Abstract
People working on water issues in the state of Florida, USA, recognize outdoor water conservation as an important area of focus. Social marketing has become increasingly accepted as a behavior change approach in Florida, but often the individuals who wish to use social marketing do not have access to the formative audience research needed. In addition to a lack of formative audience research, the prevalence of homeowners’ associations (HOAs) across the state further complicates outdoor water conservation initiatives. This study’s purpose was to evaluate how those who live in HOAs might be considered distinct segments for residential outdoor water conservation interventions. An electronic survey instrument was used to collect data to examine the relationship between theory of planned behavior variables and landscape water conservation behavioral intent in HOA and non-HOA segments. The model fits the two segments similarly. Then, descriptive norms drawn from four distinct referent groups (close-peer, neighborhood, state, and national) were introduced to the theory of planned behavior variables. None of the descriptive norms were significant in the non-HOA model, and close-peer, state, and national norms were significant in the HOA model. State norms had a negative relationship with behavioral intent. Findings revealed clear distinctions in factors relating to behavioral intent between the two segments. Social marketing efforts should treat non-HOA and HOA members as distinct segments and incorporate the normative beliefs that relate to intent among HOA dwellers.
Keywords
Availability of adequate water supply is an increasingly important and complex global issue (Russell & Fielding, 2010; United Nations, 2019). There is an exceptional opportunity to address water availability problems by influencing outdoor water use behaviors in home lawns and yards (Kumar Chaudhary, Warner, & Ali, 2019; Shober, Denny, & Broschat, 2010). As much as half to three-quarters of a household’s water use may occur through watering the lawn and landscape (Milesi et al., 2012), an amount totaling 9 million gallons of water daily across the United States (Environmental Protection Agency [EPA], 2013). This statistic becomes more concerning when considering most yards are watered with potable water (Kjelgren, Rupp, & Kilgren, 2000).
The need to promote water conservation practices is especially pronounced in the state of Florida, USA, due to its substantial use of water, rapidly urbanizing population, unique environment, and extreme population growth (Delorme, Hagen, & Stout, 2010; Shober et al., 2010). Florida ranks fourth in water use among the 50 United States, withdrawing about 1,500 million gallons of water each day for domestic purposes including landscape irrigation (Dieter et al., 2018). An added complexity in Florida is the extent of homeowners’ associations (HOAs), which are recognized as potential drivers of both positive and negative landscape management behaviors and which govern residents’ home and landscape management practices to a degree far more restrictive than that of local government (Felter, Irani, Monaghan, Carter, & Dukes, 2015; Locke et al., 2018).
Across the state of Florida, social marketing is being increasingly recommended and used as a behavior change strategy for residential landscape water conservation (Delorme et al., 2010; Felter et al., 2015; Kassirer, 2019; Monaghan, Ott, Wilber, Gouldthorpe, & Racevskis, 2013; Tampa Bay Water, 2008; Warner, Lamm, Rumble, Martin, & Cantrell, 2016). However, conservation professionals who might potentially conduct social marketing interventions require access to formative audience research to develop successful behavior change campaigns (Huang, Lamm, & Dukes, 2016; Warner, 2014). A lack of understanding of the target audience has been cited as one of the primary barriers to successful residential outdoor water conservation campaigns (Kumar Chaudhary et al., 2019; Warner et al., 2016).
With access to distinguishing characteristics of audience segments, people working on water use and conservation can design more effective social marketing campaigns (Ibrahim, Knox, Rundel-Thiele, & Arli, 2018). One underexplored research area is how differences among people living in HOAs could inform impactful social marketing campaigns, given a dearth of research on the social and economic implications of HOAs (McCabe & Tao, 2006; Meltzer, 2012) and that HOAs are “uniquely capable of both promoting and inhibiting water conservation” (Dyckman, 2008, p. 19). This study was conducted to inform potential audience segmentation strategies using HOA membership for residential outdoor water conservation programs because HOAs are prevalent and influential on residents.
Background
Formative Research and Audience Segmentation
Drawing upon a “fanatical emphasis on the consumer” (Andreasen, 1994, p. 111), successful social marketing campaigns need formative research to understand target audiences’ aspirations, perceptions, current behaviors, needs, and wants (Grier & Bryant, 2005). Formative research allows the practitioner to better understand their target audience, especially when conducted through the lens of behavior change models and theories (Lee & Kotler, 2011). During the formative research process, social marketers make decisions about which audience segments to target and how to strategically incorporate the salient benefits and barriers to position the behavior for each selected segment (Grier & Bryant, 2005).
Audience segmentation can play an effective role in a social marketing campaign but unfortunately “is often absent or done on the fly” (Ibrahim et al., 2018, p. 3). A growing body of recent research has addressed this deficiency and examined audience segmentation strategies for water conservation campaigns. When audience segmentation is used as part of a social marketing strategy, resulting segments should share needs, behaviors, lifestyles, and other characteristics that make them likely to respond in similar ways to a campaign (Lee & Kotler, 2011). When people within the resulting segment have similarities, they communicate more effectively, which supports the diffusion of new ideas or practices (Rogers, 2003).
An audience segmentation study of Florida residents (Warner et al., 2016) yielded three residential landscape water user segments: the unconcerned water users, the water considerate majority, and the water savvy conservationists. These segments were again generated when the study was replicated across the United States (Warner, Kumar Chaudhary, Lamm, Rumble, & Momol, 2017). From another audience segmentation study conducted among students, staff, and faculty of the University of Sharjah in the United Arab Emirates, Ibrahim, Knox, Rundel-Thiele, and Arli (2018) reported three water user segments: the regular, conscious, and careless users. The authors of these three studies reported behaviors and other distinct characteristics relative to water conservation among audience segments, concluding audience segmentation should be integrated into water conservation initiatives.
HOAs and Conservation Practices
Living within an HOA could be a characteristic used to segment a target audience. HOAs are private residential governments established to ensure conformity as a mechanism to maintain common areas (i.e., green spaces and streets) and provide services including maintenance, infrastructure development, and recreational opportunities (McCabe & Tao, 2006; Meltzer, 2012; Weinstein, 2005). Across the country, more than one in five citizens (about 70 million people) live in an HOA (Community Associations Institute [CAI], 2017a , 2017b). These institutions are “quickly becoming the most common and fastest growing units of local governance in the United States” (McCabe, 2005, p. 404). Florida has more residents living in HOAs than any other state and houses 14% of the country’s HOAs (CAI, 2017a , 2017b).
Most HOAs are private nonprofit corporations (McCabe & Tao, 2006) that “ensure a degree of conformity” (Fokidis, 2011, p. 394) through private systems of governance. HOAs control the appearance of the community with rules on landscaping and architecture that are often additive to local governmental ordinances (Meltzer, 2012; Weinstein, 2005). Opting in to a community with mandatory membership signifies the existence of “both perceived and codified social norms within HOAs [that] shape individuals’ landscape water use practices” (Lamm, Lundy, Warner, & Lamm, 2016, p. 46) and could signal important differences that distinguish HOA members from others. For example, high water users are more likely to live in an HOA (Huang et al., 2016). Researchers have reported HOA residents apply more fertilizer to their yards (Fraser, Bazuin, Band, & Grove, 2013) and are less likely to engage in water conservation practices (Monaghan et al., 2013) and recommended specifically targeting HOA residents for water conservation programs. Another Florida audience segmentation study used county population density to inform audience segments and revealed those in the more urban segments were less engaged in several outdoor water conservation practices and also more likely to live in an HOA (Warner, Diaz, & Kumar Chaudhary, 2018). Participants in qualitative studies in Florida and Montana have described HOA rules as barriers to conservation behaviors (Felter et al., 2015; Sigler, 2018).
While HOAs may not expressly require higher levels of yard inputs (i.e., chemicals, water), the look “mandated in many HOA-governed neighborhoods almost always requires the use of them (de facto)” (Fraser et al., 2013, p. 32). However, HOAs can also support water conservation practices (Dyckman, 2008). Although Lamm, Lundy, Warner, and Lamm (2016) operationalized living in an HOA as a constraint, they found more people lived in HOAs in their two water-conserving segments compared to two non-water-conserving segments. In addition, Kumar Chaudhary et al. (2019) reported that within a high-conservation segment, twice as many Florida residents living in HOAs received some type of reward related to their landscape, demonstrating another possible connection between HOAs and conservation.
Theoretical Framework
The theory of planned behavior (TPB; Ajzen, 1991) was used to examine potential differences between HOA and non-HOA residents. The TPB consists of three variables (attitudes toward a behavior, perceived behavioral control over a behavior, and subjective norms pertaining to a behavior; Ajzen, 1991) and has consistently shown good predictive power over citizen’s behavioral intentions (Fishbein, & Ajzen, 2010; Kumar Chaudhary et al., 2017). Behavioral intention is a valuable predictor of actual behaviors and may explain the majority of environmentally responsible practices (Onel, 2017).
Attitude refers to an individual’s evaluation of the outcomes of engaging in the behavior (Ajzen, 1991) or whether the benefits outweigh the costs (Stern, 2018). Perceived behavioral control refers to whether an individual feels they have the autonomy to engage in the behavior and whether it is simple enough for them to do (Ajzen, 1991). Subjective norms refer to perceptions that the behavior is approved of among one’s peers and other important persons (Ajzen, 1991; Fishbein & Ajzen, 2010). A recent Florida-based study found the three TPB variables explained 25% of residents’ intent to adopt outdoor water conservation practices, with subjective norms having the greatest influence (Kumar Chaudhary et al., 2017).
Subjective norms may be further understood as perceptions of what people commonly do (descriptive norms) and what others think should be done (injunctive norms; Cialdini, Reno, & Kallgren, 1990; Schultz, Tabanico, & Rendon, 2008). Descriptive norms tend to have more powerful influences over behavior (Cialdini, 2007) and have been used in most of the successful reported norms interventions (Berkowitz, 2004). Descriptive norms effectively influence behavior because people want to conform to a group (Stern, 2018). Further, descriptive norms give hints as to the behaviors most likely to be appropriate in a situation (Fishbein & Ajzen, 2010).
The study presented in this article evaluated the addition of descriptive norms to the TPB. From their study of green food purchasing practices in Europe, Ham, Jeger, and Ivković (2015) reported descriptive norms predicted behavior, and when added to the standard TPB model, increased its overall explanation of the variance in green food purchasing. Descriptive norms are best understood in the context of a referent group or a set of individuals that can be differentiated in sociocultural terms (Pollis & Pollis, 1970). Referent groups are critically important to the success of a behavior change campaign, but referent groups are often not defined (McKenzie-Mohr & Schultz, 2014).
The influence of norms varies among different populations, behaviors, and scales (Ajzen, 1991; Fishbein & Ajzen, 2010; Locke et al., 2018; Park, Klein, Smith, & Martell, 2009), and it is still not fully understood in which groups and behaviors this approach should be used (Silva & John, 2017). When integrating norms into social marketing campaigns, it is important to avoid highlighting norms from a group an individual dislikes which could cause adverse effects (McKenzie-Mohr & Schultz, 2014). Perceptions of descriptive norms are often different from actual descriptive norms (Larimar et al., 2011). Distinct referent groups are likely to influence people differently, and social marketers should target normative influences accordingly (Meisel & Goodie, 2014). For all of these reasons, norms approaches should not be integrated into social marketing campaigns without a deep understanding of how normative influences operate in a given context (Cho, 2006; Stern, 2018).
Perceived norms are most influential on behavior when respondents identify strongly with the referent group (Cho, 2006; Terry & Hogg, 1996). Perceived descriptive norms are also most accurate for referent groups with which people are most similar and proximal (Berkowitz, 2004; Larimar et al., 2011). For example, Blaine, Clayton, Robbins, and Grewal (2012) reported “very high associations between the lawn care practices of homeowners and those of their neighbors” (p. 266). Lede, Meleady, and Seger (2019) found appealing to an in-group (a specific university and town) to which study participants belonged resulted in greater adoption of water conservation behaviors (i.e., taking shorter showers and signing up for a water-saving retrofit program), compared to general social norms approaches without a referent group specified. Researchers have found individuals more likely to save energy after receiving feedback about people from their own neighborhood when compared to another (Graffeo, Ritov, Bonini, & Hadjichristidis, 2015) or to the region or country (Loock, Landwehr, Staake, Fleisch, & Pentland, 2012), although somewhat generic referent groups may also be effective (Schultz et al., 2008). It is important to decide how to segment an audience as well as measure norm saliency among resulting subgroups prior to designing behavior change programs (Grier & Bryant, 2005; McKenzie-Mohr & Schultz, 2014; Silva & John, 2017).
Method
Approach and Sample
The specific objectives that guided the study were to compare HOA- and non-HOA-dwelling Floridians’ attitudes, perceived behavioral control, subjective norms, and behavioral intent (TPB variables) pertaining to landscape water conservation; compare HOA- and non-HOA-dwelling Floridians’ perceptions of descriptive norms from various referent groups; compare the influence of TPB variables on HOA- and non-HOA-dwellers’ intent to conserve water; and compare how the addition of descriptive norms from various referent groups could improve on the TPB’s prediction of water conservation on HOA- and non-HOA-dwellers’ intent to conserve water. This study employed purposive sampling methods to access a statewide sample of Floridians.
A professional survey sampling company was used to invite potential respondents to participate. After screening to ensure these individuals were Florida residents 18 years and older, there were 1,546 eligible respondents. Of these, 1,173 provided completed surveys that computed to a 75.9% participation rate (participation rate reported as it is impossible to calculate response rate when using an opt-in panel; Baker et al., 2013). Prior to conducting the study, the research protocol was reviewed and approved by the University of Florida institutional review board. As part of the approved protocol, respondents were presented with an informed consent document that outlined the purpose of the study and potential risks and benefits. Participants had to indicate agreement with the informed consent to participate in the study.
The average respondent was White (82.0%, f = 962) and non-Hispanic (80.6%, f = 946). Most respondents (79.8%, f = 936) owned their homes. The most common education level was a 4-year college degree (27.8%, f = 326) followed by some college without a degree (22.6%, f = 264). The most common income category was 25,000–49,999 (23.2, f = 272) followed by 50,000–74,999 (22.7%, f = 267). The average respondent was 49 years of age and had lived in the state for 24 years. About half of the respondents (46.2%) belonged to an HOA. Of those belonging to an HOA, 77.0% said their HOA had policies or requirements related to their landscaping, 74.0% indicated their HOA imposed penalties such as warning letters and fines for the look of their landscape.
Measures
A web-based, researcher-designed survey instrument was used to collect the data needed to address the research objectives. The instrument was programmed to ensure it appeared comparably across different devices and browsers (Dillman, Smyth, & Christian, 2009). Prior to measuring the variables associated with this study, respondents were asked whether they had a lawn and/or landscape that they make decisions about or personally care for. Only those respondents who indicated yes proceeded to answer the study questions. A total of seven input variables included the core TPB variables (attitude, perceived behavioral control, subjective norms) and descriptive norms of four referent groups (close-peers, neighborhood, state, and national).
The TPB items were adapted from Warner, Lamm, and Kumar Chaudhary (2018) who developed their scales following Ajzen (2006) and Warner, Lamm, Rumble, Martin, and Cantrell (2016). Attitude and perceived behavioral control were each measured using a 5-point Semantic Differential Scale. There were 6 items for attitude (good–bad, important–unimportant, foolish–wise, beneficial–harmful, positive–negative, unnecessary–necessary) and 5 for perceived behavioral control (possible for me–not possible for me, easy for me–not easy for me, in my control–not in my control, up to me–not up to me, practical for me–not practical for me). Indexes were created by averaging the responses for both attitude and perceived behavioral control, and they could theoretically range from −2.00 to 2.00.
Subjective norm was measured with a 4-item Likert-type scale with instructions for respondents to indicate their level of agreement or disagreement with four statements, each beginning with the people who are important to me…(…expect me to minimize my use of water when taking care of my lawn/landscape,…expect me to conserve water in my yard,…expect that I will not waste water when taking care of my lawn/landscape,…expect that I will take care of my landscape using the smallest amount of water possible). There was a 5-point response scale: strongly disagree (−2), disagree (−1), neither disagree nor agree (0), agree (1), and strongly agree (2). A subjective norm index was created by averaging the four responses and could therefore theoretically range from −2.00 to 2.00.
The descriptive norms variables were researcher-developed for the purpose of this study. These variables were measured for the four referent groups using a similar 4-item Likert-type scale. Each scale’s 4 items began with an orientation to the referent group: the people who are important to me…(close-peer), the people in my neighborhood…(neighborhood), the people in Florida…(state), the people in the United States…(national). The 4 items for each scale ended in (…minimize their use of water when taking care of their lawn/landscape,…conserve water in their yards,…do not waste water when taking care of their lawn/landscape,…take care of their landscape using the smallest amount of water possible). These all used the same response scale ranging from strongly disagree to strongly agree. The four descriptive norm indexes were created by averaging each scale’s four responses, and each could therefore theoretically range from −2.00 to 2.00.
The outcome variable was behavioral intent, which was adapted from Warner et al. (2016). Respondents received instructions to indicate how unlikely or likely they were to engage in 12 water conservation behaviors in the future. The list of behaviors included practices such as replacing high-water plants with drought-tolerant species, calibrating irrigation sprinklers, and using a rain gauge to monitor rainfall to reduce irrigation. Responses included very unlikely (−2), unlikely (−1), undecided (0), likely (1), and very likely (2). A subjective norm index was created by averaging the four responses and could therefore theoretically range from −2.00 to 2.00.
Prior to conducting the study, a seven-member expert panel was asked to review the instrument to establish face and content validity (Vaske, 2008). The expert recommended minor changes in the wording of a few items. After these changes were completed, a pilot test was conducted with 50 respondents to estimate reliability, test the survey’s programming and procedures, and ensure there were no response issues pertaining to any items (Dillman et al., 2009). All pilot test reliabilities exceeded .70 (data not presented), as estimated using Cronbach’s α, indicating they were appropriate for use (Vaske, 2008). Variable means, standard deviations, correlations, and reliabilities from the full study are presented in Table 1.
Descriptive Statistics, Correlations, and Reliabilities.
Note. Reliabilities reported on diagonal. ATT = attitude; PBC = perceived behavioral control; SN = subjective norm; CPDN = close-peer descriptive norm; NDN = neighborhood descriptive norm; SDN = state descriptive norm; USDN = U.S. descriptive norm; BI = behavioral intent.
** Significant at p ≤ .001.
Data Analysis
Prior to conducting these analyses, data were weighted with post-stratification weighting methods to make sure the data reflected the actual population (Baker et al., 2013, Kalton & Flores-Cervantes, 2003; Lamm & Lamm, 2019). Specifically, individual responses were weighted to match the sex, ethnicity, race, age, and county population density of the 2010 U.S. Census reports for Florida.
Respondents were asked whether their home belonged to an HOA. After removing 33 individuals who were unsure, a sample of 1,140 remained, and it was split closely between those who belonged to HOA (n = 541, 46.2%) and those who did not (n = 599, 51.1%). This variable was used to subdivide the data set into two groups that were labeled HOA and non-HOA segments for the purpose of data analyses.
In the unweighted data set, an independent t-test for equality of means revealed the HOA segment was slightly older (M = 59.40, SD = 16.07) on average compared to the non-HOA segment (M = 56.77, SD = 15.60, t = 2.88, p = .004). In the weighted data set, there was no difference in mean age between the HOA segment (M = 50.73, SD = 19.48) and the non-HOA segment (M = 49.07, SD = 17.27, t = 1.51, p = .131). There was no relationship between living in an HOA and education, family income, rural–urban continuum code.
Independent t-tests for equality of means were used to achieve Objectives 1 and 2. When a statistically significant difference was detected, Cohen’s (1988)d was calculated as a measure of effect size. To address Objective 3, the data set was first split by segment and then identical hierarchical multiple linear regression analyses were conducted for each. In the first block, the core TPB variables (attitude, perceived behavioral control, subjective norm) were entered. In the second block, the four descriptive norms (close-peer, neighborhood, state, and national) variables were introduced to the model, resulting in a non-HOA Models 1 (TPB variables) and 2 (TPB plus descriptive norms) and HOA Models 1 (TPB variables) and 2 (TPB plus descriptive norms). All data were analyzed using SPSS (Version 25.0; IBM Corp., Armonk, NY).
Findings
Objective 1: Compare HOA- and non-HOA-dwelling Floridians’ attitudes, perceived behavioral control, subjective norms, and behavioral intent (TPB variables) pertaining to landscape water conservation.
There were no differences in attitudes or behavioral intent between non-HOA and HOA segments (see Table 2). Non-HOA residents had significantly greater perceived behavioral control over engaging in landscape water conservation practices, while HOA residents perceived stronger subjective norms around landscape water conservation. Objective 2: Compare HOA- and non-HOA-dwelling Floridians’ perceptions of descriptive norms from various referent groups.
Comparison of TPB Variables and Behavioral Intent Among Non-HOA- and HOA-Dwelling Floridians.
Note. Cohen’s (1988) d values were interpreted as 0.2 = small, 0.5 = medium, 0.8 = large. HOAs = homeowners’ associations; TPB = theory of planned behavior.
* Significant at p = .05.
Florida residents believed their close-peers were most engaged in landscape irrigation water conservation and perceived less engagement in conservation as referent groups became more distant (neighborhood, state, national; see Table 3). There was no difference in the perceived norms among respondents’ close-peers. However, there were differences in respondents’ perceptions of neighborhood, state, and national groups’ conservation. Compared to non-HOA residents, HOA-dwellers perceived higher engagement in conservation at each of these levels. Objective 3: Compare the influence of TPB variables and descriptive norms on HOA- and non-HOA-dwellers’ water conservation intent.
Comparison of Perceived Descriptive Norms Among Different Referent Groups Among Non-HOA- and HOA-Dwelling Floridians.
Note.Cohen’s (1988)d values were interpreted as 0.2 = small, 0.5 = medium, 0.8 = large. HOAs = homeowners’ associations.
* Significant at p = .05. **Significant at p ≤ .001.
The results of the hierarchical multiple linear regression analyses (see Table 4) revealed similar predictive power of the TPB variables on HOA-dwellers’ behavioral intent, R2 = .138, F(3, 537) = 28.697, p < .001, compared to non-HOA, R2 = .133, F(3, 593) = 30.478, p < .001.
Results of Hierarchical Multiple Regression of Behavioral Intent Predicted by TPB and Descriptive Norms of Different Referent Groups Among Non-HOA- and HOA-Dwelling Floridians.
Note. Model 1 = core TPB variables (attitude, perceived behavioral control, subjective norms). Model 2 = Model 1 + close-peer descriptive norms + neighborhood. HOAs = homeowners’ associations; TPB = theory of planned behavior.
** Significant at p ≤ .001.
Among the non-HOA residents, all three TPB variables were significant and predicted between 14% and 25% of the variance in behavioral intent, with attitude having a negative relationship with behavioral intent (see Table 5). Of the TPB variables, only perceived behavioral control and subjective norms were significant in the first model among HOA residents.
TPB and Descriptive Norms Variable Coefficients From Hierarchical Multiple Regression of Behavioral Intent Among Non-HOA- and HOA-Dwelling Floridians.
Note. Reported B are unstandardized coefficients and β are standardized regression coefficients. HOAs = homeowners’ associations; TPB = theory of planned behavior.
* Significant at p = .05. **Significant at p ≤ .001.
When the four referent group descriptive norms were introduced to the models, both the non-HOA, R2 = .144, F(4, 590) = 14.190, p < .001, and HOA, R2 = .182, F(4, 533) = 16.988, p < .001, models were again significant with the HOA model having more predictive power. Among non-HOA residents, the three TPB variables remained significant, and of the newly added variables, none of the descriptive norms variables were significant predictors.
Perceived behavioral control and subjective norms remained significant in the second HOA model, and of the newly added descriptive norms variables, perceptions of close-peer, state, and national referent groups’ norms significantly predicted behavioral intent. The relationship between state descriptive norms and behavioral intent was negative while that of close-peer and national descriptive norms was positive. HOA residents with one-unit greater perceived close-peer descriptive norms would be expected to report .207 greater units of behavioral intent. HOA residents with one-unit greater perceived state descriptive norms would be expected to report .217 fewer units of behavioral intent, while HOA residents with one-unit greater perceived national descriptive norms would be expected to report .206 greater units of behavioral intent.
Discussion
The overall purpose of this research was to formatively evaluate HOA-dwellers as a distinct audience segment on a statewide level to inform future social marketing strategies. This research tested the predictive power of the TPB and descriptive norms in the context of water conservation in HOA and non-HOA audiences.
A comparison of HOA and non-HOA segments revealed no differences in attitudes between the two groups. Perceived behavioral control, however, was lower among HOA dwellers. This finding is presumably related to the greater levels of rules and restrictions applying to HOA members’ yards (Shober et al., 2010; Weinstein, 2005). Subjective norms were higher among HOA dwellers, indicating a possible conservation ethic within HOAs. Following the TPB, lower perceived behavioral control should relate to lower conservation intent in the HOA segment, and higher subjective norms could relate to greater intent in the HOA segment (Ajzen, 1991). However, there were no differences in behavioral intent between the two groups, and therefore, we cannot conclude either segment is more prone to outdoor water conservation in general.
Turning to descriptive norms differentiated to four referent groups, there was a trend among all respondents that close-peers were perceived as being most engaged in conservation with progressively lower perceived conservation reported as the referent group became more distant. There were no differences in perceptions of close-peers’ norms when these variables were compared between non-HOA and HOA segments. However, the HOA residents perceived neighborhood, state, and national referent groups were significantly more engaged in conservation when compared to perceptions from non-HOA residents.
The first regression models revealed the TPB model fits the HOA and non-HOA segments equally well, explaining about 13% of the variability in behavioral intent. All three variables were significant in the non-HOA model while only perceived behavioral control and subjective norms were significant in the HOA model. From the first models, we can conclude the non-HOA segment’s behavioral intent is most strongly related to perceived behavioral control when TPB variables are considered alone, while in the HOA segment, behavioral intent is most strongly related to subjective norms.
In regard to the second model of the hierarchical regression analysis for the non-HOA segment, none of the descriptive norms were significant additions to the TPB model. Perceptions of close-peer, neighborhood, state, and national referent groups do not seem to relate to non-HOA members’ intent to conserve, at least when considered in tandem with TPB variables. The second model of the hierarchical regression analysis for the HOA segment differed quite a bit. Descriptive norms from close-peer, state, and national referent groups were significant additions to the TPB model. However, increases in perceived conservation at the state level related to lower intent to conserve, while increases in perceived conservation at the close-peer or national level related to greater intent to conserve. Perceptions of neighborhood norms did not relate to either segment’s intent to conserve.
The current findings (particularly the relationship with close-peer descriptive norms in the HOA segment) partially align with others who have reported appealing to referent groups with which people are most similar or proximal leads to more effective behavior change campaigns (Graffeo et al., 2015; Larimar et al., 2011; Lede, Meleady, & Seger, 2019; Loock et al., 2012). However, the lack of relationship with neighborhood norms within the HOA segment and similar magnitude of relationship between intent and close-peer and (distant) national norms within the HOA segment somewhat contradicts those previous findings. These findings support Graffeo, Ritov, Bonini, and Hadjichristidis’s (2015) contention that when communicating norms, “it is important to [the target audience to] know what others do, but also who those others are” (p. 8). The lack of any relationship with norms in the non-HOA segment suggests descriptive norms may not be a driver of landscape water conservation outside of an HOA setting, at least when TPB variables can be accounted for. Results, however, reveal clear distinctions in factors that relate to behavioral intent among non-HOA and HOA residents. Viewed through the lens of the TPB, behavioral intent is clearly influenced by different factors between the two groups, leading to a conclusion that non-HOA and HOA residents should be considered as distinct segments.
The finding that none of the descriptive norms relate to behavioral intent among non-HOA residents is interesting. It could also be that people outside of an HOA structure inherently wish to be dissimilar to others or perhaps they are simply not influenced by what others do. People may deviate from perceived social norms when they see themselves as generally behaving in better ways than others (Schultz et al., 2008). Given the non-HOA segment perceives three of the four groups are conserving less, compared to the HOA group, it could also be that conservation behaviors are especially less visible and consequently, less influential, to non-HOA residents.
Within the HOA segment, it is especially interesting that neighborhood descriptive norms did not relate to behavioral intention, as would be expected in such communities where consistency is highly valued. The lack of relationship between neighborhood descriptive norms and intent within either segment is in contrast to both Locke et al.’s (2018) and Blaine et al.’s (2012) findings that neighborhood norms were strong drivers of landscape practices. It could be that desire to conform to a norm to have a green and aesthetically-pleasing lawn within a neighborhood (Felter et al., 2015) simply outweighs the desire to be like one’s neighbors in terms of conservation practices. It is also intriguing that state-level norms relate negatively to behavioral intent, while national-level norms relate positively to behavioral intent. Given that HOAs are growing rapidly in Florida, perhaps the HOA segment has, on average, lived in the state for a shorter period of time (as reported by Monaghan et al., 2013), leading to a feeling of disconnect with other Floridians. This explanation would align with Silva and John’s (2017) conclusion that newcomers may lack an awareness of norms. This potential newness and “being from somewhere else” too could explain possible feelings of connection to people on a national level.
Implications for Social Marketing
The findings of this study can be used to inform local social marketing initiatives targeting residential landscape water conservation programs. Firstly, practitioners should consider non-HOA and HOA segments as distinct communities and target them accordingly. The different relationships identified between TPB variables and behavioral intent in the two segments point to different barriers and motivators that must be integrated into social marketing programs. In non-HOA settings, social marketing campaigns might primarily work on increasing perceived behavioral control, while in HOA settings, social marketing campaigns should build greater perceptions of subjective norms. Given that high water users are more likely to live in HOAs (Huang et al., 2016) and that normative feedback can be most effective for the highest resources (i.e., energy, water) users (Alcott, 2011; Brent, Cook, & Olsen, 2015), social marketing practitioners may choose to prioritize the HOA audience segment. Delorme, Hagen, and Stout's (2010) qualitative research participants “were quite enthusiastic” (p. 33) in recommending HOAs as a channel through which to deliver water conservation messages.
Normative appeals are considered promising in the context of water conservation behaviors (Lede et al., 2019) and can be useful when people have inadequate motivation to perform a behavior (McKenzie-Mohr & Schultz, 2014). Because people might tend to underestimate engagement in behaviors because they are not publicly visible (i.e., turning off one’s irrigation), there may be an opportunity to illuminate normative conservation practices. To integrate descriptive norms into behavior change strategies, practitioners need to make residents aware of what others are doing, which can be accomplished through various forms of communication (Nolan, Schultz, Cialdini, Goldstein, & Griskevicius, 2008) including normative feedback and messages through passive media (Schultz et al., 2008). Practitioners may want to reconsider appealing to descriptive norms in non-HOA settings. While more research is needed to explore this phenomenon, the lack of relationship warrants caution in communicating descriptive norms because there seems to be no connection to conservation practices among non-HOA residents. Normative appeals targeting non-HOA members could integrate subjective norms, however. When targeting HOA residents, practitioners should exercise caution when promoting norms on a statewide scale (i.e., most Floridians conserve water in their yard) and consider highlighting norms on a close-peer (i.e., most people who are important to you conserve water in their yard) or national scale (i.e., most people across the United States conserve water in their yard).
Limitations and Recommendations for Future Research
Findings indicate the clear value of segmenting residents by whether they live in HOA or not. Follow-up field testing is needed to examine whether the recommended normative appeals are effective. Message testing with print-, video-, and social-media-based normative appeals using the referent groups tested in this study might be informative for future social marketing campaign strategies.
A few limitations should be addressed. Firstly, this study employed a non-probability sample, and while post-stratification weighting reduced the potential error associated with this type of sampling, caution should be taken when generalizing the results. It would be ideal to replicate the study with a random sample from the general population. There may be pre-existing diversity between HOA and non-HOA audiences that might also explain the differences. Next, the data were self-reported, which introduced potential for respondents to over- or under-report behaviors. While there is strong evidence intention is predictive of behavior lack of actual behavioral data is a possible limitation. Finally, this study considered a suite of both efficiency and curtailment (Russell & Fielding, 2010) water conservation behaviors, and while there is value in promoting complementary behaviors together (Rogers, 2003), there may be different barriers, drivers, and perceptions associated with each practice which need to be examined in the future.
Given HOAs encompass large audiences and geographic coverage with relative conformity, they offer exceptional opportunities for research (Fokidis, 2011). The substantial differences in influence of the referent groups within and between segments hint that there may be variations of group identity at play. Future study should examine these descriptive normative beliefs and add in identification with the referent group variables. It is likely the perceived descriptive norms from referent groups residents identify with the most will be the most influential on behaviors and intention (Fishbein & Ajzen, 2010; Terry & Hogg, 1996). It would be interesting to expand this study to a broader geographic scope and bring in locations where landscape aesthetic preferences may be different.
The outcome variable in this study was behavioral intent, which is well-suited for guiding social marketing efforts as it represents potential for change. However, respondents could feel they already comply with the norms relevant to themselves, and therefore, the normative influences on current behaviors might look different. For this reason, it could be informative to examine how current conservation behaviors relate to the descriptive norms introduced to the TPB in this study.
This study focused on residents themselves, and future researchers might examine structures and social norms of HOA governing bodies that may make key decisions about water use as separate audience segments (Felter et al., 2015). For example, nearly 100% of large HOAs (exceeding 1,000 acres) across the United States provide lawn mowing and similar maintenance of community common areas (McCabe & Tao, 2006), representing additional conservation potential. It may also be beneficial to examine how other factors such as length of time residing in Florida, homeownership, housing type and density, or sociodemographic characteristics could increase the predictive power of these models. In addition to examining the relationships between these factors and conservation behaviors, future work should explore how they may differ between HOA and non-HOA segments.
The current study evaluated how descriptive norms might build upon the TPB and be incorporated into social marketing strategies targeting HOA and non-HOA segments. There may be relationships between the descriptive norms variables and behavioral intent that are overtaken by the strength of the TPB variables, and analyzing these relationships without the TPB could present an additional perspective. For example, the correlations between pairs of study variables demonstrated significant relationships between all four referent groups’ (close-peer, neighborhood, state, national) descriptive norms and behavioral intent among all study respondents before they were segmented by HOA status.
Beyond descriptive beliefs, subjective norms are comprised of injunctive norms (perceptions of what others expect), which can be further explained by individuals’ motivation to comply (Cialdini et al., 1990). Future study should examine perceptions of injunctive norms as well as motivation to comply with those norms. While descriptive norms may be more influential on behaviors than injunctive norms (Cho, 2006; Schultz, 2008), influences are audience- and context-specific, and therefore, more research is needed to inform landscape water conservation social marketing campaigns.
Footnotes
Acknowledgment
The author would like to thank University of Florida faculty for participating in an expert panel review of the survey instrument.
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: Funding for this research was provided by the University of Florida, Center for Landscape Conservation and Ecology, and USDA National Institute of Food and Agriculture, Hatch project 1018367.
