An exploratory path analysis of attitudes, behaviors and summer water consumption in the Portland Metropolitan Area

Sustainable Cities and Society 23 (2016) 68–77

Contents lists available at ScienceDirect

Sustainable Cities and Society journal homepage: www.elsevier.com/locate/scs

An exploratory path analysis of attitudes, behaviors and summer water consumption in the Portland Metropolitan Area Jonathan Straus a , Heejun Chang b,∗ , Chang-yu Hong c a b c

Systems Science Program, Portland State University, Portland, OR 97201, United States Department of Geography, Portland State University, Portland, OR 97201, United States Nohad A. Toulan School of Urban Studies and Planning, Portland State University, Portland, OR 97201, United States

a r t i c l e

i n f o

Article history: Received 9 October 2015 Received in revised form 5 March 2016 Accepted 7 March 2016 Available online 11 March 2016 Keywords: Water consumption Water conservation Behavior Path analysis Portland

a b s t r a c t We examined the underlying attitudinal and behavioral factors of summer water consumption among Portland Metropolitan Area households by combining survey responses from households and corresponding empirical water consumption data. Path analysis shows that pro-conservation attitudes regarding water usage (even when controlling for property size and other demographic variables) were strong predictors of actual reductions in summer water consumption. Furthermore, these self-reported attitudes appear to directly impact specific water consumption behaviors identified in the survey, with potentially significant impact in two of three key areas of water conservation strategies: landscaping, adapting conservation technology, but not habitual use. We draw implications for focused educational programs promoting awareness of water conservation issues and monitoring their impacts and efficacy. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Managing our water resources with sufficient regard for environmental factors, efficient infrastructure, and the costs of continued human development is one of the ongoing key challenges in today’s world. Fortunately, water conservation efforts in the U.S. in recent decades have galvanized and mobilized (Lee, Tansel, & Balbin, 2013). Governments at national, state and city levels have implemented water conservation policies, including federal regulatory initiatives from the 1980s and 1990s introducing new standards for the flow technology of toilet, showerheads, and other fixtures consuming water. Furthermore, high-profile information campaigns, such as “turn off the tap” encouraging the more mindful use of water when brushing teeth or gardening, have proliferated across communities. Throughout this period, the country overall has seen substantial water savings from conservation (Coomes, Rockaway, & Rivard, 2010) as directly evident from surpluses monitored by water bureaus, surpassing even the expectations of many optimistic city planners. In fact, in many areas, including the entire states of Arizona, California, Colorado and Oregon, total water consumption has actually decreased, despite continuing growth of the population, the number of utility accounts, and economic activ-

∗ Corresponding author. E-mail address: [email protected] (H. Chang ). http://dx.doi.org/10.1016/j.scs.2016.03.004 2210-6707/© 2016 Elsevier B.V. All rights reserved.

ity (APA, 2013). Thus, water consumption appears to have been highly responsive to at least some of the social, technological, and structural changes resulting from these recent initiatives. While literature on water conservation has traditionally focused on the impacts of top-down structural and institutional factors on water consumption, such as physical infrastructure (property and water-saving technology) and policy (government-based initiatives and laws) (House-Peters & Chang, 2011), fewer attempts have been made to empirically measure decentralized bottom-up ‘soft’ processes leading to conservation outcomes, such as how individual residents are motivated by policy or awareness of water conservation issues and how their specific behaviors or patterns of water usage change accordingly (Sauri 2013; Chang, 2016). However, the understanding we can glean of the dynamics behind water consumption (a function of many individual users’ behaviors at the micro-level) from the vantage point of policies, physical infrastructure (and other solely macro-level variables) is rife with limitations. Given numerous potential intermediary variables by which policies may effect change, most previous studies on such a macro-level is accompanied by the typical problems of spurious correlations; it is not clear, for instance, if areas with strong water-conservation policies incentivize more conservation from its residents, or if the strong water conservation policies themselves are the outgrowth of an already water-conscious community. Furthermore, such ‘hard’ variables may have, at best, complex correlations with water conservation behavior without direct plausible causal links; even using

J. Straus et al. / Sustainable Cities and Society 23 (2016) 68–77

predictors like education and income are problematic, since they may predict better awareness and access to conservation technologies, but also correlate with larger property sizes and more water-consuming amenities (Chang, Parandvash, & Shandas, 2010; Halper, Scott, & Yool, 2012; March & Sauri, 2010; Polebitski & Palmer, 2010). Moreover, from a practical standpoint, government policy or incentives are not necessarily the most natural leverage point for many conservation efforts. In fact, this tortuous translation from public policy to private behavior often results in the phenomenon of “policy resistance” (Sterman, 2000), where a policy or measure’s intended effect is annulled, or even pushed in the opposite direction, through other social-economic feedbacks triggered by the change. A large government project aimed at energy conservation in Australia found that mandating and subsidizing improved thermal insulation in houses, rather than reducing heating energy consumption, and induced residents (presumably motivated by increased comfort and convenience) to wear less clothing indoors (Browne, Jones & Compston, 2011). Examples like these highlight the importance of norms or attitudes about the value of conservation, which are propagated throughout communities and internalized by individuals. Indeed, many contemporary models of conservation behavior incorporate personal attitudes (and the related norms and motivations) as important causal determinants for conservation outcomes. These models are often informed by the psychology of Planned Behavior (Ajzen & Fishbein, 1975; Lam, 1999; Perren & Yang 2015), which recognize attitudes as important antecedents to adopting conservation behavior, as well as field studies where self-identified voluntary and civic engagement appears to have been a catalyst for conservation behaviors such as household recycling (Oskamp, Edwards, & Okuda, 1997). In other words, individual positive attitudes about conservation (its efficacy, importance, etc.) are needed to overcome barriers to conscious conservation behaviors that reduce household consumption and waste. It may be argued that the dynamics of Water Conservation results (at least in large part) from a similar causal chain of attitudes and behaviors. This paper proposes such an Attitude-Behavior-Consumption framework, in particular to predict summer water consumption in the Portland Metropolitan Area, where we explore the plausibility of a causal chain of environmental action through residents’ attitudes, their individual conservation efforts, and empirical water conservation outcomes.

2. Attitudes and water consumption behaviors As the social sides of environmental problems are increasingly recognized, growing pockets of research have been addressing the effects of pro-conservation attitudes (for example, believing that it is a civil responsibility to conserve water by turning off the faucet when not in use) on associated conservation behaviors (i.e. actually turning off the tap). While our understanding of how such attitudes translate into behavior is largely incomplete, general findings corroborate a positive, albeit complex and nonlinear relationship between the strength of pro-conservation attitudes and their associated behaviors. Bamberg, Moser, and Moeser, (2007) found a statistically significant (but modest) main effect between strength of pro-conservation attitudes and active conservation behavior on the part of individuals in their meta-analysis of general environmental conservation behaviors. This limited effect size seems, in large part, due to the dependence of attitude effect on knowledge and constellation of other factors; Basic understanding of conservation and environmental issues, for instance, was an essential requirement to begin adapting these behaviors (Burgess, 1988). It should be noted, however, that additional training and technical

69

knowledge beyond the basics delivered decreasing returns, with personal attitudes instead becoming the most important cognitive determinant of pro-environmental behavior (Kollmuss & Agyeman, 2002). Nonetheless, the impact of attitudes is limited and sensitive to the basic material and practical means of individuals institutionally have (Blake, 1999). Costs, such as economic burden of buying conservation equipment were needed to be perceived as affordable (Lee & Paik, 2011). Normative drives were major antecedents to the effects of pro-conservation attitudes, where people identified increased community identification and engagement, neighbors’ behavior, and sense of collective moral obligation as their primary motivation for reuse, recycling and other conservation behaviors (Burn, 1991; Gamba & Oskamp, 1994). Along the same vein, most people seem adverse to compromising appearances of social normality for the sake of adapting new environmental practices (Costanzo, Archer, Aronson, & Pettigrew, 1986; Sadalla & Krull, 1995). Finally, cognitive accessibility of conservation’s importance (regardless of available information or advanced knowledge of the environment) through personal experiences, such as immersive exposure to pristine natural environments (rivers, forests, etc.) and first-hand experiences of environmental degradation and scarcity also increase the applicability of attitudes to actions (Chawla, 1988). Broadly speaking, these findings suggest the impact of attitudes on pro-environmental behaviors increase with their ease of economic, social, and psychological adaptation (Diekman & Preisendoerfer, 1992), which are in turn reliant on basic technical and community infrastructure satisficing all these dimensions. Thus, when communities can provide these prerequisite conditions, they empower individuals to act and change their daily behaviors according to their ecological values. A few studies extend these findings to water conservation (most notably, Dolnicar, Hurlimann, & Grun, 2012 Moser, Navarro, Ratiu, & Weiss, 2010, and Willis, Stewart, Panuwatwanich, Williams, & Hollingsworth, 2011). Consistent with the general conservation and attitudes literature, cultural and community-level factors were major determinants of water conservation behavior. In a large international study comparing the cultural attitudes and water consumption patterns of major cities in Europe, Asia, and the Americas, Moser et al. (2010) found cultural attitudes toward water and water conservation predicted differences in magnitude of water conservation among the study cities. Areas that did not culturally emphasize water conservation as a pressing mainstream issue, or trust that others would conserve water, hindered individual and organized conservation efforts, even from those who agreed on its importance. However, individual attitudes appeared highly consequential in other regions (including the UK, Australia, and metropolitan areas of the U.S.) having already above-average awareness and information about water conservation, as well as a generally progressive ethos about the environment. Where scarcity was a salient issue (or water issues were otherwise highly publicized) differences in the strength of positive attitudes played a significant mediating effect on water conservation behaviors and sociodemographic factors within cities (Dolnicar et al., 2012). Although these cities typically had fairly high average reported pro-conservation attitudes, differences in conservation behavior were still highly discernible by strength of those attitudes. Along the urban areas of Australia’s Gold Coast, Willis et al. (2011) found tangible differences in water conservation between households reporting very high perception of importance towards water conservation versus those with only moderately strong agreement. Moreover, modest differences in water consumption by income gaps appeared to be explained by these attitudinal clusters. Gilg and Barr (2006) found a similar pattern among households espousing strong pro-environmental and conservation attitudes in Devon, UK. Their findings suggest an environmentalist commitment engen-

70

J. Straus et al. / Sustainable Cities and Society 23 (2016) 68–77

dered by strong attitudes, reinforced by general conservation habits and norms (e.g. energy) may be the keys to substantial water conservation. In sum, the bottlenecks for establishing reliable water conservation behavior seem to differ depending on the stage of development. Places that are relatively uninitiated to conservation efforts, under this theoretical frame, need to ideologically and behaviorally acclimate to environmentally-conscious lifestyles (as we see in the adoption, normalization, and practice of using recycling facilities, a la Berger, 1997). However, in areas with adequately developed physical and educational infrastructure, attitudes towards water conservation appear to be the crucial limiting factor. It is especially in these places that inspiring change in attitudes could catalyze desirable improvements in water conservation behaviors. Furthermore, research suggests such targeted conservation behaviors might be usefully differentiated as Consumption Habits, Technology Adaptation, and Land Use (Monroe, 2003). Habits include behaviors that are a result of regular ongoing routine decisions to use or curtail use of water. For example, some people – as a result of habit – tend to take longer showers than other individuals who might also be more selective about when to wash the car, or deciding when the yard requires watering. Technological Adaptation refers to embracing utilization of infrastructure and equipment that efficiently makes less use of water: for example, installing low-flow, or alternative means of recycling water (e.g. rain and grey-water systems). Land Use pertains to how the property is allocated to different functions and amenities—for instance, how much of a person’s lawn should be covered in grass (as opposed to native plants or dirt), gardening area, or even how big the swimming pool in the backyard is. Intuitively, habits (which require ongoing conscious regard for water conservation and potential lifestyle sacrifices) would seem to have the closest natural correspondence to internalized attitudes, in contrast to technology and land use, which lead to more passive water savings. However, adaptation of technology and ecologically sound landscaping practices typically require some degree of active individual initiative and deliberate adaptation that is usually voluntary (rather than mandated). Community psychologists Costanzo et al. (1986) further argue that voluntary technological adaptation is in fact most responsive to local social influences, neighborhood norms, and ultimately internalized attitudes. Larson, Polsky, Gober, Chang, and Shandas, (2013) point out the various ways in which land-use practices are an often ongoing and intentional exercise of beliefs and values about the environment, although a clear relationship between attitudes and water conservation friendly landscaping practices so far appears mostly confined to ground cover (e.g. types of vegetation/earth/rocks used to cover the area), and is dependent on the flavor of the attitude (i.e. only specific conservation attitudes but not many general pro-environmental attitudes). Given the relative youth and immense practical value of these lines of research, we hope to further investigate the effects of pro-conservation attitudes on these different dimensions of water consumption. 3. Attitudes and summer water consumption in Portland Portland, Oregon, is a highly compelling and potentially illuminating area to further exploratory research on water consumption in light of the above literature. Situated in the Pacific Northwest along the basin of the Columbia and Willamette rivers, Portland is a major city with an estimated population of 619,360 (the greater metropolitan area encompassing over 2.35 million) (Census Gov, 2014). Portland is well known for its progressive environmental practices, from its emphasis on green space with urban grown boundaries, to its integrated water management, making it thus

seen in many respects be a model city for sustainable development by urban planners. Accordingly, Portland has led technological and progressive infrastructural innovation for water conservation (especially since the 1990s) whose developments include adaptation of rain/soil moisture sensors, irrigation controllers, hose timers, faucet aerators, and rebates for ultra-low-flow high efficiency shower-heads, toilets, and washing machines. However, this city faces continuing challenges to managing its water. Portland must contend with one of the steepest population growth rates in the nation at over 1.1% per year (Christensen, 2014). Despite enjoying a water-rich temperate climate with fairly generous precipitation during most of the year, the Portland area is subject to dry spells with intense continuous sun during the summer, often lasting over two months (Peel, 2007), and summer is likely to get drier in the future (Chang & Jung 2010). The ensuing seasonal return of summer residents and tourism (attracted by the weather), and accompanying surge in outdoor activity (including flora-heavy landscaping and other recreational water use) amounts to the summer water consumption nearly doubling that of winter months in the Portland Metropolitan Area (see Fig. 1 below). This co-occurence of both increased water consumption and diminishing water supply with hotter days, in conjunction with the rising average temperatures accompanying climate change, exacerbate this sharp pressure on the PMA’s water supply during summers (Chang, Praskievicz, & Parandvash, 2014; Larson et al., 2013). However, Portland has several socio-demographic characteristics that, in light of the previous literature, are possibly crucial to their community water conservation efforts and may continue to play a mitigating role in discretionary water consumption and the yearly water shortage problem. As individuals, Portlanders are generally renowned for having highly progressive attitudes, especially with respect to environmental and social issues. This may reflect the history of the city and consequent surrounding normative culture of conservation, where Portland has seen the evolution of self-organized networks around citizen involvement of social and environmental issues (Johnson, 2002). This general commitment to environmental attitudes and grassroots action, combined with Portlanders’ strong community identity, has furthermore focused much of its attention on local water issues. PMA residents’ concerned about their water (and pristine water quality) have engaged their community in political action, with recent protests against Nestle’s bottling plants (Food & Water Watch, 2009) in (CBS News Staff, 2012), covering reservoirs (Mesh, 2013) and mixing well-water with the Bull Run water supply. Unique social factors like these may elucidate why, in contrast to studies such as Pintar (2009) where higher education level is generally associated with more water consumption in most urban areas (given such education is linked to more income and larger size properties), highly educated Portlanders do not show higher household water consumption despite having larger property sizes (Hong & Chang, 2014), supporting our notion that ideationallydriven conscious water conservation (as a response to education) may also play a significant role in Portland where the technological and social support for such measures is established. Given these circumstances and the growing interest of conservation literature in attitudes, Portland appears to be an interesting case study on the effects of its resident’s progressive attitudes, strong community orientation and (at least an ostensible) shared sense of urgency on their water conservation behaviors. Assuming the pertinence of the above literature on attitudes to this situation, the strength of progressive pro-environmental attitudes, in conjunction with the community-level infrastructural and social supports, may be meaningfully consequential to PMA residents’ willingness (and actual steps) to conserve water.

J. Straus et al. / Sustainable Cities and Society 23 (2016) 68–77

71

Fig. 1. Single-family residential household seasonal water consumption patterns from 2001 to 2005 with contrast between winter and summer water consumption averages. Based on metered data from 460 households (321 served by the PWB and 139 served by TVWD).

4. Research purpose and questions We conduct an exploratory study of the relations among proconservation attitudinal factors, water conservation strategies, and empirically-measured household summer water consumption in the PMA. This study is divided into three main parts. In the first part, we verify that pro-conservation attitudes (as operationalized in our survey) predicts overall actual reductions in water consumption. Secondly, we test whether those pro-conservation behaviors correspond with water conservation strategies identified in the survey (as they apply to conservation habits, technology, and land use). Finally, we examine which of the water conservation strategies detected in the survey serve as the appropriate intermediary variables to explain the variation in summer water consumption. This is intended to complement and bolster the knowledge gleaned from more traditional modeling approaches of Portland’s water consumption that tend to focus on physical infrastructure (Chang et al., 2010), land cover and climate variables (Breyer, Chang, & Parandvash, 2012; Chang et al., 2010) and regression on sociodemographic variables used by Hong and Chang (2014), all which are more concerned with prediction of water consumption rather than causation. We seek to answer the following research questions.

(1) Are PMA residents pro-conservation attitudes reflected in affirmative responses to water conservation on the survey? (2) Are self-reported pro-conservation attitudes significant predictors of water conservation at the household level in the PMA? Do these remain predictive when controlling for property size and political views? (3) Which of Monroe’s components of water conservation behavior (Habits, Technology, and Landscaping) increases with stronger pro-conservation attitudes?

(4) Do any of the water conservation areas identified in the survey mediate the relationship between pro-conservation attitudes and water consumption?

5. Data and methods 5.1. Data collection Surveys were distributed to 664 households in the Portland Water Bureau (PWB) and Tualatin Valley District (TVWD) service areas, based on their participation in the Consumer Demand Monitoring (CDM) Program (ongoing from 2000 to 2007). The CDM project collected data on water use for a total of 680 households, 442 served by Portland Water Bureau (PWB) and 238 served by Tualatin Valley Water District (TVWD). The sample size was determined based on geographical representativeness and average household water use. The participants’ average annual water use was not statistically different from the single family customer class water use in each water provider area. To cover the geographical representativeness of samples, spatially stratified random sampling was used. With this sampling, samples were taken proportionally by household numbers in each different neighborhood that represents different sociodemographic characteristics in each water provider area. The CDM project began in 1998, initially involving only PWB households, and was expanded in 2001 to include TVWD households. Data collection was most consistent and involved the largest sample of households between 2002 and 2005. The average duration of household participation was 5.1 years, with PWB households tending to participate longer than TVWD households (6.0 years compared to 3.6 years). Individual household participation ranged from less than one year to over eight years. The number of participating households began declining shortly before 2004 and

72

J. Straus et al. / Sustainable Cities and Society 23 (2016) 68–77

Table 1 Operational definitions of variables used in path analysis. Variable

Operationalization

Pro-conservation attitudes

Mean agreement score with water conservation items (listed in Fig. 2) Hours weekly resident reports watering their lawn Hours weekly resident reports using the shower Proportion of low-flow (or energy star) toilets sinks, and dishwashers Use of rain water and recycled water facilities on property Proportion of yard space dedicated to native plants over total proportion of yard dedicated to plants (native plants + potted plants + grass) Negative of summer water consumption (Jun-Sep 2008)

Conservative lawn watering Conservative Washing Low-flow technology Water recycling Native plant scaping

Water conservation

continued to fall through 2008, at which point the project concluded. Because one of the main goal of the CDM study is to investigate outdoor water consumption patterns in relation to weather conditions, only single family residences were considered the CDM program. We sent surveys to all households who participated in the CDM study. Items on the CDM survey broadly included questions about general residence, demographics, attitudes, and water consumption/land-use behaviors (with mostly multiple choice, item list, and Likert-scale responses). Water providers (PWB and TVWD) supplied address-specific water consumption data metered from the households, which we matched to survey responses. Only 162 surveys were returned at the time of the report, 157 which had fully usable data points for our analysis. The questionnaire was mailed to each participating household in mid-July in 2010 with a return envelope enclosed. Reminder post cards were sent to households that initially did not return the survey after three weeks. The survey period was closed at the end of August. The summer of 2010 was not substantially different from other summers in terms of weather and hydrology. According to the streamflow measurement at Bull Run River Near Bull Run (USGS gaging station #14140000), the 2010 water year (563.6 cfs) was slightly wetter than the average flow (508.2 cfs) of the 2000s (2000–2009), but the annual streamflow was very close to the long term average since 1960 (566.8 cfs). However, the summer flow in 2010 (35.9 cfs) was lower than that of the 2000s (51.3 cfs). Given that the surveyed year 2010 was not much different from the other years in the 2000s, it is unlikely that the weather conditions in the survey year did not have any big influence on survey responses.

5.2. Variables We examined several basic geographical features (location, building and lot size, etc.) and socio-demographic attributes (political views, education, income, etc.) of the household respondents. In addition to those variables reported in the survey, we identified several clusters of questions that seemed to best conceptually fit our attitudes-behaviors-WC framework. In particular, we computed composite scaled scores for each household respondent with regard to their explicit endorsement of (water-related) proconservation attitudes, as well as their self-reported conservation behaviors with respect to washing, low-flow technology use, water recycling, and the use of native plants for landscaping (See Table 1 for more detailed operational definition of each variable). Table 2 summarizes the descriptive statistics of these variables and sources.

5.3. Data analysis First, we performed a summary univariate analysis on respondents’ geographic and socio-demographic characteristics (income, household size, geographical distribution, etc.) to augment our confidence that the respondent data on our sample were reasonably representative of the PMA population of interest. Other characteristics, such as specific pro-conservation attitudes, were also examined. Next, we explored bivariate correlations between commonly proposed predictors of water consumption (property-size, income, education, progressive political views, and pro-conservation attitudes) and empirical water consumption, using both the standard Pearson’s r correlation and Spearman’s Rho for purposes of verifying the correlations’ robustness to variable parameterization. The correlation between pro-conservation attitudes and water consumption was further examined with controls, especially for property size. Finally, we used a path analysis (see: Pedhuzar, 1982; Guhathakurta & Gober, 2010) to explore potential sequential relationships between general water conservation attitudes, water conservation behavior, and actual metered water consumption. Specifically, the path analysis examined the impact of stronger proconservation attitudes on various water conservation behaviors reported in the survey (conservative washing, low-flow technology use, water recycling, and native plant landscaping), and how each of these behavioral categories in turn contributes to – or otherwise predict – actual water conservation (i.e. reduced water consumption). We changed the sign of (and accordingly renamed) several of the variables in our path analysis for consistency with our expected positive causal relationship between desirable water consumption behaviors and outcomes (e.g., recoding the negative of Water Consumption simply as Water Savings) and ran one-tailed tests on all the directed relations hypothesizing ␤ > 0. Note these transformations were strictly linear and thus simply changed the sign—but not the magnitude or significance of the standardized beta coefficients. While path analysis has been used for understanding the major biophysical drivers of water use in Phoenix (Guhathakurta & Gober 2010), it has not been used for understanding both attitudinal and behavioral aspects of water use.

6. Results 6.1. Socio-demographics and pro-conservation attitudes Our univariate socio-demographic analysis on household respondents mostly showed characteristics common to PMA residents, including the overall geographic distribution, progressive political views, above-average education, with predominantly White (and low African-American) population. However, the completed surveys appeared to select for individuals of overall higher socioeconomic standing than those indicated by census data of general Portland-Area residents, suggesting that our results may thus be most generalizable to middle and upper class PMA residents. Among the highlights, the median reported household income was nearly $70,000 (compared to a PMA median of less than $52,000) (US Bureau of Census, 2010). Survey respondents also came from more educated backgrounds, with 98% of households having at least a high school graduate, and 76% having one or more members graduating college (in contrast to 90% and 44%, respectively reported for all PMA residents*) (US Bureau of Census, 2010). Respondents’ racial composition was comparable to that of the PMA, although somewhat more white (91% of respondents compared to 79% of PMA residents) (US Bureau of Census, 2010). More central to our study, we found the emergence of some interesting patterns concerning explicit pro-conservation attitudes

J. Straus et al. / Sustainable Cities and Society 23 (2016) 68–77

73

Table 2 Summary statistics of the variables used for analysis.

Water consumption (L/day) Pro-conservation attitudesa Conservative lawn wateringa Conservative washinga Low-flow technology (%) Water recyclingb Native plant scaping (%) Education (degree) Income ($) Property size (m2 ) a b

N (Valid)

N (Missing)

Mean

Median

S.D.

Source

157 162 155 162 162 162 157 157 148 157

5 0 7 0 0 0 5 5 14 5

1218.36 3.565 1.825 2.150 52.1 0.111 14.7 Bachelor 50 K–99 K 774.6

1052.41 3.700 2.000 2.000 50.0 0.000 0.000 Bachelor 50 K–99 K 661.2

778.08 0.657 0.898 0.790 37.8 0.315 22.5 NA NA 464.1

PWB, TVWD Survey Survey Survey Survey Survey Survey Survey Survey Survey

Based on likert scale (1–5) data. Based on binary response.

Fig. 2. Pro-conservation attitudes. Includes items pertaining to awareness, sense of community responsibility, and personal responsibility regarding water conservation.

(see Fig. 2 below). Respondents were most affirmative about their knowledge regarding their water supply (M = 4.21, SD = 0.74), while being least inclined to claim that they actually thought about such issues on a daily basis (M = 3.03, SD = 1.13). However, responses to most other issues tended to cluster around categories: PMA residents in general agreed highly with a sense of personal responsibility – in the abstract – of conserving water through personal actions such as installing efficient water-saving equipment (M = 3.91, SD = 0.79) and through reducing use for baths and showers (M = 3.89, SD = 0.77). They indicated a moderately high awareness of water shortage issues pertaining to summer (M = 3.78, SD = 0.91) and the reasons for them (M = 3.78, SD = 0.92). Respondents, however, tended to see themselves as somewhat less likely to actually to join conservation efforts through community involvement, whether through their water provider company (M = 3.53, SD = 0.91) or even their own neighborhood (M = 3.27, SD = 1.02). Although there appears to be somewhat systematic rank ordering of these agreement-levels between attitude subcategories, it should be noted that the absolute magnitude of these differences between most of these categories were not extremely different, with responses to most attitude items tending largely toward moderately-high agreement, consistent with acquiescence

bias (Watson, 1992) as well as PMA’s reputation for progressive pro-environmental attitudes.

6.2. Exploratory correlations with water consumption Among simple bivariate correlations (Fig. 3), we found, most importantly, that our selected pro-conservation attitudes were significantly correlated with reduced water consumption (r = −0.401, p < 0.001). This effect was also fairly significant with the non-parametric Spearman’s Rho measure (␳ = −0.189, p = 0.018). Progressive political attitudes were also statistically significant predictors of reduced water consumption in the non-parametric case (␳ = −0.209, p = 0.009), albeit less so with the standard Pearson’s correlation coefficient (r = −0.132, p = 0.107). Education was a somewhat less consistent predictor of reduced water consumption, only being significant with Pearson’s r (r = −0.179, p = 0.025) but not with Spearman’s Rho (␳ = −0.047, p = 0.559), while income alone actually predicted somewhat increased water consumption (r = 0.143, p = 0.074, and ␳ = 0.186, p = 0.020) under Pearson’s r and Spearman’s Rho, respectively. Finally, as supported by numerous other studies (Chang et al., 2010; House-Peters, Pratt, & Chang, 2010; House-Peters & Chang, 2011; Breyer et al., 2012; Kontokosta

74

J. Straus et al. / Sustainable Cities and Society 23 (2016) 68–77

Fig. 3. Simple bivariate correlations between commonly associated socioeconomic variables and Water Consumption, including parametric (Pearson) and non-parametric (Spearman Rank) correlation coefficients, with 95% confidence intervals. Note pro-conservation attitudes and property-size are significant predictors of Water Consumption under both metrics.

& Jain, 2015), larger property-size was significantly correlated with increased water consumption in both the normal Pearson’s r (r = 0.183, p = 0.022) and non-parametric case (␳ = 0.291, p < .001). In summary, pro-conservation attitudes were the most robust single predictor for empirical water consumption, having moderately strong (and highly significant) correlation with reduced water consumption. Furthermore, this correlation remained robust when controlling for property-size (r = −0.393, p < 0.001) as well as for political views (r = −0.419, p < 0.001), suggesting not only that pro-conservation attitudes are a major determinant of water consumption in their own right, but also that they may well account for almost all the correlation initially found between political views and water consumption.

6.3. Path analysis Fig. 4 shows the results of path analysis on the sequential relationships between pro-conservation attitudes, intermediary behavioral variables, and water consumption outcomes. From the onset, we found that reported pro-conservation attitudes were predictive of most water-conserving behaviors identified in the survey, with pro-conservation sentiments significantly predicting adaptation of low-flow technology (␤ = 0.250, p = 0.001), Water recycling infrastructure (␤ = 0.131, p = 0.048) and use of native plants in landscaping (␤ = 0.162, p = 0.021). However, private consumption habits did not exhibit any direct positive relation, let alone a significantly

positive relation. The contributions from reported behaviors to savings in water consumption were less clear, with explicitly outdoor measures such as conservative yard watering (␤ = 0.155, p = 0.029) and use of native plants in landscaping (␤ = 0.144, p = 0.040) being significant, but no discernible direct input from any of the indicators in conservation technology. Among the overall paths, then, consumption habits were the most anomalous, showing mostly beta coefficients with a negative sign (opposite the hypothesized positive relationships), suggesting that any causal role these identified habits play between attitude and measurable conservation behaviors are dominated by other underlying variable effects. In contrast, land use (native plant-scaping in our case) showed a continuous path of positive relations, suggesting land use to be a likely mediating factor between pro-conservation attitudes and reduced water consumption. Beyond all behavioral indicators from the survey, the reported property-size was nonetheless still the strongest direct predictor of water consumption (␤ = 0.335, p < 0.001).

7. Discussion and conclusions We have long suspected that peoples’ attitudes (i.e., normative beliefs and values) are important in motivating and shaping their behaviors toward tangible conservation outcomes. Recent research in the psychology of planned behavior has shown that people who feel empowered to make a difference (e.g., through perceived community support and approval of their efforts) are likely

J. Straus et al. / Sustainable Cities and Society 23 (2016) 68–77

75

Fig. 4. Path analysis diagram.

to act on their pro-conservation attitudes in this effective manner. The results of our study suggest that these circumstances apply to water conservation in the PMA, where individual household members’ attitudes were reasonably powerful predictors of actual household water consumption. Personal attitudes in general (e.g., conservation attitudes and political views) were stronger predictors of water conservation than other sociodemographic variables (such as income and education), although pro-conservation attitudes regarding water in particular seem to be the key predictors of actual reduced water consumption, even when controlling for political views and property-size. We also potentially identified more explicitly causal relationships between pro-conservation attitudes and targeted selfidentified water conservation behaviors through our path analysis. Out of Monroe’s three prongs of water conservation behaviors, we found that pro-conservation attitudes strongly predicted household adaptation of conservation technology (both low-flow technology and water recycling) as well as conservation-friendly land use (i.e., more native plants in place of grass and potted plants on one’s property). These results appear congruent with the theory of planned behavior, implying that pro-conservation attitudes (in conjunction with civic power accorded to PMA residents to self-organize neighborhood conservation efforts) compel citizens to adapt the technology and landscaping practices toward intentional conservation outcomes. However, the fact that the opposite relation with personal consumption habits (e.g., showering and watering the lawn) was found most likely reflects (consistent with Domene and Sauri (2006)) the underlying correlation of proconservation attitudes with income and education, which may accompany emphasis on personal hygiene and lawn presentation. Our path analysis was somewhat less successful in showing how individual water conservation behaviors reported in the survey

amounted to reduced water consumption. The only two behavior items that significantly predicted water savings both pertained to outdoor water use (Yard Watering and Native Plant Scaping) which only became apparent when controlling for property size (itself a significant determinant of water consumption). Nonetheless, these findings support the assertions of previous studies in the PMA that the large increases in summer water consumption are mainly attributed to spikes in outdoor water use (e.g., Breyer et al., 2012; Lee, Chang, & Gober, 2015). A likely explanation is that our current survey or model may not identify the target behaviors most impactful to summer water consumption (despite the obvious increase in water conservation from strong pro-conservation attitudes), and thus could benefit from considering a wider array of self-reportable outdoor conservation behaviors. Perhaps most promisingly, we found that land use (specifically, cultivation of native plants) was a significant intermediary factor between pro-conservation attitudes and water conservation. It thus may be plausible to assert that people who are strongly motivated by water conservation attitudes are more inclined to consciously foster native plants (as opposed to potted plants and grass that typically require additional water) in their yards, which may in turn, explain the significant reduction in their household water consumption. Several aspects of this analysis could be improved. As implied by the above results, peoples’ motivations for some water conservation behaviors cannot be understood by general conservation attitudes without considering underlying socioeconomic factors (such as income and education) which alter peoples’ cognitive framing of conservation issues and priorities (Frederiks, Stenner, & Hobman, 2015) and thus should be analyzed as determinants of other attitudinal factors (e.g. the perceived importance of frequent personal washing or presenting a green lawn). Futhermore, while we focused on the effects of individual planned behavior by

76

J. Straus et al. / Sustainable Cities and Society 23 (2016) 68–77

restricting our analysis to single-family residences, the actual data available for use was comprised of measures at different levels, mixing responses from surveys (combining eminently individuallevel variables like attitudes with household-level ones like low-flow technology) with physical water consumption (metered only at the household-level). Jorgensen, Martin, Pearce, and Willis, 2014 show how individual-level behaviors are much better predictors of individual water consumption, which can be difficult to measure. This lack of a clean delineation between strictly residentlevel and shared household variables complicates our prospects of inferring continuous causal chains between attitudes, behaviors and water consumption from our simple path analysis. Moreover, many outdoor water consumption behaviors rely on establishing cumulative infrastructure (including automated irrigation systems, landscaping, and integrated water recycling) or long-term property features influenced by previous owners and neighborhood associations (e.g. population of native plant/trees on the land) over time. Our design may thus greatly benefit from incorporating these into a multi-level analysis (potentially with time-series data) at individual, household, neighborhood, and (if this study is replicated in other geographic locations) city/regional level into our path analysis. A more fundamental challenge, however, is integrating our individual-centered model of water conservation behavior into the larger established body of knowledge on consumer demand management of water. Personal attitudes and other psychological mediators that we propose to motivate conservation are themselves largely responses to the larger surrounding socioecological environment. Obviously, peoples’ sense of urgency to conserve water (e.g., in Salvaggio, Futrell, Batson, & Brents, 2014) reflects conditions of actual water shortages in those communities. Likewise, areas where people share a strong sense of civic engagement and particularly strong pro-conservation attitudes (Willis et al., 2011) reflect the coherence of community social networks and accessibility of neighborhood conservation programs (Berger, 1997). On an even higher-level, much of the current consumer demand for water is also driven by the surrounding legacy of local, state and national policies, which consumers continually respond to not only directly (via incentives and public infrastructure that alter consumption patterns) but to changes mediated through the political ecology (e.g. community adaptation of messaging and internalization of new policies). As seen in growing parts of California (Morin & Stevens, 2015), changes in the biophysical landscape, climate, and legal consumption restrictions constrain peoples’ access to water, and thus shift the behavioral norms of what constitutes necessary water consumption across different socioeconomic classes and conservation ideologies. In sum, top-down influences work in conjunction with bottom-up and endogenous community processes to shift baseline consumer demand of water as well as the introduction of specific conscious conservation behaviors in a mutual feedback loop. Despite the limitations of our exploratory study, the results suggest potential applications of our methods to larger water conservation studies in the near future. The fact we were able to trace a significant degree of actual water consumption back to self-reported general water conservation attitudes and outdoor land use from a generic survey paired with simple metered water consumption is promising. This hints at the strong possibility of systematically tracking cascading effects from peoples’ attitudes through their behaviors to empirical water consumption by simply modifying surveys to examine a broader array of water conservation behaviors (especially pertaining to outdoor consumption) and more deliberately partitioning of attitudes and psychological attributes that drive the behaviors. We could thus evaluate existing educational community programs emphasizing pro-environmental attitudes like BARK (which cultivate appreci-

ation of the Pacific Northwest’s rich natural heritage), as well as more conservation-specific programs targeting awareness of the PMA’s pressing summer water problem, and tapping into a sense of community identity or civic duty; while also gauging the tangible effectiveness of promoting specific conservation behaviors. By iteratively tracking the effect of various attitudinal components along certain bottleneck behaviors and their marginal returns on water savings over time, we can continually refine these programs’ messages and delivery. Pairing attitudinal perception surveys with water consumption data can also offer us insight into the underlying effects and implementation of new household interventions such as Eco-Feedback technologies (Froelich, Findlater, & Landay, 2010), where the users’ perception and motivation to use conservation strategies is influenced in innumerable ways by real-time access to their own consumption data. By implementing these tools, we can effectively coordinate research efforts of local water bureaus (that collect water data) and other city governments (that make program and policy decisions) with internal community efforts toward saving water. With these adaptive programs, Portland will likely be able to mitigate its seasonal potential water shortages over time. More generally, the resulting lines of research may help us innovate better conservation programs in the Portland area (and beyond) that help catalyze positive conservation sentiments into action. Acknowledgements This research was supported by Portland Water Bureau and Institute for Sustainable Solutions at PSU. Views expressed are our own and do not necessarily reflect those of sponsoring agencies. We appreciate two anonymous reviewers whose comments helped strengthen the paper. References Ajzen, I., & Fishbein, M. (1975). Belief, attitude, intention, and behavior: an introduction to theory and research. Reading. MA: Addison-Wesley. APA, 2013. Water demand revolution, Journal of Planning: Vol. August/September, American Planning Association. Bamberg, S., Moser, G., & Moeser, S. (2007). Twenty years after Hines Hungerford, and Tomera: a new meta-analysis of psycho-social determinants of pro-environmental behaviour. Journal of Environmental Psychology, 27(1), 14–25. Berger, Ida E. (1997). The demographics of recycling and the structure of environmental behavior. Environment and Behavior, 29(4), 515–531. Blake, J. (1999). Overcoming the ‘value-action gap’ in environment policy: tensions between national policy and local experience. Local Environment, 4(3), 257–278. Breyer, B., Chang, H., & Parandvash, G. (2012). Land-use, temperature, and single-family residential water use patterns in Portland Oregon and Phoenix, Arizona. Applied Geography, 1–2, 142–151. Browne, C., Jones, H. M., & Compston, P. (2011). ‘Insulating expectations: a dynamical perspective of the assumptions used in Australia’s home insulation program’. In James M. Lyneis, & George P. Richardson (Eds.), 29th international conference of the system dynamics society (pp. 1–11). USA: System Dynamics Society. Burgess, H. (1988). Practical considerations for conservation bleaching. Journal of the International Institute for Conservation–Canadian Group, 13, 11–26. Burn, S. (1991). Social psychology and the stimulation of recycling behaviors: the block leader approach? Journal of Applied Social Psychology, 21(8), 611–629. CBS News Staff (2012). Portland, OR to add fluoride to city’s water supply, CBS News, http://www.cbsnews.com/news/portland-oregon-to-add-fluoride-tocitys-water-supply/. Census Gov, 2014. Population of Portland: http://quickfacts.census.gov/qfd/states/ 41/4159000.html. Chang, H. (2016). Water conservation. In The international encyclopedia of geography. Wiley. Chang, H., & Jung, I. (2010). Spatial and temporal changes in runoff caused by climate change in a complex large river basin in Oregon. Journal of Hydrology, 388(3–4), 186–207. Chang, H., Parandvash, G., & Shandas, V. (2010). Spatial variations of single-family residential water consumption in Portland, Oregon. Urban Geography, 31(7), 953–972. Chang, H., Praskievicz, S., & Parandvash, H. (2014). Sensitivity of urban water consumption to weather and climate variability at multiple temporal scales:

J. Straus et al. / Sustainable Cities and Society 23 (2016) 68–77 the case of Portland, oregon. International Journal of Geospatial and Environmental Research, 1(1). Article 7 Chawla, L. (1988). Children’s concern for the natural environment. Children’s Environments Quarterly, 5(3), 13–20. Christensen, N. (2014). Portland growth slows, but still among U.S. leaders. In Metro News. http://www.oregonmetro.gov/news/portland-growth-census-032714 Coomes, P., Rockaway, J., & Rivard, B. (2010). North america residential water usage trends since 1992. Denver, CO: Water Research Foundation. Costanzo, M., Archer, D., Aronson, E., & Pettigrew, T. (1986). Energy conservation behavior: the difficult path from information to action. American Psychologist, 41(5), 521–528. http://dx.doi.org/10.1037/0003-066X.41.5.521 Dolnicar, S., Hurlimann, A., & Grun, B. (2012). Water conservation behavior in Australia. Journal of Environmental Management, 105, 44–52. Domene, E., & Sauri, D. (2006). Urbanization and water consumption: influencing factors in the metropolitan region of barcelona. Urban Studies, 43(9), 1605–1623. Food & Water Watch. (2009). Help keep a Nestle water bottling plant out of the Columbia River gorge. In Food & Water Watch Fact Sheet. https://keepnestleout. files.wordpress.com/2011/02/oregonlr.pdf Frederiks, E. R., Stenner, K., & Hobman, E. V. (2015). The sociodemographic and psychological predictors of residential energy consumption: a comprehensive review. Energies, 8, 573–609. Froelich, J., Findlater, L., & Landay, J. (2010). The design of eco-feedback technology. Proceedings of the SIGCHI conference on human factors in computing systems, 199, 9–2008. Gilg, A., & Barr, S. (2006). Behavioural attitudes towards water saving? Evidence from a study of environmental actions. Ecological Economics, 57(3), 400–414. Guhathakurta, S., & Gober, P. (2010). Residential land use, the urban heat island, and water use in Phoenix: a path analysis. Journal of Planning Education and Research, 31(1), 40–51. Halper, E., Scott, C. A., & Yool, S. (2012). Correlating vegetation, water use and surface temperature in a semi-arid city: a multi-scale analysis of the impacts of irrigation by single-family residences. Geographical Analysis, 44, 235–257. Hong, C., & Chang, H. (2014). Uncovering the influence of household sociodemographic and behavioral characteristics on summer water consumption in the Portland Metropolitan Area. International Journal of Geospatial and Environmental Research, 1(2). Article 2. House-Peters, L., & Chang, H. (2011). Urban water demand modeling: review of concepts, methods, and organizing principles. Water Resources Research, 47(5). W05401. House-Peters, L., Pratt, B., & Chang, H. (2010). Effects of urban spatial structure, sociodemographics, and climate on residential water consumption in hillsboro, oregon. Journal of the American Water Resources Association, 46(3), 461–472. Johnson, S. (2002). The transformation of civic institutions and practices in Portland, Oregon 1960–1999. pp. 337. Portland, OR. Jorgensen, B. S., Martin, J. F., Pearce, M. F., & Willis, E. M. (2014). Predicting household water consumption with individual-level variables. Environment and Behavior, 46, 872–897. Kollmuss, A., & Agyeman, J. (2002). Mind the Gap: why do people act environmentally and what are the barriers to pro-environmental behavior? Environmental Education Research, 8(3), 239–260. Kontokosta, C., & Jain, R. (2015). Modeling the determinants of large-scale building water use: implications for data-driven urban sustainability policy. Sustainable Cities and Society, 18, 44–55. Lam, S.-P. (1999). Predicting intentions to conserve water from the theory of planned behavior, perceived moral obligation, and perceived water right. Journal of Applied Social Psychology, 29, 1058–1071. http://dx.doi.org/10.1111/j. 1559-1816.1999.tb00140.x

77

Larson, K., Polsky, C., Gober, P., Chang, H., & Shandas, V. (2013). Water education for sustainability: criteria and recommendations. Society & Natural Resources, 27(11), 1213–1222. Lee, M. S., Tansel, B., & Balbin, M. (2013). Urban sustainability incentives for residential water conservation: adoption of multiple high efficiency appliances. Water Resources Management, 27(7), 2531–2540. Lee, S. J., Chang, H., & Gober, P. (2015). Space and time dynamics of urban water demand in Portland, Oregon and Phoenix, Arizona, stochastic. Environmental Research & Risk Assessment, 29, 1135–1147. March, H., & Sauri, D. (2010). The suburbanization of water scarcity in the barcelona metropolitan region: sociodemographic and urban changes influencing domestic water consumption. The Professional Geographer, 62(1), 32–45. Mesh, A. (2013). City will cover its reservoirs, and drain the ones on Mount Tabor, Willamette Week. http://www.wweek.com/portland/blog-30207-city will cover its reservoirs and drain the ones on mount tabor.html Monroe, M. (2003). Two avenues for encouraging conservation behaviors. Human Ecology Review, 10(2), 113–125. Morin, M., & Stevens, M. (2015). California residents cut water use by hefty 29% in May, officials say. Los Angeles Times., retrieved from www.latimes.com Moser, G., Navarro, O., Ratiu, E., & Weiss, K. (2010). Cultural background and environmental context of water perception and use. In Corral Verdugo, V. Garci´ıa Cadena, H. Cirilo, & M. Fri´ıas Armenta (Eds.), Psychological approaches to sustainability: current trends in theory, research and applications (environmental science, engineering and technology series). New York: Nova Science. Oskamp, H., Edwards, S., & Okuda, S. (1991). Factors influencing household recycling behavior? Environment and Behavior, 23(4), 494–519. Pedhuzar, E. J. (1982). Multiple regression in behavioral research: explanation and prediction. New York: Holt, Rinehart & Winston. Peel, M. (2007). Updated world map of the Koppen-Geiger climate classification. Hydrology and Earth System Sciences, 5, 1633–1644. Perren, K., & Yang, L. (2015). Psychosocial and behavioural factors associated with intention to save water around the home: a Greek case study. Procedia Engineering, 119, 1447–1454. Pintar, K. (2009). Water consumption habits of a south-western Ontario community. Journal of Water and Health, 2, 276–292. Polebitski, A., & Palmer, R. (2010). Seasonal residential water demand forecasting for census tracts. Journal of Water Resources Planning and Management-Asce, 136(1), 27–36. Sadalla, E., & Krull, J. (1995). Self-presentational barriers to resource conservation. Environment and Behavior, 27(3), 328–353. Salvaggio, M., Futrell, R., Batson, C., & Brents, B. (2014). Water scarcity in the desert metropolis: how environmental values, knowledge and concern affect Las Vegas residents’ support for water conservation policy. Journal of Environmental Planning and Management, 57(4), 588–611. Sauri, D. (2013). Water conservation: theory and evidence in urban areas of the developed world. Annual Review of Environment and Resources, 38, 227–248. Sterman, J. (2000). Business dynamics: systems thinking and modeling for a complex world. Boston, MA: McGraw Hill. US Bureau of Census, 2010. 2010 US Census retrieved from www.census.gov. Watson, D. (1992). Correcting for acquiescent response bias in the absence of a balanced scale an application to class consciousness. Sociological Methods Research, 21(1), 52–88. http://dx.doi.org/10.1177/0049124192021001003 Willis, R., Stewart, R., Panuwatwanich, K., Williams, P., & Hollingsworth, A. (2011). Quantifying the influence of environmental and water conservation attitudes on household end use water consumption. Journal of Environmental Management, 92, 1996–2009. http://dx.doi.org/10.1016/j.jenvman.2011.03.023