Perceptions of Quality and Household Water Usage: A Representative Study in Jacksonville, FL.
Keywords Water usage - Natural resource management * Risk perception * Survey research * Multivariate Probit
Although estimates differ, bottled water costs consumers between 240 and 10,000 times more than tap water (Font-Ribera et al. 2017; He et al. 2008). Nevertheless, the average consumption of bottled water among Americans continues to increase. According to a 2016 report by the Beverage Marketing Corporation, per capita bottled water consumption was 39 gallon in 2016, while per capita soft drink consumption was 38.5 gallon (Beverage Marketing Corporation 2017). The growth in bottled water usage is widespread, and it is accompanied by significant negative externalities. When compared to tap water, bottled water consumes 1000 to 2000 times more energy to produce (plastic is made from crude oil) and transport, releasing substantial carbon dioxide into the atmosphere (Gleick and Cooley 2009). Furthermore, only about 13% of the empty plastic bottles eventually reach the recycling stream (Didier 2013). Those that end up in landfills generate incinerated toxic byproducts, such as chlorine gas and ash containing heavy metals. Others enter streams, rivers, and the oceans, contributing to contamination of water bodies and harming wildlife. Research has focused on ways of increasing the recycling rate (Viscusi et al. 2012), but so far organized anti-bottled water campaigns have had little impact on consumer behavior, despite the relative price differences and environmental costs.
Previous research suggests that the factors influencing preferences for bottled water include perceptions of safety and quality, convenience, household composition, and skepticism towards public officials and institutions (Bontemps and Nauges 2016; Jeuland et al. 2016; Vasquez et al. 2015; Brox et al. 2003; Urn et al. 2002). Perceptions represent the formation of an individual's state of mental awareness and can change over time due to personal, social and economic factors. For example, both aesthetic and non-aesthetic qualities are closely related to perceptions of risk and safety of drinking water drawn from city water systems (Anadu and Harding 2000). By contrast, attitudes such as trust and confidence are influenced strongly by experience (Tanellari et al. 2015). For example, Chatterjee et al. (2017) found that perceptions of tap water quality are connected to trust and confidence in public authorities. In addition to basic competence and trustworthiness, trust in public authorities also involved a belief that they were capable of effectively addressing any concerns (Alexander et al. 2008). However, trust and confidence can be undermined by a negative reputation and media coverage of common complaints (Abrahams et al. 2000; Hu et al. 2011).
Consumer faith in the superior quality of bottled water is often misplaced (National Realty & Development Corporation 2010; Saleh et al. 2001; Lalumandier and Ayers 2000; Hunter 1993). In the United States, bottled water is considered a food product, and thus is subject to the packaging and quality requirements of the Food and Drug Administration (FDA). Tap water supplied by municipal utilities, by contrast, is regulated by the Environmental Protection Agency (EPA). There is some disagreement as to which regulatory regime is more restrictive (Duncan 2010; U.S. Government Accountability Office 2009). Indeed, the maximum allowable levels for both agencies are identical across 83 different contaminants, and in many cases commercial bottled water is more exposed to plastic contamination than filtered tap water (Royce 2008; Westerhoff et al. 2008). Nevertheless, the perception of quality differences continues to influence consumer choices concerning water usage. In many settings, consumers are more likely to use bottled water as their primary drinking water source when they perceive their drinking water is not safe (Hu et al. 2011; Moffat et al. 2011; Yoo 2005; Abdalla et al. 1992). Viscusi et al. (2015) found that in addition to perceptions of safety, taste and convenience also are major reasons behind consumption of bottled water.
This study examines the factors that influence households' mode of water usage utilizing data drawn from an extensive survey. The methodological approach is similar to Viscusi et al. (2015), but there are important differences. First, they used data from a nationally representative sample, while this study focused on Jacksonville, FL. By concentrating on a single locality, this study is able to hold constant the municipal utility provider and ultimate water source across all of our respondents. This helps ensure that the behaviors modeled are not attributable to municipal or regional differences, and the inferences are driven by within-city variation. In addition, much of the previous research investigated individual water consumption in general across multiple user contexts and did not distinguish between usage in the home and outside of the home. For this study the survey was tailored to examine water usage in the home specifically. Insofar as preferences for bottled water are driven by convenience and portability, this analysis isolates and minimizes these features and affords us the opportunity to explore other relevant factors. Finally, much of the existing research compares bottled water usage to tap water usage. This study contributes to the literature by including and analyzing a third option, filtered tap water, and draws comparisons across the three modes. In particular, evidence is presented that respondents' concerns regarding safety, contamination and sickness linked to unfiltered tap water are associated with increased bottled water usage in the home, but they have no effect on water filter usage. In contrast, relatively aesthetic complaints, such as foul-smelling water, are associated with increased usage of filtered tap water in the home.
Jacksonville is the largest city by population in the state of Florida, with an estimated population of 913,010 (U.S. Census Bureau 2015). The Jacksonville Electrical Authority (JEA) was established in 1895 to own, operate and manage the city electric system. As a consequence of city-county consolidation, JEA was reconstituted as an independent authority in 1967. In 1997 the water and sewer systems also became part of JEA's utility service offerings. Today, JEA is the primary water utility provider in Jacksonville/Duval County, and parts of neighboring St. Johns, Nassau, and Clay counties.
The JEA drinking water system consists of 135 artesian wells, 37 water treatment plants, over 4300 miles of water lines for distribution, and customers' meters. JEA supplies approximately 114 million gallons of water every day to more than 240,000 customers. The ultimate source of water in this area is the Florida aquifer, an underground river that courses through Florida's limestone bedrock. As stipulated by the State of Florida, in a typical year JEA collects more than 45,000 water samples and tests for more than 100 bacteriological and chemical components. Despite JEA's efforts to maintain high quality tap water, it has encountered criticism and fights a constant battle to improve public perceptions. For example, the Environmental Working Group (EWG) examined the quality of water supplies in a large group of major American cities between 2005 and 2009 and the city of Jacksonville ranked 10th worst in their analysis (Environmental Working Group 2012). In addition, the presence of sulfur and iron bacteria in the local water supply, while presenting no major health hazards to humans, nevertheless generates an unpleasant smell that is widely noted within the community. This kind of negative publicity has led to serious and lingering concerns about the quality of local tap water in the area.
Against this backdrop, a telephone survey of JEA customers within Jacksonville/Duval County was designed with the purpose of learning about public attitudes and mode of water use in the home. The Public Opinion Research Lab at the University of North Florida conducted the survey during March 2016. The telephone data bank for the survey was purchased from Scientific Telephone Samples, and the numbers called were randomly generated (random digit dial) using 18,045 Duval County numbers. In total, 529 complete responses were collected, resulting in a 7.1% response rate based on the response rate calculator created by the American Association for Public Opinion Research (2016). This response rate is typical for recent telephone surveys conducted within the U.S. (Kohut et al. 2012).
Definitions and descriptive statistics for the variables used in the analysis appear in Table 1. The first three variables listed in the table record water usage in the home (Unfiltered Tap Water, Filtered Tap Water, Bottled Water). All three alternatives are well-represented in the sample; notably, over half of the respondents reported that they use bottled water. Following the literature, the survey captures respondents' perceptions of water quality with the next four variables (Belief Safe, Belief Contaminated, Concern Sick, Foul Smell). Attitudes toward publicly provided goods and services are sensitive to trust in authorities. Therefore, respondents were asked how much trust they placed in the city government, state government, EPA and JEA. Given that these measures were highly correlated, an average measure was calculated (Distrust Average). In addition, to assess the importance of information and media coverage of common complaints on water usage (Abrahams et al. 2000; Hu et al. 2011), respondents were asked if they heard anything about any potential problems with the local water supply (Heard). The final group of variables records household economic circumstances and demographic details. The typical respondent was between 45 and 64 years old, had some college, lived in owner-occupied housing, and had an annual household income between $50,000 and $75,000. These characteristics closely resemble the 2015 Census report (U.S. Census Bureau 2015), giving confidence that the sample is a fair representation of the targeted population.
Multinomial logit is the traditional framework for modelling individual choice with multiple outcomes, but the well-known independence of irrelevant alternatives (IIA) assumption is inappropriate in this case. The three outcomes in this study are not mutually exclusive in principle, and in the sample approximately 22% of respondents reported more than one mode of water usage in the home. Specifically, the sample correlation between unfiltered water usage and bottled water usage is -0.22 (p < 0.01), the sample correlation between filtered water usage and bottled water usage is -0.40 (p < 0.01), and the sample correlation between unfiltered water usage and filtered water usage is -0.25 (p<0.01). The data do not lend themselves to nested logit or a mixed logit approach, as there are no measures that are 'alternative specific' Instead, we turn to multivariate probit and model mode of water use as a 3-equation system. The multivariate probit framework, of which bivariate probit is a special case, has proven useful across a wide range of applications, including in transportation economics (Castillo-Manzano 2010; Choo and Mokhtarian 2008), energy economics (Baskaran et al. 2013), information and communications economics (Koo et al. 2014) and in agricultural economics (Kpadonou et al. 2017; Weber and McCann 2015).
Let the unobserved latent variable of choice be determined according to an index model
[mathematical expression not reproducible] (1)
where the observed outcomes [mathematical expression not reproducible] and 0 otherwise. Perceptions of tap water quality, health concerns, trust in government and household demographics are collected in the vectors Xm. The errors em are assumed distributed trivariate normal with means equal to zero and variances equal to 1. The key departure from the standard multinomial setting is that the error correlations across the equations are free parameters to be estimated. Evaluating the trivariate normal integrals and their derivatives can be computationally difficult. The approach of Cappellari and Jenkins (2003) is followed and the Geweke-Hajivassiliou-Keane (GHK) simulator employed to estimate the parameters [beta]m and error correlations, [rho]l2 = [rho]21, [rho]23 = [rho]32, [rho]13 = [rho]31, via maximum simulated likelihood.
Table 2 presents the first set of results. At the bottom of the table are the estimates and standard errors for the cross-equation error correlations. All three are negative and statistically significant. This reinforces the choice of modelling approach insofar as neither a traditional multinomial model, nor a series of independent probit models, would adequately capture these cross correlations. Specifically, the negative values indicate general substitutability across the modes of water use. For example, a shock that increases the probability of using unfiltered tap water simultaneously decreases the probability of using bottled water, and vice versa. This pattern extends to all three pairwise comparisons.
The value of the simultaneous equation approach is also demonstrated in the table of coefficients where the perceptions of water quality have asymmetric effects on the modes of water use. The variables Belief Safe, Belief Contaminated, Concern Sick, and Foul Smell were constructed in such a way that higher values correspond to agreement with the implied question. Thus, higher values on Belief Safe indicate that the respondent agrees that, in general, their unfiltered tap water is safe to drink. Such a belief increases the probability of using unfiltered tap water in the home (significant positive coefficient) and decreases the probability of using bottled water in the home (significant negative coefficient). However, Belief Safe does not have a statistically significant effect on the use of filtered tap water. Our interpretation of this result is that respondents do not view residential water filtering systems as an adequate solution to perceived problems of water safety. That is. as respondents become more concerned with perceived tap water quality, they rum to bottled water rather than water filters. This interpretation is reinforced by the results from a somewhat stronger form of this question, Belief Contaminated. In this case, respondents' concerns over contaminated water led them to be less likely to use unfiltered tap water and less likely to use filtered tap water. Finally, those who responded that they were most concerned about themselves or family members falling ill from drinking tap water. Concern Sick, were more likely to rum to bottled water also.
More aesthetic concerns are captured by the variable Foul Smell. The more respondents agree that the unfiltered tap water in their homes smells bad, the less likely they are to use unfiltered tap water (significant negative coefficient). Interestingly, this belief also makes respondents less likely to use bottled water (significant negative coefficient). Instead, these respondents are more likely to use filtered tap water (significant positive coefficient). Whereas water filtering systems are not deemed as adequate substitutes when respondents are concerned about safety, they are viewed as the preferred alternative when respondents are more concerned with cosmetic features.
The measure of distrust in government agencies and institutions did not have a statistically significant effect on any of the modes of water use.' However, the variable Heard had a significant positive effect on the use of unfiltered tap water. This is a puzzling result, as this variable captures whether respondents have "heard about any potential problems that your town or city supplier has, or had, with their water supply." The local water supplier, JEA, has at times acquired a rather negative reputation for water quality, and frequent anecdotal episodes tend to reinforce this perception in the local news media. We originally hypothesized that those who had been most exposed to this information would be the least likely to use unfiltered tap water. The estimates suggest just the opposite. There are several potential explanations for this counterintuitive result. First, it may be that distrust of the local media to some degree offsets any negative message about water quality, although this would tend to support the null hypothesis. Second, JEA has engaged in an active public relations campaign to correct misperceptions and reassure local residents about the quality of their tap water (Jacksonville Electric Authority 2017). It may be that this message has become the dominant narrative within the community, and so those who have been exposed to this message respond in a positive way towards the tap water. Finally, this unexpected estimate may be the result of differences in the salience of this information across respondents. That is, those who are more likely to use unfiltered tap water in the home are also more likely to pay attention to news stories and articles about water quality. These respondents have not switched their mode of water use, possibly for economic or other reasons, and thus they are vulnerable to poor tap water quality. As a result, they are more likely to internalize the negative message and recall it when asked by an interviewer. By contrast, information about the quality of local tap water is not as salient to those who have already substituted away from unfiltered tap water. They do not pay close attention to local news stories about this issue and do not recall hearing about them when asked by an interviewer. This asymmetric response resembles examples of 'intentional blindness' (Simons 2000). Unfortunately, the current data are not sufficiently detailed to distinguish between these and other potential explanations. Nevertheless, we intend to continue to explore this intriguing result in future research.
The model also included a series of variables designed to capture demographic and economic characteristics. Higher levels of education increased the probability of using filtered water, as did home ownership, while households with a larger fraction of children were more likely to use unfiltered tap water. We suspect these results may be driven in part by household income, so we re-estimated the model including a control for household income. The new estimates appear in Table 3. Including income in the model reduced the sample size by almost 25% due to missing data, so precise comparisons were not possible. However, the overall results are qualitatively similar to the original model, while education, home ownership and family composition ceased to have any predictive power. Higher levels of household income decreased the probability of using unfiltered tap water in the home and increase the probability of using filtered water. This is to be expected, as purchasing, installing and maintaining water filtering systems in the home is a quality upgrade that involves costs over and above that of unfiltered tap water. Bottled water is more expensive still, yet the evidence indicates that income does not appear to have any effect on the probability that bottled water is used in the home.
One final result of note, potentially related to the disparity in income effects noted above, is the estimate for the indicator of African-American respondents. The baseline racial category is Caucasian, so in both specifications of the model we interpret this coefficient as suggesting that African-American respondents are more likely to use bottled water and less likely to use filtered tap water than are Caucasian respondents. This result persists even after controlling for household income, home ownership, family size and composition. The city of Jacksonville has a long history of racially segregated residential patterns. We speculate this finding reflects a relative cost and access differential for tap water alternatives across neighborhoods within the JEA service area. Unfortunately, this is an area where the data are not detailed enough to pinpoint the source of this effect. However, it does suggest another promising avenue for future research that examines the economic and racial geography of water usage.
The results of this study raise several important questions that cannot be addressed with the current data. For example, focusing on a single municipal area enhances internal validity while limiting external validity. The estimation results are driven by within-city variation, allowing us to isolate individual influences on behavior at the microeconomic level, but they do not speak to differences across water sources or municipal systems. In addition, the survey is limited to measuring perceptions of quality at a point in time, but perceptions likely are formed over time and are influenced by accumulated experience. Similarly, health conditions and associated concerns fluctuate in intensity. These data also do not capture geography or position within the service network, though neighborhood age and household construction are surely important considerations. Finally, these data do not include any administrative records on billing or usage. These are among the issues we plan to address in follow-up research.
The widespread use of bottled water produces significant negative environmental externalities. Although convenience and portability are surely important reasons for the rise in popularity of bottled water usage outside of the home, this paper set out to investigate the factors that are associated with the mode of water usage in the home. Using data generated by a telephone survey of municipal water customers in Jacksonville, FL, a model of residential water usage designed to accommodate nonexclusive alternatives was estimated. Consistent with prior research (Viscusi et al. 2015), the results suggest that residential bottled water usage is strongly associated with concerns and perceptions related to the quality of tap water.
Some analysts have proposed that the use of in-home water filters and purifiers might offer a cost effective and environmentally-friendly alternative to bottled water (Kumar Kumar Kaushal 2014; Sobsey et al. 2008). However, we argue that consumers' demand for filtered tap water and their demand for bottled water are driven by different considerations. Specifically, the evidence presented here suggests that the use of water filters is associated with aesthetic complaints, and that filtered tap water is not, in general, viewed as an adequate solution to more serious concerns related to perceptions of poor tap water quality. This finding has important implications for public policy. While we do find that water filter usage is sensitive to income constraints, municipal programs designed to promote, subsidize, or supply in-home water filters are unlikely to address concerns related to safety or perceived quality, and hence are unlikely to have much influence on behavior in these cases.
In addition, this paper presents evidence of differences in water usage across racial groups. In particular, African-American respondents are more likely to use bottled water in the home than are other racial groups. Insofar as this is linked to concerns over perceived water quality, this result is consistent with a broad literature on the legacy of residential segregation and consequent variation in the quality of public services. In particular, this literature has identified predominantly African-Americans and Hispanic neighborhoods as being especially likely to experience poor tap water quality (Gorelick et al. 2011). While the present study is necessarily limited in its ability to address this topic directly, this is clearly a fertile area for additional research and policy analysis.
Acknowledgements We acknowledge support from the Environmental Center, University of North Florida (Seed Grant, 2015). We arc thankful to Dr. Michael Binder (Faculty Director, Public Opinion Research Laboratory) and Andrew Hopkins (Coordinator, Public Opinion Research Laboratory) for their support in pursuing the survey, and to Art Sams for providing excellent research assistance. Lastly, we are very much thankful to Dr. Mary Beal for her involvement and all the students who volunteered their time for the survey. Dr. David Lambert at the University of North Florida (Director. Environmental Center) provided very useful comments. However, the opinions expressed here are solely those of the authors.
Abdalla, C. W., Roach, B. A, & Epp, D. J. (1992). Valuing environmental quality changes using averting expenditures : An application to groundwater contamination. Land Economics, 163(2), 163-169, May 1992.
Abrahams, N. A.. Hubbell, B. J., & Jordan, J. L. (2000). Joint production and averting expenditure measures of willingness to pay: Do water expenditures really measure avoidance costs? American Journal of Agricultural Economics. 82(2), 427-137, May 2000.
Alexander, K. S., Price, J. C, Browne, A. L., Leviston. Z., Bishop, B. J., & Nancarrow, B. E. (2008). Community perceptions of risk , trust and fairness in relation to the indirect potable use of purified recycled water in south East Queensland : A scoping report Urban water security research Alliance, technical report no. 2. Queensland. Australia: CSIRO. http://citeseerx.ist.psu.cdu/viewdoc/download?doi=l0.1.1.424.508&rcp=repl&type=pdf. Accessed 2 February 2017.
American Association of Public Opinion Researchers. (2016). Response rates - an overview. https://www.aapor.org/Education-Resources/For-Researchers/Poll-Survey-FAQ/Responsc-Rates-An-Ovcrview.aspx. Accessed 29 March 2017.
Anadu, E. C, & Harding, A. K. (2000). Risk perception and bottled water use. Journal of the American Water Works Association, 92(11), 82-92, November 2000.
Baskaran, R., Managi, S., & Bendig, M. (2013). A public perspective on the adoption of microgeneration technologies in New Zealand: A multivariate probit approach. Energy Policy. 58, 177-188, July 2013.
Beverage Marketing Corporation. (2017). Press Release: Bottled Water Becomes Number-One Beverage in the U.S. (p. 1). https://www.bcvcragemarketing.com/news-detail.asp?id=438. Accessed on 2 June 2018.
Bontemps, C, & Nauges, C. (2016). The impact of perceptions in averting-decision models: An application of the special regressor method to drinking water choices. American Journal of Agricultural Economics, 89(1). 297-313 January 2016.
Brox, J. A., Kumar, R. C, & Stollery, K. R. (2003). Estimating willingness to pay for improved water quality in the presence of item nonresponse bias. American Journal of Agricultural Economics. 85(2), 414-428, May 2003.
Cappcllari. L., & Jenkins, S. P. (2003). Multivariate probit regression using simulated maximum likelihood. 77ie Stata Journal, 5(3), 278-294, September 2003.
Castillo-Manzano. J. I. (2010). Determinants of commercial revenues at airports: Lessons learned from Spanish regional airports. Tourism Management, 31(6), 788-796. December 2010.
Chattcrjec. C., Triplett. R., Johnson, C. K., & Ahmed, P. (2017). Willingness to pay for safe drinking water: A contingent valuation study in Jacksonville, FL. Journal of Environmental Management, 203. 413-421| December 2017.
Choo, S., & Mokhtarian, P. L. (2008). How do people respond to congestion mitigation policies? A multivariate probit model of the individual consideration of three travel-related strategy bundles. Transportation, 35(2), 145-163 March 2008.
Didicr. S. (2013). Water Bottle Pollution Facts. San Francisco Chronicle, http://homcguides.sfgate.com/water-bottlc-pollution-79179.html. Accessed 1 January 2017.
Duncan, S. (2010). 'We have drunken our water for money': The political economy of bottled water, 1940-1995. Case Western Rcscrvcr University - Electronic Thesis, http://rave.ohiolink.edu/etdc/view?acc_num=casel275601437. Accessed on 25 March 2017.
Environmental Working Group. (2012). National Drinking Water Database, http://www.ewg.org/tapwatcr/whatsinyourwater/UT/BountifulCity/AH06015/Trichloroethylenc/2984/atabasc. Accessed on 4 February 2017.
Font-Ribera, L., Cotta. J. C, Gomez-Gutierrez, A., & Villanueva, C. M. (2017). Trihalomcthanc concentrations in tap water as determinant of bottled water use in the city of Barcelona. Journal of Environmental Sciences, 58, 77-82 August 2017.
Glcick. P. H.. & Cooley, H. S. (2009). Energy implications of bottled water. Environmental Research Letters, 4(1), 1-6, February 2009.
Gorelick, M. H., Gould, L., Nimmer, M., Wagner, D., Heath, M., Bashir, H., & Brousseau, D. C. (2011). Perceptions about water and increased use of bottled water in minority children. Archives of Pediatrics and Adolescent Medicine. 165( 10), 928-932 October 2011.
He, S., Jordan, J., & Paudel, K. (2008). Economic evaluation of bottled water consumption as an averting means: Evidence from a hedonic price analysis. Applied Economics Letters. 15(5), 337-342, April 2008.
Hu, Z., Morton, L. W., & Mahler, R. L. (2011). Bottled water United States consumers and their perceptions of water quality. International Journal of Environmental Research and Public Health, 8(2), 565-578 February 2011.
Hunter. P. R. (1993). The microbiology of bottled natural mineral waters. Journal of Applied Bacteriology, 74(4), 345-352, April 1993.
Jacksonville Electric Authority. (2017). Frequent water questions, https://www.jea.com/about/water_supply/frequent watcrqucstions/. Accessed on 30 March 2017.
Jeuland. M., Orgill, J., Shaheed. A., Rcvell, G., & Brown, J. (2016). A matter of good taste: Investigating preferences for in-house water treatment in peri-urban communities in Cambodia. Environment and Development Economics, 2/(3). 291-317 June 2016.
Kohut. A.. Keeter, S., Doherty, C. Dimock. M.. & Christian. L. (2012). Assessing the representativeness of public opinion surveys. Pew Research Center. Washington. DC. http://www.people-press.org/files/lcgacy-pdf/assessing the representativeness of public opinion survcys.pdf. Accessed on 2 March 2017.
Koo, Y., Lim, S., Kim, K. & Cho. Y. (2014). Analysis of user characteristics regarding social network services in South Korea using the multivariate probit model. Technological Forecasting and Social Change, 88. 232-240 October 2014.
Kpadonou, R. A. B., Owiyo, T., Barbier, B., Denton, F., Rutabingwa, F, & Kicma, A. (2017). Advancing climate-smart-agriculture in developing drylands: Joint analysis of the adoption of multiple on-farm soil and water conservation technologies in west African Sahcl. Land Use Policy, 61, 196-207, February 2017.
Kumar Kaushal. R. (2014). Arsenic removal from water by physico-chemical approach - A low-cost treatment option. International Research Journal of Public and Environmental Health, 1(1), 10-18. March 2014.
Lalumandier. J. a., & Ayers, L. W. (2000). Fluoride and bacterial content of bottled water vs tap water. Archives of Family Medicine, 9(3), 246-250 June 2000.
Moffat B.. Motlalcng, G.. & Thukuza, A. (2011). Households willingness to pay for improved water quality and reliability of supply in Chobe ward, Maun. Botswana Journal of Economics, 8(12). 45-61 January 2011.
National Realty & Development Corporation. (2010). Bottled water - executive summary. Bottled Water: Pure Drink or Pure Hype? http://www.nrdc.org/water/drinking/bw/cxesum.asp. Accessed on 20 March 2017.
Royce, E. (2008). Bottlemania: How water went on Sale and why we bought it (pp. 112-174). New York: Bloomsbury.
Salch. M. A., Ewane, E., Jones, J., & Wilson, B. L. (2001). Chemical evaluation of commercial bottled drinking water from Egypt. Journal of Food Composition and Analysis, 14(2), 127-152, April 2001.
Simons, D. J. (2000). Attcntional capture and inattentional blindness. Trends in Cognitive Sciences, 4(4), 147 155. April 2000.
Sobsey, M. D., Stauber, C. E., Casanova, L. M., Brown, J. M., & Elliott, M. A. (2008). Point of use household drinking water filtration: A practical, effective solution for providing sustained access to safe drinking water in the developing world. Environmental Science and Technology, 42( 12). 4261-4267, March 2008.
Tanellari, E., Bosch. D.. Boyle, K.. & Mykcrczi, E. (2015). On consumers' attitudes and willingness to pay for improved drinking water quality and infrastructure. Water Resources Research. 5/(1), 47-57, January 2015.
U.S. Census Bureau. (2015). American community survey 5-ycar estimates. American Commnity Survey - Demographic and Housing, https://factfindcr.ccnsus.gov/faces/tableservices/jsf7pages/productview.xhtml?src=CF. Accessed on 15 March 2017.
U.S. Government Accountability Office. (2009). FDA safety and consumer protections are often less stringent than comparable EPA protections for tap water, https://www.gao.gov/assets/300/291465.pdf. Accessed 15 June 2018.
Um, M., Kwak, S., & Kim, T. (2002). Estimating willingness to pay for improved drinking water quality using averting behavior method with perception measure. Environmental and Resource Economics, 27(3), 285-300 March 2002.
Vasquez, W.F.. Mozumdcr. P., & Francheschi, D. (2015). Water quality, household perceptions and averting behavior: Evidence from Nicaragua. Water Economics and Policy, 1(4), 1-21, December 2015.
Viscusi. W. K, Huber, J.. & Bell, J. (2012). Alternative policies to increase recycling of plastic water bottles in the United States. Review of Environmental Economics and Policy, 6(2), 190-211 July2012.
Viscusi, W. K. Hubcr, J.. & Bell, J. (2015). The private rationality of bottled water drinking. Contemporary Economic Policy, 33(3), 450-467 July 2015.
Weber, C. & McCann. L. (2015). Adoption of nitrogen-efficient technologies by U.S. com farmers. Journal of Environmental Quality. 44(2), 391-401, March 2015.
Westerhoff. P., Prapaipong, P., Shock, E., & Hillaireau, A. (2008). Antimony leaching from polyethylene terephthalate (PET) plastic used for bottled drinking water. Water Research, 42(3), 551-556, February 2008.
Yoo, S. H. (2005). Analysing household bottled water and water purifier expenditures: Simultaneous equation bivariate Tobit model. Applied Economics Letters, 12(5), 297-301 April 2005.
Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
[??] Russell Triplett
Russell Triplett (1*) Chiradip Chatterjee (1*) Christopher K. Johnson (1*) Parvez Ahmed (2)
(1) Department of Economics and Geography, University of North Florida, 1 UNF Drive, Jacksonville. Florida 32224, USA
(2) Department of Accounting and Finance, University of North Florida, 1 UNF Drive, Jacksonville. Florida 32224, USA
Model specifications were explored with separate coefficients for each government agency, but the high correlation among the variables introduced multicollinearity and reduced degrees of freedom. The estimation results were not qualitatively different.
Table 1 Variables constructed from March 2016 water use survey Variable Description Unfiltered Tap Water Respondent drinks water from unfiltered taps in the home. (1 if Yes, 0 if No) Filter Tap Water Respondent uses a water treatment device (such as Brita water filters, reverse osmosis system, refrigerator front filters, water ionizer) in the home. (1 if Yes, 0 if No) Bottled Water Respondent drinks bottled water (large containers or small bottles, subscription, delivery and/or on demand) in the home. (1 if Yes, 0 if No) Belief Safe Your unfiltered tap water is safe to drink. (1 - Strongly disagree, 2 - Somewhat disagree. 3 - Neither agree nor disagree, 4 - Somewhat agree 5 - Strongly agree) Belief Contaminated Your unfiltered tap water is contaminated. (1 - Strongly disagree, 2 - Somewhat disagree, 3 - Neither agree nor disagree, 4 - Somewhat agree 5 - Strongly agree) Concern Sick Using a scale of I to 5, where 1 means you are not at all concerned and 5 means you are very concerned, how concerned arc you about you or another member of your family getting sick from drinking unfiltered tap water? Foul Smell Your unfiltered tap water has a foul smell. (1 - Strongly disagree, 2 - Somewhat disagree, 3 - Neither agree nor disagree, 4 - Somewhat agree 5 - Strongly agree) Distrust Average How much of the time do you think you can trust the State Government, City Government, Environmental Protection Agency, and JEA to do what is right? (1 - Just about always, 2 - Most of the time, 3 - Only some of the time, 4 - Never) Average score calculated by authors. Heard Have you heard about any potential problems that your town or city supplier has, or had, with the water supply? (1 if Yes, 0 if No) Family Size Total number of people living in the household. Fraction Children Fraction of people in the household that are age 18 or less. Calculated by authors. Home Owner Respondent lives in owner-occupied housing (1 if Yes, 0 if No) Age Age cohort of respondent (1-18-24 yrs., 2-25-44 yrs.. 3-45-64 yrs., 4-65-80 yrs., 5 - Over 80 yrs.) Education Respondent's highest grade in school or year of college completed. (1 - Less than high school, 2 - High school graduate, 3 - Some college, 4 - College graduate, 5 - Post graduate) Income Annual household income (1 - Less than $25,000, 2 - $25,000 to $50,000, 3 - $50,000 to $75,000, 4 - $75,000 to $100,000, 5 - $100,000 to $150,000, 6-Above $150,000. Racial Categories Caucasian African American Hispanic, Asian or Native American (HANA) Other Variable Obs. Mean S.D. Unfiltered Tap Water 529 0.24 0.43 Filter Tap Water 529 0.39 0.49 Bottled Water 529 0.57 0.50 Belief Safe 529 3.74 1.12 Belief Contaminated 529 2.72 1.06 Concern Sick 529 2.63 1.62 Foul Smell 529 2.71 1.03 Distrust Average 529 2.66 0.65 Heard 529 0.13 0.34 Family Size 529 2.82 1.47 Fraction Children 529 0.14 0.22 Home Owner 529 0.69 0.46 Age 529 2.73 0.96 Education 529 2.31 0.94 Income 420 2.96 1.45 Racial Categories 529 55% 34% 6% 5% Source: Authors' original survey data collected in March 2016 Table 2 Multivariate probit, water mode choice Equations Variables 1: Unfiltcred tap water 2: Bottled water Belief safe 0.288 (***) -0.146 (**) (0.070) (0.068) Belief contaminated -0.139 (*) 0.074 (0.720) (0.068) Concern sick -0.054 0.073 (*) (0.049) (0.044) Foul smell -0.119 (*) -0.137 (**) (0.071) (0.065) Distrust average -0.041 0.087 (0.105) (0.095) Heard 0.452 (**) -0.046 (0.207) (0.182) Education -0.074 -0.042 (0.066) (0.062) Age 0.006 -0.008 (0.073) (0.066) Family size -0.088 0.019 (0.059) (0.053) Fraction children 0.736 (*) 0.078 (0.374) (0.361) Home owner -0.304 (**) -0.193 (0.145) (0.133) African american -0.214 0.477 (***) (0.147) (0.134) HANA+ -0.041 -0.082 (0.274) (0.238) Other -0.339 0.007 (0.329) (0.266) Constant -0.366 0.553 (0.601) (0.565) Fit statistics Error Correlations Log psudolikelihood -851.94 Pl2 Wald x2 134.48 Pl3 p value 0.00 P23 N 529 Draws 100 Variables 3: Filtered tap water Belief safe 0.025 (0.065) Belief contaminated -0.109 (*) (0.064) Concern sick -0.047 (0.044) Foul smell 0.129 (**) (0.062) Distrust average -0.046 (0.093) Heard -0.065 (0.184) Education 0.189 (***) (0.064) Age 0.049 (0.067) Family size 0.048 (0.052) Fraction children -0.169 (0.344) Home owner 0.219 (*) (0.131) African american -0.368 (***) (0.132) HANA+ 0.211 (0.249) Other -0.145 (0.270) Constant -0.904 (*) (0.548) Fit statistics Log psudolikelihood -0.319 (***) (0.071) Wald x2 -0.393 (***) (0.066) p value -0.488 (***) (0.062) N Draws Source: Authors' original survey data collected in March 2016. Notes: Robust standard errors in parentheses. (*), (**), (***) signify staAsAcal significance at the 1%. 5% and 10% levels, respectively. (*) Hispanic. Asian or Native American Table 3 Multivariate probit, water mode choice (with income) Equations Variables 1: Unfiltcrcd tap water 2: Bottled water Belief safe 0.272 (***) -0.133 (*) (0.082) (0.078) Belief contaminated -0.145 (*) 0.037 (0.082) (0.081) Concern sick -0.104 (*) 0.083 (0.057) (0.053) Foul smell -0.078 -0.125 (*) (0.078) (0.075) Distrust average 0.125 0.037 (0.114) (0.104) Heard 0.501 (**) -0.028 (0.228) (0.202) Education -0.031 -0.011 (0.084) (0.079) Age 0.015 -0.029 (0.093) (0.078) Family size -0.051 0.037 (0.067) (0.062) Fraction children 0.495 0.029 (0.422) (0.403) Home owner -0.251 -0.123 (0.169) (0.157) African amcrican -0.182 0.466 (***) (0.164) (0.153) HANA (*) -0.056 -0.024 (0.295) (0.259) Other -0.251 -0.099 (0.419) (0.346) Income -0.154 (**) -0.031 (0.063) (0.055) Constant -0.483 0.591 (0.689) (0.631) Fit statistics Error correlations Log psudolikelihood -642.77 Pl2 Wald [chi square] 118.25 Pl3 p value 0.00 P23 N 400 Draws 100 Variables 3: Filtered tap water Belief safe 0.002 (0.075) Belief contaminated -0.124 (0.077) Concern sick -0.038 (0.053) Foul smell 0.129 (*) (0.712) Distrust average -0.046 (0.105) Heard -0.022 (0.214) Education 0.123 (0.081) Age 0.027 (0.082) Family size 0.074 (0.062) Fraction children -0.089 (0.382) Home owner 0.091 (0.158) African amcrican -0.317 (**) (0.152) HANA (*) 0.435 (0.279) Other -0.305 (0.341) Income 0.113 (**) (0.054) Constant -0.966 (0.635) Fit statistics Log psudolikelihood -0.337 (***) (0.081) Wald [chi square] -0.371 (***) (0.077) p value -0.502 (***) (0.069) N Draws Source: Authors' original survey data collected in March 2016. Notes: Robust standard errors in parentheses. (*), (**), (***) signify staAsAcal significance at the 1%, 5% and 10% levels, respectively. ([dagger]) Hispanic, Asian or Native American
|Printer friendly Cite/link Email Feedback|
|Author:||Triplett, Russell; Chatterjee, Chiradip; Johnson, Christopher K.; Ahmed, Parvez|
|Publication:||International Advances in Economic Research|
|Date:||May 1, 2019|
|Previous Article:||Two-Sided Information Asymmetry in the Healthcare Industry.|
|Next Article:||Impact of Increasing Minimum Wage on Homeownership and Home Equity Loans.|