to address methodological inconsistencies in the spillover literature. Data are from a 2015 longitudinal field experiment to increase composting. Participants (N = 284) were residents of Costa Mesa, California, who received a structural intervention (i.e., curbside organic waste bins) and procedural information about composting. Positive spillover was observed. Additionally, cognitive accessibility partially mediated the relationship between composting and energy and water waste–prevention behaviors. Future research should adopt a consistent definition of spillover and explore additional pathways.

Introduction

In recent years, significant advances have been made toward understanding the emerging topic of behavioral spillover. Spillover occurs when engagement in an initial target behavior, often the target of an intervention, is linked to performance of another, seemingly unrelated behavior (Austin, Cox, Barnett, & Thomas, 2011; United Kingdom Department for Environment, Food, and Agricultural Affairs, 2008). For instance, outdoor irrigation restrictions targeting water conservation may also result in increased recycling behavior. Spillover may be positive, when an increase in one proenvironmental behavior (PEB) is associated with an increase in another PEB, as in the example provided, or negative (Thøgersen & Crompton, 2009), such as when a homeowner reduces recycling after adopting water conservation measures in response to irrigation restrictions. Indeed, although several studies have found positive associations between PEBs (Thøgersen, 1999; Thøgersen & Ölander, 2003; Whitmarsh & O’Neill, 2010), potentially suggesting positive spillover, evidence from other work showing inverse relationships among PEBs may indicate negative spillover (Barr, Shaw, Coles, & Prillwitz, 2010; Weber, 1997).
Given this is a nascent area, a limited range of behaviors and psychological pathways through which the spillover process occurs have been studied. From a theoretical standpoint, further understanding of these topics is needed to identify the conditions under which any spillover, positive or negative, occurs. From a practical standpoint, a better understanding of spillover can aid practitioners deploying real-world PEB interventions, illuminating opportunities to maximize program benefits (i.e., positive spillover) and avoid unintended consequences (i.e., negative spillover). The current study examined PEB spillover as part of a larger intervention project that encouraged composting through a curbside collection program (Geislar, 2017).

The Problem of Food Waste

Food waste constitutes the largest material stream sent to landfills by municipalities (22% by weight) and has the lowest recovery rate (i.e., 3.9% vs. 57.3% of recovered yard trimmings) of any recycled or composted material in the United States (U.S. Environmental Protection Agency [EPA], 2013a). The 50 million metric tons of organic waste entering landfills annually, 70% of which is food waste, is not simply a problem of limited landfill space; its decomposition generates 11% of U.S. greenhouse gas (GHG) emissions (U.S. EPA, 2013a, 2013b). Thus, reducing organic matter in landfills can considerably reduce GHG emissions. To address food waste challenges, municipalities have taken various measures, including community drop-off systems in several Swiss and Canadian cities. U.S. cities tend to integrate food waste into existing curbside waste collection programs (Platt & Goldstein, 2014).

The challenge of composting

Curbside organics collection programs require households to dispose of organic materials separately from other household wastes (i.e., compost), making their success heavily dependent on household behavior change. For instance, a recent study found that over half of San Francisco’s landfill-cart contents could be recycled or composted despite separate carts for each (San Francisco Department of the Environment, 2013).
Further, the nature of food waste poses unique challenges as it begins decomposing immediately, generating unpleasant odors and attracting pests. Lapses in food waste removal can exacerbate these effects and heighten responses of “disgust,” characterized as “defense against the harm from potential foods, or things that can easily contaminate foods” (Rozin, Haidt, & McCauley, 2008, p. 758). Food decay is considered a “prototypical odor of disgust” that signals potential contamination (Rozin et al., 2008, p. 761). Hence, its associated sights and smells can trigger attempts to avoid food-borne illness. As composting involves interaction with food waste in the kitchen, composting can inherently increase the likelihood of avoidance rather than continued participation, which we argue makes it a difficult behavior.
Studies on household food waste management have largely focused on decision making around food waste generation (Parizeau, von Massow, & Martin, 2015) and reduction (Visschers, Wickli, & Siegrist, 2016). Related to composting specifically, one study examined intentions to participate in backyard composting and found that participants preferred curbside pickup (Ghani, Rusli, Biak, & Idris, 2013); another found that separate waste bins in kitchens promoted greater food waste separation than information leaflets (Bernstad, 2014). We identified no studies on the influence of composting on other types of household waste management (spillover).

Evidence for Behavioral Spillover

Spillover can occur within and/or across different behavioral domains (e.g., one water conservation behavior to another vs. water conservation to energy conservation behavior; De Young, 2000; Maiteny, 2002). Although most spillover research has been cross-sectional (Bratt, 1999; Cornelissen, Pandelaere, Warlop, & Dewitte, 2008; Evans et al., 2013; Stern, Dietz, Abel, Guagnano, & Kalof, 1999), limiting conclusions about causality, a few recent studies have used designs that shed light on directionality of effects. For instance, Lanzini and Thøgersen (2014) applied an intervention to promote green purchasing among university students using either monetary incentives or praise. Positive spillover was observed whereby students in the monetary condition increased green purchasing behavior (i.e., target PEB), which predicted increases in several spillover PEBs, including turning off lights and recycling. In an experiment with another university student sample, Truelove, Yeung, Carrico, Gillis, and Raimi (2016) found that Democrats who recycled a plastic bottle expressed less support for campus environmental policy than controls who did not dispose of a water bottle, indicating negative spillover.
Beyond these university-based efforts, several community field experiments were identified. Hotel guests who made a specific commitment to reuse towels (or received a lapel pin to increase salience of environmental identity) demonstrated positive spillover; not only were they more likely to reuse towels, but they also turned off lights more than those who read the same message but did not commit to it (Baca-Motes, Brown, Gneezy, Keenan, & Nelson, 2013). In another study (Steinhorst, Klöckner, & Matthies, 2015), German households received energy-saving tips either in terms of environmental impacts (CO2) or monetary savings. Although each reported intentions to conserve electricity (the target PEB), those in the environmental framing group also reported stronger intentions to engage in nontarget PEBs, such as reducing driving and beef consumption. A third study found that Australian residents reporting curtailed water use were more likely to subsequently install water efficient appliances (Lauren, Fielding, Smith, & Louis, 2016). Finally, following a Welsh policy charging a plastic shopping bag fee, Thomas, Poortinga, and Sautkina (2016) observed spillover from increased frequency of bringing reusable bags to sustainable transportation choices and home water and energy conservation.

Spillover process model

Although spillover has been broadly conceptualized as involving an initial target behavior and a subsequent spillover behavior, one limitation in the literature is the somewhat inconsistent manner in which “spillover” has been measured across individual studies. Specifically, some studies did not examine psychological mechanisms of spillover (Baca-Motes et al., 2013; Lanzini & Thøgersen, 2014; Thomas et al., 2016). Others reported the co-occurrence of target behaviors and spillover behaviors, but did not take into account temporal effects to discern which may have occurred first (Baca-Motes et al., 2013), and few have examined change over time (Lanzini & Thøgersen, 2014; Lauren et al., 2016; Thomas et al., 2016). Still others have fallen a bit short of assessing spillover to behaviors, and have instead measured hypothetical behavior, including behavioral intentions (Steinhorst et al., 2015), as well as support (Lacasse, 2016; Truelove et al., 2016) and willingness to pay (Margetts & Kashima, 2016) for environmental policy. Summarizing, most studies have examined pieces of the spillover process; we found only one that investigated it in its entirety using a behavioral measure of spillover in a repeated-measures study (Lauren et al., 2016).
Integrating pieces of spillover that prior studies have measured, we propose a broad, integrative spillover process model that outlines several approaches to capturing spillover. In conceptualizing the model, we consider that spillover is, by definition, a causal process whereby the performance of one behavior causes a secondary behavior; therefore, spillover studies must make a case for causality. The integrative spillover process model that we define begins with a target behavior that influences the performance of a secondary (spillover) behavior (see Figure 1). Of critical importance to determine causality, the temporal order of target and spillover behaviors must be established; this can be done by measuring change in target behavior, change in spillover behavior, or both. Hence, spillover can be modeled as (a) change in a target behavior that affects change in spillover behavior (Model A); (b) change in a target behavior that affects spillover behavior (Model B, used in current study); or (c) a target behavior that affects change in spillover behavior (Model C). In all models, the link between the target and spillover behaviors is mediated by the spillover pathway—the psychosocial mechanisms that account for the relationship between seemingly independent behaviors. Whereas correlational studies can provide illuminating evidence consistent with spillover, because they omit temporal ordering of behaviors, causal processes are veiled and hence these studies are excluded from our model.
Figure 1. Integrative spillover process model with study design variants.
We argue that intervention studies offer a methodologically advantageous approach for studying spillover. Broadly, an intervention is a systematic attempt to encourage change in a target behavior (Truelove, Carrico, Weber, Raimi, & Vandenbergh, 2014), such as provision of energy use feedback, financial incentives, or structural elements (e.g., curbside waste carts). Intervention studies help build cases for causality by establishing the temporal order of behaviors and eliminating confounds (e.g., history effects), making them strong methodological vehicles.

Prior work on spillover pathways

Various psychological processes may be at play in the spillover pathway. Although a full review is beyond the scope of this article, we briefly summarize key pathways below.
For negative spillover, previous studies have found moral licensing (Mazar & Zhong, 2010; Merritt, Effron, & Monin, 2010; Sachdeva, Iliev, & Medin, 2009; Thøgersen & Crompton, 2009), guilt (Truelove et al., 2016), and single-action bias (Weber, 1997) to mediate the relationship between target and spillover PEBs. Relevant to positive spillover, self-efficacy and personal norms (Lauren et al., 2016; Steinhorst et al., 2015), as well as goal orientation (Susewind & Hoelzl, 2014) and striving (Margetts & Kashima, 2016) can act as pathways.
The positive spillover pathway that has received the most empirical attention relates to identity and people’s desire to think and act in consistent ways. Informed by multiple theories in psychology (Bem, 1972; Cialdini, Trost, & Newsom, 1995; Festinger, 1957), this pathway rests on the notion that when one performs an identity-consistent behavior, subsequently engaging in inconsistent behavior can generate cognitive dissonance; attempts to resolve this unpleasant state can include behavioral change (Festinger, 1957; Freedman & Fraser, 1966). Similarly, Self-Perception Theory suggests that when presented with a decision to engage in a PEB (or not), people “look back” to their past behavior as a signal of their self-concept (Bem, 1972). Hence, if a person looks back and recalls past instances when she/he has performed PEBs, she/he may be more likely to perform future PEBs. Lending support to this theory, reminding individuals of their previous PEBs or labeling people as environmentalists can result in their making future eco-friendly decisions (Lacasse, 2016; van der Werff, Steg, & Keizer, 2014). Reminders strengthen environmental self-identity, or the extent to which one sees oneself as an environmentally friendly person (Whitmarsh & O’Neill, 2010). Similarly, prosocial identity has been found to mediate spillover between prosocial behaviors (Gneezy, Imas, Brown, Nelson, & Norton, 2012).
Spillover processes may be influenced by the level of difficulty of the initial PEB (Truelove et al., 2014). For instance, when people consider a difficult past PEB, the salience of their environmental self-identity increases (van der Werff et al., 2014), and they later perform PEBs in line with this self-identity. The opposite is true for easy initial behaviors, suggesting that difficult PEBs serve as stronger signals of identity and may enhance positive spillover.

Cognitive accessibility: A new spillover pathway

Self-Perception Theory’s “looking back” process implicated in the identity pathway is complex and multistep. Because cognitive processes are involved, cognitive biases may influence which behaviors are more likely to be recalled. For instance, the availability heuristic suggests that more readily accessible information is more likely to be recalled (Tversky & Kahneman, 1973, 1974) Indeed, many studies provide evidence of this bias, whereby more cognitively available risks are overestimated and less available risks are underestimated (Combs & Slovic, 1979; Lichtenstein, Slovic, Fischhoff, Layman, & Combs, 1978; Pachur, Hertwig, & Stienmann, 2012). As the availability heuristic was initially used to describe judgments of likelihood or frequency, we use “cognitive accessibility”—the frequency with which people interact with or think about something, coined by Schley and Dekay (2015)—to describe similar processes in judgments of environmental impacts.
In particular, research suggests that easily accessible information plays a key role in judgments of environmental problems. For example, beliefs about global warming correlate with outdoor temperature on the study day, and priming with heat-related cognitions can influence such beliefs (Joireman, Truelove, & Duell, 2010). Also, perceived temperature fluctuations influence global warming opinions more than actual fluctuations (Li, Johnson, & Zaval, 2011; Zaval, Keenan, Johnson, & Weber, 2014). Moreover, physically experiencing warmth facilitates mental representations of warmth, including global warming (Risen & Critcher, 2011).
In summary, prior work suggests that more cognitively accessible information is more likely to influence perceptions and opinions. It follows that interventions that augment information availability—or frequency of behaviors that influence information availability—may improve subsequent environmental perceptions and choices (Arkes, 1991). For instance, a composting intervention can offer repeated opportunities to engage in PEBs, which may be considered relatively visceral in nature. Frequent engagement in such behavior could also increase the availability of waste-related cognitions. Therefore, someone who participates in such an intervention may not only have a fairly large catalogue of examples of when she/he has performed PEBs, but also more waste-related cognitions. In “looking back” on past behavior, she/he may use these more available cognitions to infer self-concept (Bem, 1972), which can influence environmental identity. Particularly if such PEBs are difficult, which we suggest is the case with composting given the “disgust” factor (Rozin et al., 2008), they may strengthen environmental identity and lead to more future PEBs (van der Werff et al., 2014).

Normative messaging

This study also examines the effect of normative messaging in the spillover process. The larger study, of which the current research is a part, used descriptive social norms messaging to encourage composting (i.e., the target behavior) among randomly selected participants (Geislar, 2017), offering the opportunity to test for the additive effects of normative messaging on spillover behavior. Although a wealth of previous work has found that communicating descriptive social norms, or the actual behavior of a social group (Cialdini, Kallgren, & Reno, 1991), can have a positive effect on that behavior (Goldstein, Cialdini, & Griskevicius, 2008), few have studied their effect on spillover behaviors.
We found only one study in which nontarget behaviors were measured following a normative messaging intervention (Kormos, Gifford, & Brown, 2015). Whereas descriptive norms were effective in promoting PEBs targeted by messaging, other PEBs were not influenced. However, this study did not explicitly examine the spillover process. Further, other work suggests moderation effects involving proenvironmental identity and norms, whereby norms may result in negative spillover among those with weaker proenvironmental identities and positive spillover among those with stronger proenvironmental identities (Truelove et al., 2014). Norms messages may serve as reminders of identity. Prior work finds that conveying norms characteristic of one’s in-group can strengthen intentions to perform norm-consistent behavior (Terry & Hogg, 1996). Hence, it is possible that a composter who receives norms messages about neighbors’ composting may engage in other in-group PEBs, or positive spillover, to maintain positive social identity (Hogg & Reid, 2006). It is unclear whether norms may influence spillover via the target behaviors alone or via direct effects on both target and spillover behaviors.

Contributions

This article makes several contributions. First, from a theoretical standpoint, we investigate cognitive accessibility as a possible new factor in the spillover pathway. Second, on a methodological level, spillover has been defined in varied ways, revealing portions of the spillover process, but few studies have examined the entire process. We refine and investigate an integrative spillover process model (see Figure 1, Model B). Furthermore, the majority of spillover research has been conducted on university campuses; we found only four field studies, two of which did not measure spillover pathways (Baca-Motes et al., 2013; Thomas et al., 2016) and another that assessed behavioral intentions rather than behavior (Steinhorst et al., 2015). Our project builds on this body of work using a 6-month longitudinal field experiment. Finally, whereas difficult behaviors are hypothesized to result in greater spillover (Truelove et al., 2014), most spillover studies have targeted relatively easy initial behaviors (e.g., recycling a plastic bottle once, reusing towels during a hotel visit). In contrast, our target behavior (i.e., composting, which has received limited attention in studies of PEB overall) is a more difficult behavior.

Study Objectives and Hypotheses

Our first objective was to test whether adopting a new, difficult PEB—composting—would increase performance of household waste prevention behaviors. We hypothesized that positive spillover would occur, whereby an intervention providing curbside organics carts and procedural information would increase composting behavior (results in Geislar, 2017), which would in turn be associated with other waste prevention behaviors (i.e., food, energy, and water waste prevention; Hypothesis 1 [H1]). Second, we investigated cognitive accessibility as a possible new mediator of the spillover process. Composting involves repeated interaction with waste, so when people “look back,” they may be more likely to recall such frequent behaviors, as well as related cognitions (i.e., thoughts about waste). We predicted that cognitive accessibility of waste would partially mediate the relationship between adopting a new target behavior (i.e., beginning to compost) and subsequent spillover behaviors (Hypothesis 2 [H2]). Finally, we expected descriptive norms to be associated with spillover (Hypothesis 3a [H3a], nondirectional), and to moderate the relationship between beginning to compost and spillover by amplifying spillover effects (Hypothesis 3b [H3b]).

Method

Procedure

The present study is part of a larger project on household composting (Geislar, 2017) in which a repeated-measures experimental design followed participants over 6 months in 2015. The timeline for the larger study is available online in Appendix A.

Recruitment

At baseline, 7,400 addresses were randomly selected from a complete list of 20,000 single-family residences served by the Costa Mesa Sanitary District (the District) to receive a recruitment mailer. Each mailer included an information sheet, baseline survey, incentive information, and postage-paid return envelope (and a URL for online survey option).

Surveys

Participants completed a baseline survey, a midpoint assessment 4 months later, and a follow-up survey 2 months after that (i.e., 6 months following baseline). Survey measures are summarized in the Measures and Variable Coding section below. Reminder postcards were sent to nonresponders 2 weeks after distribution of the follow-up survey. Respondents received a $10 gift card incentive to a chain grocery store for completing each survey and four were randomly selected to receive a $75 grocery gift card for completing all surveys. Only data from the baseline and follow-up surveys were used in the present study.

Intervention procedures and materials

Following recruitment, the District launched an intervention to divert waste from landfills. Specifically, the 20,000 homes received 64-gallon organics carts, collection service, and flyers describing the types of food and yard waste to be placed in the organics cart (see online Appendix B). In this intervention, composting involved several individual behaviors. Residents were asked to keep food (e.g., plate scraps, inedible trimmings, and spoilage) and yard waste intended for organics carts separate from landfill waste intended for trash carts. Organic waste was then to be placed in the organics cart, which was to be placed at the curb on collection day (both organics and trash carts were collected weekly on the same day). Prior to this, residents had a single cart for all waste (recyclables separated after collection). Residents were also offered free 2-gallon pails for indoor food scrap collection. Roughly 75% of the sample reported using either the free pail (38%) or an improvised item (38%; e.g., bowl, empty container) for indoor food waste separation.
Additionally, following the midpoint assessment, half of respondents were randomly assigned to receive weekly descriptive norms messaging for 8 weeks. Messages read: “75% of households in Costa Mesa separated all of their food scraps this week,” varying from 73-84% each week (see online Appendix C). As in previous work (Goldstein et al., 2008), these figures were designed to test the effect of conveying positive norms and did not reflect actual rates. Participants who returned print baseline surveys received print messaging (i.e., a magnet and weekly stickers with updated values) and digital respondents received email messaging with the same information. Intervention results are reported in Geislar (2017).

Participants

Residents were eligible for the larger project if they were adults residing in one of the 7,400 randomly selected homes to receive a recruitment mailer. For the current study, participants also had to return baseline (n = 1,089) and follow-up surveys (n = 399), and have complete data (n = 351) for key variables. Because the current study sought to test the effect of adopting a new target behavior on performance of spillover behaviors, those already performing the target behavior at baseline were excluded (n = 67), yielding a final sample of N = 284. A subset of 109 respondents from this sample was randomly assigned to receive norms messaging.
In the final sample, 64% were female, and roughly half of respondents were over the age of 55. Approximately 62% of respondents had at least a bachelor’s degree, and 80% identified racially as White, with 7% identifying as Asian, and smaller proportions identifying as other races. Relative to Costa Mesa and broader U.S. populations, the sample was older, more highly educated, and had more female and non-Hispanic White respondents (see online Appendix D).

Measures and Variable Coding

The survey tool was developed for the larger project. Relevant to the current study, the baseline survey assessed demographic characteristics, composting behavior, and food waste prevention (spillover) behaviors. The follow-up survey additionally assessed energy and water waste prevention behaviors and cognitive accessibility related to waste prevention. This latter set of items was not included at baseline because the objective of studying spillover was added to the project after baseline. See online Appendix E for a correlation matrix among key variables.

Composting

The target behavior, composting, was operationalized as (a) disposing of food waste in a correct location and (b) separating food waste with adequate frequency. A binary variable was created to represent correct location (0 = no, 1 = yes) based on two baseline survey items, the first being “Do you currently separate food scraps for composting?” Response options included “yes,” “no,” and “don’t know”; only “yes” responses were coded 1 for correct. The second location item asked “Where do you place your food scraps?” with response options including “compost” or “vermicompost bins,” “off-site composting,” “trash,” or “other.” Any option except “trash” and “other” was coded 1 for correct. Some respondents replied “no” to “currently compost,” but wrote in correct locations, such as backyard compost bins or piles. These responses were reclassified as correct locations. In the follow-up survey, only the question about where food was placed was asked and used to determine correct location; “curbside organics cart” was an additional response option and was coded 1 for correct.
A second binary variable was created to represent adequate frequency (0 = no, 1 = yes), using responses to the item “When at home, I dispose of my food scraps separately from my other household trash.” Response options ranged on a frequency rating scale from “never” (1) to “always” (5); responses of 4 or 5 were assigned a score of 1 for adequate frequency.
Based on the scores for correct location and adequate frequency, a third binary variable representing correct target behavior (i.e., composting) was created. Only respondents with scores of 1 for correct location and adequate frequency received a score of 1 for “began composting.”
Began composting
In line with the approach of Thomas et al. (2016) that examined change in a spillover behavior following a change in target behavior (consistent with Figure 1, Model A), we measured adoption of a new target behavior (i.e., beginning to compost) to shed light on causality in the spillover process. Only individuals who were not composting at baseline and hence could adopt this as a new behavior were included in this study. This enabled us to study the effect of performing a new behavior, which may involve new psychological processes, on spillover behaviors (consistent with Figure 1, Model B). Accordingly, those who began composting by follow-up were considered to have begun composting; others were considered as not having begun composting (see Table 1). Differing from Thomas et al. (2016), however, we did not examine change in spillover behavior, which is recommended for future work.
Table 1. Variable Generation for Target Behavior (Began Composting), Including Descriptive Statistics for Relevant Dichotomous Variables at Baseline and Follow-Up (N = 284).
  Baseline Follow-up
  n % n %
Adequate food waste separation frequency
 1 = Separated almost always or more (score >3) 69 24.0 214 75.4
 0 = Separated less than almost always (score < = 3) 215 76.0 70 24.6
Correct food waste disposal location
 1 = Organics cart, compost/vermicompost bin or pile, garden, take to someone with compost bin or pile 10 3.5 220 77.5
 0 = Trash cart, garbage disposal 274 96.5 64 22.5
Correct target behaviora
 1 = Adequate frequency and correct location 0 0.0 203 71.5
 0 = Inadequate frequency or incorrect location 284 100 81 28.5
a
Respondents scoring 1 at baseline were excluded from study.

Other survey items

Other key variables used in the study are listed below (see Table 2)
Table 2. Survey Items and Descriptive Statistics for Key Variables at Follow-Up (N = 284).
Variable M (SD)
Food waste prevention behaviors (1 = strongly disagree, 5 = strongly agree)
 I plan out meals before I shop 3.70 (1.04)
 I try to use the food I have at home before I purchase more of the same food 4.40 (0.78)
 I look to see what food I have at home before I shop 4.42 (0.75)
Energy and water waste prevention behaviors (1 = never, 5 = very often)
 Prompt: How often did you do each of the following in the past 2 months?
  Walk or bike to conserve gas (walked or biked) 2.23 (1.05)
  Unplug devices not in use (unplugged devices) 2.47 (1.18)
  Left lights on in empty roomsa (left lights on) 3.79 (0.87)
  Encouraged others to save home energy (encouraged others) 3.64 (1.10)
  Turned on heater when colda (used heater) 2.50 (0.96)
  Let water run while brushing teetha (let water run) 4.09 (1.03)
  Take shorter showers (took shorter showers) 3.64 (1.08)
Cognitive accessibility (r = .91) 2.85 (1.29)
 Prompt: I consider the water and energy that was used to store and prepare the food:  
  When I throw away leftover meals 2.84 (1.30)
  When I throw away food that I intended to use 2.86 (1.33)
a
Item was reverse coded.
Dependent variables
Food waste prevention. Three items assessed food waste prevention behaviors. These items focused on planning out meals, and assessing/using food at home before shopping. Responses were provided on a Likert scale from “strongly disagree” (1) to “strongly agree” (5). A scale score was created, but due to low internal consistency (Cronbach’s α = .57), each behavior was evaluated as its own dependent variable.
Energy and water waste prevention. Respondents indicated how often they performed seven household water- and energy-related conservation behaviors. Response options ranged on a frequency rating scale from “never” (1) to “very often” (5). A scale score was created, but due to a low Cronbach’s α (.62), we evaluated each as a separate dependent variable.
Mediating variable
Cognitive accessibility. Two items assessed thoughts pertaining to the embedded energy and water wasted when food is disposed of (see Table 2). Responses were provided on a Likert scale ranging from “strongly disagree” (1) to “strongly agree” (5). A mean scale score was calculated as the average of the two items and was used in all mediation analyses.

Dropout Analyses

To test for potential selection bias, we conducted a set of dropout analyses by comparing scores on available baseline variables (i.e., food waste prevention behaviors) among the final sample to scores of respondents who were dropped. Baseline survey respondents were dropped because they were either missing data on key variables (n = 738) or already composting (n = 67). To determine whether dropped respondents significantly differed from the final sample, given nonnormal distributions, we conducted Wilcoxon-Mann-Whitney tests, the nonparametric analogue to the independent-samples t test (Mann & Whitney, 1947; Wilcoxon, 1945).
No significant differences in baseline food waste prevention scores were found between the final sample (plan meals: M = 3.56, Mdn = 4, SD = 1.04; use existing food first: M = 4.38, Mdn = 5, SD = 0.79; check food before shopping: M = 4.39, Mdn = 5, SD = 0.72) and those dropped due to missing data (plan meals: M = 3.68, Mdn = 4, SD = 1.12, U = 92,864, p = .05; use existing food first: M = 4.34, Mdn = 5, SD = 0.76, U = 100,223, p = .93; check food before shopping: M = 4.35, Mdn = 5, SD = 0.82, U = 100,226, p = .87). The final sample scored significantly lower than respondents dropped due to baseline composting on checking food before shopping (dropped: M = 4.63, Mdn = 5, SD = 0.55, U = 7,978, p = .21) but there were no differences for planning meals (dropped: M = 3.75, Mdn = 4, SD = 1.15, U = 8,410, p = .13) or using existing food first (dropped: M = 4.51, Mdn = 5, SD = 1.02, U = 8,673, p = .21) See online Appendix D for demographic characteristics of the final sample relative to other respondents.

Results

STATA IC version 13 was used for all analyses (Statacorp, 2013). We tested each of our hypotheses as follows.

Spillover Behaviors

We tested H1, that beginning to compost would be associated with performance of spillover behaviors both within (food waste prevention) and outside (energy and water waste prevention) the target behavior domain. Specifically, a series of Wilcoxon-Mann-Whitney tests evaluated the impacts of beginning to compost on the set of 10 dependent variables (i.e., seven energy and water waste prevention behaviors and three food waste prevention behaviors). To adjust for multiple comparisons arising from our 10 dependent variables, we applied Holm’s sequential Bonferroni procedure following the equation p H B = ( c − i + 1 ) × p , where pHB is the adjusted p value, C denotes the number of tests per hypothesis (10 in this case), p is the p value prior to the Holm-Bonferroni correction, and i denotes the sequential order of each test, ascending from the lowest uncorrected p value (Abdi, 2010; Cabin & Mitchell, 2000; Holm, 1979). Corrected p values of less than .05 were considered significant. We calculated effect sizes for Wilcoxon-Mann-Whitney tests using r = z /√N, where N is the total number of observations.
Evidence for positive spillover was observed. In particular, beginning to compost was associated with significantly more energy/water spillover behaviors at follow-up. This significant positive spillover was observed for all seven energy and water behaviors. Effect sizes were small, ranging from r = .15 for using the heater (less) to r = .26 for letting water run (less). None of the three food spillover behaviors were significant, except for a marginal effect for checking food before shopping (with a small effect size of r = .14; see Table 3 and Figure 2).
Table 3. Results of Wilcoxon-Mann-Whitney Tests of Spillover From Beginning to Compost to Household Waste Prevention (N = 284).
Dependent variable Median (began composting) Median (did not begin composting) Effect size (r) U p pHBa
Walked or biked 2 2 .19 6,321 .002 .013*
Unplugged devices 3 2 .17 6,444 .003 .017*
Left lights onb 4 4 .18 6,445 .002 .013*
Encouraged others 4 4 .18 6,357 .002 .013*
Used heaterb 3 2 .15 6,677 .009 .036*
Let water runb 5 4 .26 5,654 .0001 .0001***
Took shorter showers 4 3 .24 5,760 .0001 .0001***
Plan meals 4 4 .08 7,404 .165 .165
Use existing food first 5 4 .10 7,306 .101 .201
Check food before shopping 5 4 .14 6,938 .021 .062†
a
pHB refers to p values with Holm-Bonferroni correction.
b
Reverse scored.
†
pHB < .10. *pHB < .05. **pHB < .01. ***pHB < .001.
Figure 2. Mean scores on spillover behaviors among participants who did versus did not begin composting over the course of the study (N = 284).
Note. Holm-Bonferroni-corrected p values from Wilcoxon-Mann-Whitney tests: †pHB < .10. *pHB < .05. **pHB < .01. ***pHB < .001.
aReverse scored item.

Spillover Process

To test H2, that cognitive accessibility will partially mediate the spillover process from composting to household waste prevention behaviors, we conducted mediation analyses with began composting as the independent variable, seven dependent variables (i.e., energy and water waste prevention behaviors) and cognitive accessibility mean scale score as the mediator. Given the null results of beginning to compost on the food waste prevention behaviors, they were excluded from mediation analyses. We used a single-mediator model, defined by three equations,
Equation 1 : y i = β 01 + τ x i + e 1 i
Equation 2 : m i = β 02 + α x i + e 2 i
Equation 3 : y i = β 03 + β x i + τ ′ x i + e 3 i
where yi represents a given spillover behavior, xi represents the independent variable, mi represents the mediator, β01−β03 represent intercept terms, and e1i−e3i represent error terms. In Equation 1, which estimates the impact of the independent variable (i.e., began composting) on the dependent variable (e.g., taking shorter showers), τ represents the direct effect of the independent variable on spillover behavior. In Equation 2, α represents the impact of the independent variable on the mediator (i.e., cognitive accessibility). Equation 3 estimates the simultaneous effects of the independent variable and mediator on the dependent variable (Baron & Kenny, 1986), where τ′ represents the effect of the independent variable on spillover behavior holding the effect of the mediator constant, and β represents the effect of the mediator.
Prior studies of spillover processes have inconsistently reported on whether data met statistical assumptions of normality. For the present study, Cameron and Trivedi’s (2009) omnibus test was used to determine whether our data met normality and heteroscedasticity assumptions. Results indicated that skewness could not be ruled out for any of the seven behaviors (.0001 < p < .02) and heteroscedasticity could not be ruled out for several (i.e., walking or biking, letting water run, encouraging others, and taking shorter showers, .0003 < p <.05). The bootstrapping method (Hayes, 2013) was therefore not suitable, as it is not equipped to handle these issues (Yuan & MacKinnon, 2014).
We examined mediation effects using two methods suited to data with high heteroscedasticity and skewness. First, we used ordinary least squares (OLS) regression with robust Huber-White standard errors in estimating Equations 1 to 3. OLS methods are commonly used in mediation analyses, making results more comparable to mediation analyses in previous work. We also used median regression, which better controls Type 1 error and is recommended as a more robust method to examine mediation relationships when e2i and e3i do not meet the assumptions of homoscedasticity or normality (Yuan & MacKinnon, 2014). Median regression uses Least Absolute Deviations (vs. OLS methods) to estimate Equations 1 to 3. For both methods, we estimated Equations 1, 2, and 3 for each dependent variable. The coefficient of the indirect effect of the mediator (αβ) was calculated from α and β in Equations 2 and 3 above using the product of coefficients method outlined in MacKinnon, Lockwood, Hoffman, West, and Sheets (2002). Additionally, the significance of the indirect effect was tested using the distribution of the product of coefficients method (Tofighi & MacKinnon, 2011). Confidence intervals for indirect effects were computed based on α and β and their standard errors (Tofighi & MacKinnon, 2011). Although median regression is expected to be more robust for our data, it is less common in mediation analyses, so we present both sets of results.
Results from the OLS models with Huber-White standard errors and median regression mediation models are presented in Table 4. OLS results supported cognitive accessibility as a partial mediator of the relationship between beginning to compost and many household waste prevention behaviors. Specifically, significant indirect effects of cognitive accessibility were observed for all energy and water waste prevention behaviors except letting water run while brushing one’s teeth. Median regression results shared some similarities to those based on OLS regression with Huber-White errors. Specifically, median regression models found cognitive accessibility to partially mediate the relationship between beginning to compost and several energy and water waste prevention behaviors (see Table 4). However, there are differences in which behaviors were influenced, as well as the sizes of the indirect effects. Whereas median regression-based indirect effect sizes were zero for many of the behaviors, they were larger for letting water run while brushing one’s teeth, using the heater, and unplugging devices.
Table 4. Indirect Effects of Cognitive Accessibility in Mediation Analyses Between Beginning to Compost and Energy and Water Waste Prevention (N = 284).
Dependent variable Indirect effect size αβ (standard error) Percent (%) reduction in direct effecta
OLS with Huber-White Median regression OLS with Huber-White Median regression
Walked or biked 0.11 (0.04)** n/ab 22.17 n/ab
Unplugged devices 0.09 (0.05)* 0.25 (0.08)*** 18.61 75.00
Left lights onc 0.06 (0.03)* n/ab 17.73 n/ab
Encouraged others 0.14 (0.05)*** n/ab 27.45 n/ab
Used heaterc 0.07 (0.03)* 0.25 (0.10)** 20.74 100.00
Let water runc 0.05 (0.03) 0.25 (0.10)** 6.71 37.50
Took shorter showers 0.09 (0.04)* 0.00 (0.08) 15.75 0.00
Note. OLS = ordinary least squares.
a
When mediator is added to model, defined as (τ − τ′) / τ.
b
n/a because τ = 0, indicating there is no direct effect of beginning to compost on spillover behavior and thus no direct effect to mediate nor to be reduced.
c
Reverse scored.
†
p < .10. *p < .05. **p < .01. ***p < .001.
Figure 3 illustrates median regression results using unplugging devices as a sample behavior. It shows a significant direct effect of beginning to compost on the accessibility of waste cognitions (αca), and a significant direct effect of cognitive accessibility (βca) on unplugging devices not in use. The test of the distribution of the product of the coefficients (Yuan & MacKinnon, 2014) indicates a significant indirect effect of cognitive accessibility on unplugging devices, supporting our mediation hypothesis (H2). This effect was partial: the total direct effect of composting on unplugging devices (τca) decreased considerably (19%) yet remained significant with the mediator in the model (τ′ca). In summary, using both OLS and median regression approaches, cognitive accessibility was supported as a partial mediator of the relationship between composting and some household waste prevention behaviors.
Figure 3. Results of median regression testing cognitive accessibility as a mediator of the association between beginning to compost and unplugging devices when not in use (N = 284).
Note. Coefficient labels α, β, and τ correspond to Equations 1, 2, and 3. CI = confidence interval.
†p < .10. *p < .05. **p < .01. ***p < .001.

Normative messaging and spillover

Wilcoxon-Mann-Whitney tests were used to test the impact of descriptive normative messaging about composting on spillover behaviors (H3a), using the same Holm-Bonferroni correction and effect size calculation as for H1. We found very small effects on spillover behaviors, ranging from r = .01 for use existing food first to r = .11 for unplugging devices, none of which were significant (.59 < pHB < .99; see online Appendix F).
To test whether the norms intervention moderated the impact of beginning to compost on any of the 10 spillover behaviors (H3b), moderation tests were conducted. We regressed each of the 10 outcomes on three independent variables: (a) the “began composting” variable; (b) the dummy-coded moderator variable representing whether or not a participant received social norms messaging (no = 0, yes = 1); and (c) the interaction of these two variables. OLS regression with Huber-White standard errors was used because Cameron and Trivedi’s omnibus test (2009) revealed skewness among all 10 behaviors (.0001 < p < .02) and heteroscedasticity for use existing food first (p = .041), conserve gas (p = .022), and let water run (p = .0001). We applied the same Holm-Bonferroni correction as described for H1 (pHB denotes corrected p values). A number of overall models were significant (let water run: F(3, 280) = 8.44, pHB < .0001; unplugged devices: F(3, 280) = 6.38, pHB < .0003; took shorter showers: F(3, 280) = 6.23, pHB < .0003; walked or biked: F(3, 280) = 5.23, pHB< .02; encouraged others: F(3, 280) = 4.02, pHB < .05. Among these, the only one with a significant Norms × Began composting interaction was for unplugging devices (β = 0.56, p < .05). Among participants who began composting, those who also received norms messaging unplugged devices significantly more frequently than all others (i.e., groups b and a respectively in Figure 4; p < .05); no other comparisons were significant (see Figure 4).
Figure 4. Mean frequency of “unplugged devices not in use” by normative messaging and began composting.
Note. “a” significantly different from “b” at p < .05.

Discussion

The current study found that beginning to manage waste by composting can spill over to energy and water waste prevention, supporting H1. Furthermore, across all water and energy behaviors, results showed positive (vs. negative) spillover. We also found that the relationship between beginning to compost and certain energy and water waste prevention behaviors can be partly explained by greater accessibility of thoughts about waste, providing some support for H2.
Although effect sizes were rather small (ranging from .08-.26), the positive spillover effect was relatively robust across types of behaviors, ranging from taking shorter showers to walking/biking to conserve gas, to encouraging others to save energy. Importantly, these represent different choices that occur in unique contexts, suggesting spillover is not limited to a particular domain or setting. The largest effect sizes were associated with the two water-related spillover behaviors that were assessed (i.e., taking shorter showers: r = .24, letting water run while brushing teeth: r = .26). This may be due in part to the study setting and California’s historic drought, which occurred during the study period. It is possible that beginning to compost and thoughts of waste enhanced the saliency of California’s water shortage and reinforced water conservation behaviors. Surprisingly, no significant spillover was observed among any of the food waste prevention behaviors, with effect sizes ranging from .08 to .14. Given that beginning to compost and food waste prevention can be thought of as falling into the same behavioral domain, we expected to observe stronger spillover for food-related behaviors. Ceiling effects may have contributed to these null results, particularly given that these behaviors were quite prevalent among those who did not begin composting (see Figure 2). Alternatively, preventing food waste often occurs temporally and spatially distal from the point at which food becomes waste, whereas preventing energy and water waste occurs more proximally to the points at which waste is recognized (e.g., leaving water running) and waste prevention acts can occur. Our findings have methodological, theoretical, and practical implications, discussed below.

Methodological Implications

Given inconsistencies in how prior work has examined spillover, the current research sought to advance an integrative process model for studying spillover, and tested one variant of the model (Model B; see Figure 1). Spillover is an inherently causal process in which performing a target behavior influences a seemingly unrelated secondary behavior. The next component of the model, the spillover pathway, represents the psychological processes that explain how adoption of target behaviors leads to spillover behaviors. Whereas many spillover studies omit this component, understanding it is critical to isolating the conditions under which spillover will and will not occur, and whether it will be positive or negative spillover (Truelove et al., 2014). Moreover, to enable attributions of causality, testing this process model requires establishing the temporal order of behaviors, including change in either or both. Thus, a repeated-measures research design is methodologically advantageous, and we encourage its use in future research.
Whereas prior spillover studies inconsistently reported on whether data met statistical assumptions of mediation, we reported this information. Based on the findings, we used both OLS models with Huber-White errors, the results of which are likely more comparable to prior work, as well as median regression, the results of which are likely more robust. Comparing the two sets of findings, differences in effect sizes and patterns of significance are observed, with median regression finding fewer but larger effects. Effects may have differed across models because OLS generally has less power to detect mediation when high nonnormality is present, and tends to inflate Type 1 error when heteroscedasticity is present. These types of errors may be quite common but undetected in mediation analyses due to the rarity of normality tests in previous studies (Yuan & MacKinnon, 2014). The results for unplugging devices and using a heater were supported using both approaches, providing more confidence in mediation involving these behaviors, versus other behaviors that were significant using one method or the other.
Finally, spillover studies often involve multiple comparisons, yet very few address family-wise Type 1 error rates. That we observed significant spillover with a conservative Type 1 error correction speaks to the robustness of our findings. We recommend that future studies using multiple comparisons similarly control Type 1 error (Armstrong, 2014; Perneger, 1998).

Theoretical Implications

Supporting H2, those who began composting reported more thoughts about embedded energy and water in food waste, which in turn was associated with significantly more frequent energy and water spillover for some behaviors. It is possible that cognitive accessibility and the established identity pathway (Truelove et al., 2014) may work together. Because composting involves repeated interaction with waste, when people “looked back” at past behavior relevant to waste (van der Werff et al., 2014), they may have recalled behaviors involved in the composting process, which per our operational definition were performed with relatively high frequency (similar to “many behaviors” condition in Lacasse, 2016), as well as cognitions connected to those behaviors (i.e., waste-related cognitions). Such cognitions may have activated other psychological processes contributing to spillover, discussed below.
First, whereas most spillover studies have targeted relatively easy initial behaviors, such as recycling a plastic bottle once, or reusing towels during a hotel visit, our target behavior (i.e., composting) was relatively more difficult among our sample who were not regular composters at baseline. Difficult target behaviors have been found to strengthen environmental identity when people look back on past PEBs (van der Werff et al., 2014), and to result in greater spillover (Truelove et al., 2014). Difficult behaviors may also activate different spillover processes. Further research is needed to refine our understanding of spillover with “easy” versus “difficult” behaviors (McKenzie-Mohr, 1999; Schultz, 2014).
Second, individuals may construe the notion of waste in a variety of ways, appealing to a broad range of motivations. For individuals with strong proenvironmental identities, waste cognitions may motivate waste prevention behavior due to motivation to protect the environment. However, thoughts of waste may activate alternative motivations such as saving money or making the most of resources one’s household already has. Hence, people who do not consider themselves environmentalists may also find motivation to engage in subsequent waste prevention behaviors following salient thoughts about waste.
Finally, affect may play a role. A plethora of clinical psychology research has found that people can experience automatic thoughts of which they may be only vaguely aware or even not notice. Although fleeting, such cognitions can play a powerful role in giving rise to emotions (Beck, Rush, Shaw, & Emery, 1979). For instance, in a person with highly accessible waste cognitions as a result of composting, one of two hypothetical processes could occur. In the first scenario, she/he may experience automatic thoughts that she/he generates a lot of waste, consumes a lot of resources, or perhaps that she/he is wasteful. If she/he also looks back at past behavior and recalls composting, or engaging in other PEBs, particularly if s/he has a strong environmental identity, new thoughts of wastefulness would conflict with this identity and may lead to negative emotions such as guilt. In other words—thoughts that “I am a wasteful person; however, I’m an environmental person, so I shouldn’t be generating so much waste” could contribute to cognitive dissonance and negative affect. To alleviate this negative state of tension, one may engage in other waste prevention behaviors. In the second possible scenario, thoughts of waste may be tied to waste management, and generate automatic thoughts such as “I am helping with society’s waste management problem,” or “I am a good person,” which could lead someone with an environmental identity to feel proud and perform additional spillover PEBs per the consistency mode (Truelove et al., 2014). In either case, automatic thoughts that arise from performing target PEBs can contribute to emotions that influence subsequent PEBs. Therefore, interventions that aim to modify such thoughts may alter the resulting emotions and later PEBs.

Normative messaging

Descriptive norms did not directly influence performance of spillover behaviors, failing to support H3a. These results roughly align with previous research (Kormos et al., 2015), where normative messaging influenced target behaviors, but had no effect on nontarget behaviors. It is possible that descriptive norms may have limited impacts on behavior beyond specific target behaviors (Schultz, 2014). Alternatively, norms may moderate the spillover process. To this end, Hypothesis 3b received limited support after applying Holm-Bonferroni corrections, but normative messaging did moderate the impact of beginning to compost on frequency of unplugging devices. Specifically, those who began composting and received normative messaging unplugged devices significantly more often than other participants, suggesting that norms “boosted” this particular spillover behavior. That norms boosted spillover only among those who began composting (vs. the whole sample or those who did not) is consistent with prior work finding that in-group norms can enhance intentions to perform norm-consistent behavior (Terry & Hogg, 1996). It is possible that those who began composting perceived the referent group of composters as an in-group, and were motivated to perform other behaviors consistent with group norms, perhaps to maintain positive social identity (Hogg & Reid, 2006). On the other hand, those who did not begin composting may not have viewed the referent group as an in-group, and hence did not perform norm consistent (i.e., spillover) behavior. These processes are speculative given that the present study did not measure social or self-identities; future work should explore these possibilities.

Practical Implications

Our findings suggest that curbside organics collection programs, which have been found to enjoy high participation levels (Geislar, 2017), can have unintended—yet desirable—consequences among those who adopt the target behavior. Results suggest that municipalities should take into account target and spillover behaviors when quantifying the impacts resulting from environmental programs (in this case, waste diversion plus modest water and energy savings). This information can be used to understand how the pursuit of one goal (e.g., landfill diversion) may support or impede the achievement of others (e.g., water conservation).

Limitations and Future Directions

This study should be viewed in light of several limitations. First, the study had considerable attrition. This is a common limitation of longitudinal research, and our attrition numbers align with those from other longitudinal studies using mail-based surveys (Spears, Houston, & Boarnet, 2013). Second, our final sample differed from baseline composters on one key variable (see Dropout analysis section). Related, our sample was older, more educated, and included fewer men and racial minorities relative to the Costa Mesa and U.S. populations (see online Appendix D). Hence, these results may have limited generalizability. Additionally, effect sizes were relatively small, suggesting practical utility is likely modest.
Methodological limitations of the present study include cross-sectional assessment of cognitive accessibility and household waste prevention behaviors (at follow-up only), which limits conclusions about directionality of spillover processes. Future work should assess target and spillover behaviors, as well as potential mediators at multiple time points to flesh out causal processes. In addition, behavior was self-reported and subject to biases, such as social desirability, which may have impacted some participants more than others. For instance, those who began composting may have been more engaged in the study and more likely to be influenced by social desirability than noncomposters, which may have impacted self-reported scores differentially. Also, cognitive accessibility was assessed using double-barreled questions (asking about energy and water waste); future work would be well advised to employ a cleaner measurement approach that separates thoughts about energy versus water waste. Finally, we could not isolate the unique impact of norms messaging on spillover because all participants also received the structural intervention. Future work should examine the unique contributions of individual intervention components to identify tools most likely to result in spillover.

Strengths and Conclusions

This study makes several contributions. First, we advance an integrative process model for studying spillover, and provide one of few experimental field tests of the entire spillover process. Second, this was the first study to test cognitive accessibility as a new factor in the PEB spillover process. We discuss ways in which cognitive accessibility may operate with other established pathways of spillover. Finally, the current study focused on composting, a difficult and understudied target behavior. Future studies should build on this to examine additional difficult target behaviors and spillover pathways, which can help practitioners maximize program benefits (i.e., positive spillover) and avoid unintended consequences (i.e., negative spillover).

Acknowledgments

The authors would like to thank our two blind reviewers for their comments that have greatly strengthened the paper. We would also like to thank Paul Caporaso, Diedre Carr, Lauren Hackney, Katherine Hanna, Bella Kim, Clair Mansour, Sam Sharvini, Mariana Junqueira, Elliott Wezerek, and Noemi Wyss for their work in preparing the data, and Scott Carroll and the Board of Directors at the Costa Mesa Sanitary District for their assistance with project implementation.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Schwarzenegger Institute for State and Global Policy, the Haynes Foundation, the Costa Mesa Sanitary District, the Stanley Behrens Public Impact Fellowship, and the University of California at Irvine Graduate Division and Data Science Initiative.

References

Abdi H. (2010). Holm’s sequential Bonferroni procedure. In Salkind N. (Ed.), Encyclopedia of research design (pp. 574-577). Thousand Oaks, CA: Sage.
Crossref
Google Scholar
Arkes H. R. (1991). Costs and benefits of judgment errors: Implications for debiasing. Psychological Bulletin, 110, 486-498.
Crossref
ISI
Google Scholar
Armstrong R. A. (2014). When to use the Bonferroni correction. Ophthalmic & Physiological Optics, 34, 502-508.
Crossref
PubMed
ISI
Google Scholar
Austin A., Cox J., Barnett J., Thomas C. (2011). Exploring catalyst behaviours: Executive summary: A report to the Department for Environment, Food and Rural Affairs. London, England: Department for Environment, Food and Rural Affairs.
Google Scholar
Baca-Motes K., Brown A., Gneezy A., Keenan E. A., Nelson L. D. (2013). Commitment and behavior change: Evidence from the field. Journal of Consumer Research, 39, 1070-1084.
Crossref
ISI
Google Scholar
Baron R. M., Kenny D. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51, 1173-1182.
Crossref
PubMed
ISI
Google Scholar
Barr S., Shaw G., Coles T., Prillwitz J. (2010). A holiday is a holiday: Practicing sustainability, home and away. Journal of Transport Geography, 18, 474-481.
Crossref
ISI
Google Scholar
Beck A. T., Rush A. J., Shaw B. F., Emery G. (1979). Cognitive therapy of depression. New York, NY: Guilford Press.
Google Scholar
Bem D. J. (1972). Self-perception theory. In Berkowitz L. (Ed.), Advances in experimental social psychology (Vol. 6, pp. 1-60). New York, NY: Academic Press.
Crossref
Google Scholar
Bernstad A. (2014). Household food waste separation behavior and the importance of convenience. Waste Management, 34, 1317-1323.
Crossref
PubMed
ISI
Google Scholar
Bratt C. (1999). Consumers’ environmental behavior: Generalized, sector-based, or compensatory? Environment & Behavior, 31, 28-44.
Crossref
ISI
Google Scholar
Cabin R. J., Mitchell R. J. (2000). To Bonferroni or not to Bonferroni: When and how are the questions. Bulletin of the Ecological Society of America, 81, 246-248.
Crossref
Google Scholar
Cameron A. C., Trivedi P. K. (2009). Microeconometrics using Stata. College Station, TX: Stata Press.
Google Scholar
Cialdini R. B., Kallgren C. A., Reno R. R. (1991). A focus theory of normative conduct: A theoretical refinement and reevaluation of the role of norms in human behavior. Advances in Experimental Social Psychology, 24, 201-234.
Crossref
ISI
Google Scholar
Cialdini R. B., Trost M. R., Newsom J. T. (1995). Preference for consistency development of a valid measure and the discovery of surprising behavioral implications. Journal of Personality and Social Psychology, 69, 318-328.
Crossref
ISI
Google Scholar
Combs B., Slovic P. (1979). Newspaper coverage of causes of death. Journalism Quarterly, 56, 837-849.
Crossref
Google Scholar
Cornelissen G., Pandelaere M., Warlop L., Dewitte S. (2008). Positive cueing: Promoting sustainable consumer behavior by cueing common environmental behaviors as environmental. International Journal of Research in Marketing, 25, 46-55.
Crossref
ISI
Google Scholar
De Young R. (2000). New ways to promote proenvironmental behavior: Expanding and evaluating motives for environmentally responsible behavior. Journal of Social Issues, 56, 509-526.
Crossref
ISI
Google Scholar
Evans L., Maio G. R., Corner A., Hodgetts C. J., Ahmed S., Hahn U. (2013). Self- interest and pro-environmental behaviour. Nature Climate Change, 3, 122-125.
Crossref
ISI
Google Scholar
Festinger L. (1957). A theory of cognitive dissonance. Palo Alto, CA: Stanford University Press.
Crossref
Google Scholar
Freedman J. L., Fraser S. C. (1966). Compliance without pressure: The foot-in-the-door technique. Journal of Personality and Social Psychology, 4, 195-202.
Crossref
PubMed
ISI
Google Scholar
Geislar S. (2017). The new norms of food waste at the curb: Evidence-based policy tools to address benefits and barriers. Waste Management, 68, 571-580.
Crossref
PubMed
Google Scholar
Ghani W., Rusli F., Biak D., Idris A. (2013). An application of the theory of planned behavior to study the influencing factors of participation in source separation of food waste. Waste Management, 33, 1276-1281.
Crossref
PubMed
Google Scholar
Gneezy A., Imas A., Brown A., Nelson L. D., Norton M. I. (2012). Paying to be nice: Consistency and costly prosocial behavior. Management Science, 58, 179-187.
Crossref
ISI
Google Scholar
Goldstein N. J., Cialdini R. B., Griskevicius V. (2008). A room with a viewpoint: Using norms to motivate environmental conservation in hotels. Journal of Consumer Research, 35, 472-482.
Crossref
ISI
Google Scholar
Hayes A. F. (2013). Introduction to mediation, moderation, and conditional process analysis. New York, NY: Guilford Press.
Google Scholar
Hogg M. A., Reid S. A. (2006). Social identity, self-categorization, and the communication of group norms. Communication Theory, 16, 7-30.
Crossref
ISI
Google Scholar
Holm S. (1979). A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics, 6, 65-70.
ISI
Google Scholar
Joireman J., Truelove H. B., Duell B. (2010). Effect of outdoor temperature, heat primes and anchoring on belief in global warming. Journal of Environmental Psychology, 30, 358-367.
Crossref
ISI
Google Scholar
Kormos C., Gifford R., Brown E. (2015). The influence of descriptive social norm information on sustainable transportation behavior: A field experiment. Environment & Behavior, 47, 479-501.
Crossref
ISI
Google Scholar
Lacasse K. (2016). Don’t be satisfied, identify! Strengthening positive spillover by connecting pro-environmental behaviors to an “environmentalist” label. Journal of Environmental Psychology, 48, 149-158.
Crossref
Google Scholar
Lanzini P., Thøgersen J. (2014). Behavioural spillover in the environmental domain: An intervention study. Journal of Environmental Psychology, 40, 381-390.
Crossref
ISI
Google Scholar
Lauren N., Fielding K. S., Smith L., Louis W. R. (2016). You did, so you can and you will: Self-efficacy as a mediator of spillover from easy to more difficult pro-environmental behavior. Journal of Environmental Psychology, 48, 191-199.
Crossref
Google Scholar
Li Y., Johnson E. J., Zaval L. (2011). Local warming: Daily temperature change influences belief in global warming. Psychological Science, 22, 454-459.
Crossref
PubMed
ISI
Google Scholar
Lichtenstein S., Slovic P., Fischhoff B., Layman M., Combs B. (1978). Judged frequency of lethal events. Journal of Experimental Psychology: Human Learning, and Memory, 4, 551-578.
Crossref
Google Scholar
MacKinnon D. P., Lockwood C. M., Hoffman J. M., West S. G., Sheets V. (2002). A comparison of methods to test mediation and other intervening variable effects. Psychological Methods, 7, 83-104.
Crossref
PubMed
ISI
Google Scholar
Maiteny P. T. (2002). Mind in the gap: Summary of research exploring “inner” influences on pro- sustainability learning and behaviour. Environmental Education Research, 8, 299-306.
Crossref
Google Scholar
Mann H., Whitney D. (1947). On a test whether one of two random variables is stochastically larger than the other. Annals of Mathematical Statistics, 18, 50-60.
Crossref
Google Scholar
Margetts E. A., Kashima Y. (2016). Spillover between pro-environmental behaviors: The role of resources and perceived similarity. Journal of Environmental Psychology, 49, 30-42.
Crossref
Google Scholar
Mazar N., Zhong C. B. (2010). Do green products make us better people? Psychological Science, 21, 494-498.
Crossref
PubMed
ISI
Google Scholar
McKenzie-Mohr D. (1999). Fostering sustainable behavior: An introduction to community-based social marketing. Vancouver, British Columbia, Canada: New Society Publishers.
Google Scholar
Merritt A. C., Effron D. A., Monin B. (2010). Moral self-licensing: When being good frees us to be bad. Social & Personality Psychology Compass, 4, 344-357.
Crossref
Google Scholar
Pachur T., Hertwig R., Stienmann F. (2012). How do people judge risks: Availability heuristic, affect heuristic, or both? Journal of Experimental Psychology: Applied, 18, 314-330.
Crossref
PubMed
Google Scholar
Parizeau K., von Massow M., Martin R. (2015). Household-level dynamics of food waste production and related beliefs, attitudes, and behaviors in Guelph, Ontario. Waste Management, 35, 207-217.
Crossref
PubMed
Google Scholar
Perneger T. V. (1998). What’s wrong with Bonferroni adjustments. British Medical Journal, 316, Article 1236.
Crossref
PubMed
Google Scholar
Platt B., Goldstein N. (2014). State of composting in the U.S. BioCycle, 55, 19.
Google Scholar
Risen J. L., Critcher C. R. (2011). Visceral fit: While in a visceral state, associated states of the world seem more likely. Journal of Personality and Social Psychology, 100, 777-793.
Crossref
PubMed
ISI
Google Scholar
Rozin P., Haidt J., McCauley C. R. (2008). Disgust. In Lewis M., Haviland-Jones J. M., Barrett L. F. (Eds.), Handbook of emotions (3rd ed., pp. 757-776). New York, NY: Guilford Press.
Google Scholar
Sachdeva S., Iliev R., Medin D. L. (2009). Sinning saints and saintly sinners: The paradox of moral self-regulation. Psychological Science, 20, 523-528.
Crossref
PubMed
ISI
Google Scholar
San Francisco Department of the Environment. (2013). Zero waste. Retrieved from www.sfenvironment.org/zero-waste/overview/zero-waste-faq
Google Scholar
Schley D. R., DeKay M. L. (2015). Cognitive accessibility in judgment of household energy consumption. Journal of Environmental Psychology, 43, 30-41.
Crossref
Google Scholar
Schultz P. W. (2014). Strategies for promoting proenvironmental behavior. European Psychologist, 19, 107-117.
Crossref
Google Scholar
Spears S., Houston D., Boarnet M. G. (2013). Illuminating the unseen in transit use: A framework for examining the effect of attitudes and perceptions on travel behavior. Transportation Research Part A, 58, 40-53.
Crossref
ISI
Google Scholar
StataCorp. (2013). Stata statistical software: Release 13. College Station, TX: Author.
Google Scholar
Steinhorst J., Klöckner C. A., Matthies E. (2015). Saving electricity—For the money or the environment? Risks of limiting pro-environmental spillover when using monetary framing. Journal of Environmental Psychology, 43, 125-135.
Crossref
ISI
Google Scholar
Stern P. C., Dietz T., Abel T., Guagnano G. A., Kalof L. (1999). A value-belief-norm theory of support for social movements: The case of environmentalism. Human Ecology Review, 6, 81-97.
Google Scholar
Susewind M., Hoelzl E. (2014). A matter of perspective: Why past moral behavior can sometimes encourage and other times discourage future moral striving. Journal of Applied Social Psychology, 44, 201-209.
Crossref
Google Scholar
Terry D. J., Hogg M. A. (1996). Group norms and the attitude-behavior relationship: A role for group identification. Society for Personality and Social Psychology, 22, 776-793.
Crossref
ISI
Google Scholar
Thøgersen J. (1999). Spillover processes in the development of a sustainable consumption pattern. Journal of Economic Psychology, 20, 53-81.
Crossref
ISI
Google Scholar
Thøgersen J., Crompton T. (2009). Simple and painless? The limitations of spillover in environmental campaigning. Journal of Consumer Policy, 32, 141-163.
Crossref
Google Scholar
Thøgersen J., Ölander F. (2003). Spillover of environment-friendly consumer behaviour. Journal of Environmental Psychology, 23, 225-236.
Crossref
ISI
Google Scholar
Thomas G. O., Poortinga W., Sautkina E. (2016). The Welsh single-use carrier bag charge and behavioral spillover. Journal of Environmental Psychology, 47, 126-135.
Crossref
Google Scholar
Tofighi D., MacKinnon D. P. (2011). RMediation: An R package for mediation analysis confidence intervals. Behavior Research Methods, 43, 692-700.
Crossref
PubMed
ISI
Google Scholar
Truelove H. B., Carrico A. R., Weber E. U., Raimi K. T., Vandenbergh M. P. (2014). Positive and negative spillover of pro-environmental behavior: An integrative review and theoretical framework. Global Environmental Change, 29, 127-138.
Crossref
ISI
Google Scholar
Truelove H. B., Yeung K. L., Carrico A. R., Gillis A. J., Raimi K. T. (2016). From plastic bottle recycling to policy support: An experimental test of pro-environmental spillover. Journal of Environmental Psychology, 46, 55-66.
Crossref
Google Scholar
Tversky A., Kahneman D. (1973). Availability: A heuristic for judging frequency and probability. Cognitive Psychology, 5, 207-233.
Crossref
ISI
Google Scholar
Tversky A., Kahneman D. (1974). Judgment under uncertainty. Science, 185, 1124-1131.
Crossref
PubMed
ISI
Google Scholar
United Kingdom Department for Environment, Food, and Agricultural Affairs. (2008). A framework for p-environmental behaviours. London, England: Department for Environment, Food and Rural Affairs.
Google Scholar
U.S. Environmental Protection Agency. (2013a). Municipal solid waste generation, recycling, and disposal in the United States. Washington, DC: Author.
Google Scholar
U.S. Environmental Protection Agency. (2013b). Overview of greenhouse gases. Washington, DC: Author.
Google Scholar
van der Werff E., Steg L., Keizer K. (2014). Follow the signal: When past pro-environmental actions signal who you are. Journal of Environmental Psychology, 40, 273-282.
Crossref
Google Scholar
Visschers V. H., Wickli N., Siegrist M. (2016). Sorting out food waste behavior: A survey on the motivators and barriers of self-reported amounts of food waste in households. Journal of Environmental Psychology, 45, 66-78.
Crossref
Google Scholar
Weber E. U. (1997). Perception and expectation of climate change: Precondition for economic and technological adaptation. In Bazerman M. H., Messick D. M., Tensbrunsel A., Wade-Benzoni K. (Eds.), Environment, ethics, and behavior: The psychology of environmental degradation (pp. 314-341). San Francisco, CA: Jossey-Bass.
Google Scholar
Whitmarsh L., O’Neill S. (2010). Green identity, green living? The role of pro-environmental self-identity in determining consistency across diverse pro-environmental behaviours. Journal of Environmental Psychology, 30, 305-314.
Crossref
ISI
Google Scholar
Wilcoxon F. (1945). Individual comparisons by ranking methods. Biometrics, 1, 80-83.
Crossref
ISI
Google Scholar
Yuan Y., MacKinnon D. P. (2014). Robust mediation analysis based on median regression. Psychological Methods, 19, 1-20.
Crossref
PubMed
ISI
Google Scholar
Zaval L., Keenan E. A., Johnson E. J., Weber E. U. (2014). How warm days increase belief in global warming. Nature Climate Change, 4, 143-147.
Crossref
ISI
Google Scholar

Biographies

Nicole Sintov is assistant professor of Behavior, Decision-making and Sustainability at The Ohio State University’s School of Environment and Natural Resources. As an environmental psychologist, her work focuses on developing and evaluating interventions to promote pro-environmental behavior, and understanding the social and psychological mechanisms that explain processes of behavior change. She holds a PhD in psychology from the University of Southern California.
Sally Geislar is a post-doctoral researcher at the Veritas Research Center at Yonsei University in South Korea. Her research examines how socio-technical systems (e.g., decision-making, technology, trans-national narratives, the built environment,) influence environmental policy outcomes at the food, energy, water nexus. She earned a PhD in Urban Planning and Public Policy from the School of Social Ecology at the University of California, Irvine.
Lee V. White is a PhD candidate at the University of Southern California. Her research focuses on understanding how policy can remove barriers to individual-level pro-environmental behavior, and on how pro-environmental behavior can be encouraged. Much of her research focuses on adoption of technologies such as electric vehicles and solar photovoltaic panels that can displace fossil fuel use.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

Cite article

Cite article

Cite article

OR

Download to reference manager

If you have citation software installed, you can download article citation data to the citation manager of your choice

Share options

Share

Share this article

Share with email
EMAIL ARTICLE LINK
Share on social media

Share access to this article

Sharing links are not relevant where the article is open access and not available if you do not have a subscription.

For more information view the Sage Journals article sharing page.

Information, rights and permissions

Information

Published In

Volume 51, Issue 1
Pages: 50 - 80
Article first published online: November 10, 2017
Issue published: January 2019

Keywords

  1. spillover
  2. behavior change
  3. curbside compost
  4. household conservation
  5. cognitive accessibility

Rights and permissions

© The Author(s) 2017.
Request permissions for this article.

Authors

Affiliations

Nicole Sintov*
The Ohio State University, Columbus, OH, USA
View all articles by this author
Sally Geislar*
University of California, Irvine, CA, USA
Yonsei University, Seoul, South Korea
[email protected]
View all articles by this author
Lee V. White
University of Southern California, Los Angeles, CA, USA
View all articles by this author

Notes

Sally Geislar, Veritas Research Center, Underwood International College, Yonsei University, 85 Songdogwahak-ro IVHB 430, Yeonsu-gu, Songdo-dong, Incheon 406-840, South Korea. Email: [email protected]
*
Shared first authorship.

Metrics and citations

Metrics

Journals metrics

This article was published in Environment and Behavior.

VIEW ALL JOURNAL METRICS

Article usage*

Total views and downloads: 4808

*Article usage tracking started in December 2016

Altmetric

See the impact this article is making through the number of times it’s been read, and the Altmetric Score.
Learn more about the Altmetric Scores


Articles citing this one

Receive email alerts when this article is cited

Web of Science: 43 view articles Opens in new tab

Crossref: 0

  1. How Do Curbside Feedback Tactics Impact Households’ Recycling Performa...
    Go to citation Crossref Google Scholar
  2. Food waste behavioral intention in Islamic universities: the role of r...
    Go to citation Crossref Google Scholar
  3. Food waste tectonics: Points of friction between policy push and pract...
    Go to citation Crossref Google Scholar
  4. If You Build It, Will They Compost? The Effects of Municipal Compostin...
    Go to citation Crossref Google Scholar
  5. The impact of voluntary sustainability adjustments on greenhouse gas e...
    Go to citation Crossref Google Scholar
  6. The influencing mechanism of pro-environmental behaviors in the privat...
    Go to citation Crossref Google Scholar
  7. Emotional responses to COVID-19 stressors increase information avoidan...
    Go to citation Crossref Google Scholar
  8. Trajectory of brain-drain and quality of work-life amongst Nigeria’s u...
    Go to citation Crossref Google Scholar
  9. The Impact of Artificial Intelligence Technology Stimuli on Sustainabl...
    Go to citation Crossref Google Scholar
  10. From past to present (for a better future): The moderating role of cog...
    Go to citation Crossref Google Scholar
  11. The Impact of Emissions Reduction Awareness on Moral Self-Concept: Sus...
    Go to citation Crossref Google Scholar
  12. Examining the Relationship between Consumers’ Food-Related Actions, Wi...
    Go to citation Crossref Google Scholar
  13. Does personal climate change mitigation behavior influence collective ...
    Go to citation Crossref Google Scholar
  14. A Take-Home Message: Workplace Food Waste Interventions Influence Hous...
    Go to citation Crossref Google Scholar
  15. What predicts and prevents source separation of household food waste? ...
    Go to citation Crossref Google Scholar
  16. The more involved, the more willing to participate: An analysis of the...
    Go to citation Crossref Google Scholar
  17. Commitment and intergenerational influence: A field study on the role ...
    Go to citation Crossref Google Scholar
  18. La mise en place d’un système de collecte des déchets ménagers à l’éch...
    Go to citation Crossref Google Scholar
  19. Understanding Food Waste Produced by University Students: A Social Pra...
    Go to citation Crossref Google Scholar
  20. Once Good, Always Good? Testing Nudge’s Spillovers on Pro Environmenta...
    Go to citation Crossref Google Scholar
  21. A Nudge to Quit? The Effect of a Change in Pension Information on Annu...
    Go to citation Crossref Google Scholar
  22. A Randomized Controlled Trial to Address Consumer Food Waste with a Te...
    Go to citation Crossref Google Scholar
  23. Energy Prosumers’ Spillovers and the Policy Effect: Comparing Two Alpi...
    Go to citation Crossref Google Scholar
  24. Guilty pleasures: Moral licensing in climate-related behavior
    Go to citation Crossref Google Scholar
  25. If You Build it, They Will Compost: The Effects of Municipal Compostin...
    Go to citation Crossref Google Scholar
  26. More green than gray? Toward a sustainable overview of environmental s...
    Go to citation Crossref Google Scholar
  27. Personal or planetary health? Direct, spillover and carryover effects ...
    Go to citation Crossref Google Scholar
  28. Climate change, behavior, and the possibility of spillover effects: re...
    Go to citation Crossref Google Scholar
  29. #LetsUnlitterUK: A demonstration and evaluation of the Behavior Change...
    Go to citation Crossref Google Scholar
  30. How Pride Triggered by Pro-environmental Technology Adoption Spills Ov...
    Go to citation Crossref Google Scholar
  31. Positive spillover of consumers’ sustainable behaviors: The mediating ...
    Go to citation Crossref Google Scholar
  32. It’s all about community: On the interplay of social capital, social n...
    Go to citation Crossref Google Scholar
  33. Pro-environmental interventions and behavioral spillovers: Evidence fr...
    Go to citation Crossref Google Scholar
  34. Spillover effects from energy conservation goal-setting: A field inter...
    Go to citation Crossref Google Scholar
  35. Identity and Guilt as Mediators of Pro-environmental Spillover
    Go to citation Crossref Google Scholar
  36. Spillover in the context of forced behaviour change: observations from...
    Go to citation Crossref Google Scholar
  37. “Reduce Food Waste, Save Money”: Testing a Novel Intervention to Reduc...
    Go to citation Crossref Google Scholar
  38. Nudging to promote household waste source separation: Mechanisms and s...
    Go to citation Crossref Google Scholar
  39. A Systematic Literature Review of Concepts and Factors Related to Pro-...
    Go to citation Crossref Google Scholar
  40. Towards a Psychology of the Food‐Energy‐Water Nexus: Costs and Opportu...
    Go to citation Crossref Google Scholar
  41. Behavioral prediction of environmentally oriented anticonsumption and ...
    Go to citation Crossref Google Scholar
  42. Human Behavior and Adaptation
    Go to citation Crossref Google Scholar
  43. 厌恶情绪与消费者行为
    Go to citation Crossref Google Scholar
  44. The influence of environmental identity labeling on the uptake of pro-...
    Go to citation Crossref Google Scholar
  45. The Role of Attitude Strength in Behavioral Spillover: Attitude Matter...
    Go to citation Crossref Google Scholar
  46. How to Measure Behavioral Spillovers: A Methodological Review and Chec...
    Go to citation Crossref Google Scholar
  47. Meta-analysis of pro-environmental behaviour spillover
    Go to citation Crossref Google Scholar
  48. Understanding Contextual Spillover: Using Identity Process Theory as a...
    Go to citation Crossref Google Scholar
  49. Nudge for Good? Choice Defaults and Spillover Effects
    Go to citation Crossref Google Scholar
  50. Contest-Based and Norm-Based Interventions: (How) Do They Differ in At...
    Go to citation Crossref Google Scholar
  51. Using Behavioural Insights to Promote Food Waste Recycling in Urban Ho...
    Go to citation Crossref Google Scholar
  52. Nudge for Good? Choice Defaults and Spillover Effects
    Go to citation Crossref Google Scholar

Figures and tables

Figures & Media

Tables

View Options

View options

PDF/ePub

View PDF/ePub

Get access

Access options

If you have access to journal content via a personal subscription, university, library, employer or society, select from the options below:

Alternatively, view purchase options below:

Purchase 24 hour online access to view and download content.

Subscribe to this journal

Access journal content via a DeepDyve subscription or find out more about this option.

Start 2 week free trial
Need help?

Also from Sage