Who's greener? Comparing urban and suburban residents' environmental behaviour and concern

Abstract

While various studies have compared metropolitan and rural residents' displays of environmentalism, recent shifts in the housing and energy markets make intra-metropolitan differences in urban and suburban attitudes and actions more interesting. The housing market crash and accompanying rise in fuel costs are prompting a return of centralisation to many communities and, it is believed, are drawing environmentally conscious residents downtown closer to work and play. Past research has found that this trend – demonstrated by a re-centralising housing market – was already underway, pre-crash, in Louisville, Kentucky over the “boom” period from 2000 to 2006. Drawing on original 2006 survey data for Louisville, this study regresses two components of environmentalism (behaviour and concern) on socio-demographics and residence. Findings show urbanites in Louisville are indeed significantly more environmentally friendly in their behaviour but not in their concern. Countering some claims, residents of single-family homes also exhibit significantly higher levels of environmental behaviour and concern than those residing in apartments and condominiums. Finally, in a surprise finding, recent movers from the suburbs to the city do not differ in their behaviour and are significantly less concerned about the environment than urbanites, suburbanites, and all metropolitan residents.

Keywords:

• urban
• suburban
• environmental attitudes
• pro-environmental behaviour
• new ecological paradigm

Notes

1 All references to metropolitan Louisville in this paper refer to Louisville Metro, consolidated Louisville-Jefferson County, and not the US Census Bureau-defined, multi-county Louisville Metropolitan Statistical Area.

2 The previous research studies examining political and environmental attitudes cited in this paper exhibit quite low rates of explained variation.

3 I-264 (Watterson Expressway) and I-265 (Gene Snyder Freeway) are the inner and outer beltways, respectively.

4 The adjusted response rate for this 2006 survey was rather low (just shy of one-fifth). This problem is plaguing most contemporary surveys due to lower reliance on landline telephones and lower levels of willingness to complete phone surveys. Previous published research on environmentalism had similarly low response rates, particularly when including urban populations (Morrissey and Manning 2000). It is not possible to calculate the response rate for the cohort asked environmental questions alone from publically available materials. The quality of the data set and the environmental measures has been confirmed by the publication of other peer-reviewed research on the LMS' pro-environmental behaviour measures (Gilderbloom et al. 2009, Walton and Austin 2011). It is possible that non-response bias is affecting the findings, in that some populations could have been less likely to complete the survey than others (see note 11), but it is more important that (1) the sample was collected randomly and (2) the sample is representative of the community and approximates the population in most categories. Weighting the sample with 2006 American Community Survey data has no major effects on the findings and so no weighting was done for the results presented in this study.

5 While many central cities consist of more apartments and condominiums and less detached SFHs, Louisville is an exception. Downtown neighbourhoods are mostly composed of houses and many suburban communities feature prominent apartment complexes and an abundance of condominiums. Thus, type of residence is not a proxy for urban versus suburban. This measure is not without its weaknesses: it does not distinguish tenure, between owners and renters, but only indicates type of residence. Nonetheless, it is the only measure available for housing in the LMS.

6 Eighth grade or less (1); some high school (2); high school graduate (3); some college or technical school (4); Associate's degree (5); Bachelor's degree (6); some graduate courses (7); or advanced degree (8).

7 All environmental variables were recoded prior to the PCA so that responses are now in the same direction: higher values denote more pro-environmental responses. It should also be noted that item 10 (“The so-called ‘ecological crisis’ facing humankind has been greatly exaggerated”) from the NEP scale is an “even numbered” item, while the other five are “odd numbered” – meaning that this one item from Dunlap et al. (2000, p. 433) is the only one included in the LMS that required disagreement in the NEP scale to express higher environmental concern.

8 “It is just too difficult for someone like me to do much about the environment.”

9 Kim and Mueller (1978, pp. 70–72) state:, “ … one may ignore specific variations in the factor loadings and consider only one type of information as relevant; either a variable loads on a given factor or it does not. Consequently, a scale is built by summing all the variables with substantial loadings and ignoring the remaining variables with minor loadings. The scale created in this way is no longer a factor scale but merely factor-based.” Normalised component z-scores generated by the PCA and component-based indices summing items multiplied by their component loadings were also calculated and analysed but are not included in the models presented in this paper. For the most part, findings differ minimally when using these alternative versions of the dependent variables and these differences do not alter the conclusions of the study.

10 The urban core is more liberal than the surrounding suburban rings (mean of 2.14 versus 1.85, respectively, with a difference significant at the 0.001 level).

11 According to the 2000 Census, the median age was 37 and males composed 48% of the population. In 2005, the Census Bureau reported that the median household income was approximately $44,000. The sample also includes somewhat more SFH-dwellers than the county's rate of 70% in 2000. Although the Census Bureau defines Jefferson County as 98% urban, this analysis uses the definition derived from Louisville Metro's (2006, p. 12) Comprehensive housing strategy, which defines neighbourhoods inside the inner beltway (I-264, Watterson Expressway) as “urban”. According to this measure, 31% of the county's population resided in the urban core in 2000. This compares favourably with the LMS's zip-code based estimation of 34%. No comparable measure of political ideology exists, but the sample mean seems to reflect local political dynamics.

12 While acknowledging that the 0.05 level of significance is desirable, this analysis flags 0.10 level incidences in the tables and indicates them in the text because the 5% level is an arbitrary cutoff. It is “conventional” for econometric studies to highlight the 10% level in addition to higher levels of statistical significance (Chaudhuri 2010, p. 213). This strategy is one of full disclosure – it shows which variables “near significance” at the 0.05 level but demonstrate a lower degree of confidence.

13 Gas prices in Louisville, patterning after national trends over the past decade, doubled between mid-2006 and mid-2008 to over $4.00 per gallon. However, gas prices had already doubled from less than $1.50 to over $3.00 between 2003 and 2006 – the three-year period immediately before the LMS survey data were collected. See Gasbuddy (2014) for historical gas price data in the nation and specific metropolitan areas, including Louisville.

14 This analysis statistically tests whether the sub-sample of movers differ in concern or behaviour compared with others in the sample, not whether concern predicts the act of moving (which would require a binary logistic regression predicting a 1/0 dummy dependent variable for the movers, which is plagued by a small n in this data set).

15 The LMS questions isolate consumer behaviours exclusively motivated by “environmental reasons” rather than the behaviours themselves (see Table 1 for question wordings). It is unknown how individual respondents chose to interpret this phrase.

16 Apartment and condominium dwellers score less on both environmental behaviour and concern compared with SFH dwellers, although the difference between the groups is not significant (Table 2).