Previous studies which have compared eudaimonic and hedonic happiness have incorporated demographic information as attributes within their exploration. Levinson’s 2011 study included General Social Survey (GSS) interview responses to key questions related to happiness coded to a location where the response was recorded. By creating ‘population weighted centroids’, the study combined air quality and interview data to analyse how air quality is valued as a ‘public good’ in the United States. This approach is advantageous as it included both categorical and ordered data and offered some explanations on the economic divisions that air quality may help to reproduce. However, survey data is subject to response bias—the way in which responses are provided (Lavakras, 2008, p. 753), which is especially important when comparing hedonic and eudaimonic happiness as the results may differ based on time.

 

Previous studies have also employed a regression analysis of the data to assess the statistical significance between the chosen areas and their relative happiness (Knight & Howley, 2017 & Houlden, et al., 2019). The regression approach in the former study allowed the researchers to minimise the ‘heterogeneity’ of personality traits and assess the ‘welfare effects of exposure to nitrogen dioxide’ (Knight & Howley, 2017, p. 7). Houlden’s investigation included a Geographically Weighted Regression (GWR), to ‘adjust for underlying spatial processes within the data and investigate the geographic variation in the association between local greenspace and mental wellbeing’ (Houlden, et al., 2019, p. 1). While the GWR method is advantageous for calculating localised regression (by using distance-based weighting for each point), the applications of the method are experimental (Houlden & al, 2019, p. 3).

 

Nevertheless, adopting a similar approach to focus on the localised relationships between the variables could be more useful to test the different trends across space and time in our investigation. Our regression approach will therefore attempt to assess the localised differences of the relationship between air pollution and happiness, but be cautious not to overgeneralise, as the scale of our assessment is much smaller than existing studies.