The significance of population wellbeing is gaining widespread recognition globally, prompting governments to broaden their evaluative criteria beyond the traditional measure of GDP (Gross Domestic Product) to assess the overall success of their population [1]. While GDP and productivity measures continue to be central for policymaking, there is an emerging shift towards a more comprehensive approach that includes the assessment of wellbeing. Initiatives like the Wellbeing Economy Governments partnership (WEGo) exemplify this shift, where national and regional governments collaboratively advance the concept of Wellbeing Economies [2]. Despite sustained economic growth, New Zealand faces pressing challenges such as high rates of child poverty, homelessness, and suicide. In response, the government introduced its inaugural ‘wellbeing budget’ in 2019 [3], signifying a renewed commitment to prioritising people’s wellbeing alongside economic growth.
Understanding wellbeing presents challenges due to the evolving nature and diverse perspectives around its meaning. Initially, wellbeing was often perceived as positive human functioning, referred to as “eudaimonia,” encompassing aspects such as self-actualisation and autonomy [4]. Other researchers have integrated eudaemonic and hedonic components, combining aspects of functioning and emotions [5]. For example, Diener’s tripartite model identified cognitive, positive affect, and negative affect components [6], while Seligman’s PERMA model introduced positive emotion, engagement, relationships, meaning, and accomplishment as key dimensions [7]. Thompson et al.’s dynamic model of ‘flourishing’ further highlights the interplay between positive feelings, effective functioning, external conditions, and personal resources [8]. This comprehensive perspective suggests that ‘flourishing’ or elevated wellbeing emerges from the interplay of positive emotions and effective functioning within an individual’s unique circumstances and available resources. Thus, a ‘flourishing’ nation indicates elevated wellbeing among its citizens.
The increasing significance of incorporating wellbeing indicators into policy decisions is becoming more prominent in New Zealand. Despite this growing importance, there still exists a considerable gap in our understanding of the factors that influence population wellbeing in the country. This knowledge gap is partially attributed to the scarcity of detailed, population-level wellbeing data. The NZ General Social Survey (GSS), a biennial survey of around 9,000 individuals [9], offers wellbeing data across twelve domains: health, housing, income and consumption, jobs and earnings, leisure and free time, knowledge and skills, safety and security, social connections, cultural identity, civic engagement and governance, environmental quality, and subjective wellbeing. Designed based on the NZ Living Standards Framework [10], which in turn was drawn from the OECD’s framework [11], the GSS lays the foundation for wellbeing assessment in New Zealand. In the context of this study, we focus primarily on the subjective wellbeing domain, focussing on indicators such as life satisfaction, sense of purpose, family wellbeing and mental wellbeing.
Although the GSS sample is considered nationally representative, certain subgroups of the population (that may be of significant policy interest) remain underrepresented due to limitations in sample size. For instance, it is impractical to understand the wellbeing experiences of individuals living in government-sponsored social housing. This is because the number of people who participated in the GSS and are also residents of social housing may be very small. Therefore, to assess the impact of government initiatives targeting this specific population sub-group, comprehensive wellbeing measures applicable to the entire population are needed.
To address this challenge, two strategies offer potential solutions. One approach involves collecting regular wellbeing data for the entire population in a census activity; however, this method is resource-intensive and time-consuming. An alternative approach involves leveraging existing routinely collected data to extrapolate GSS wellbeing measures to the broader population. This may be feasible due to New Zealand's Integrated Data Infrastructure (IDI): a complex database managed by Stats NZ [12]. The IDI contains individual response data (microdata) on people and households, supplemented with anonymised information on education, income, health, justice, and housing. Notably, the IDI facilitates dataset linkage across these areas using a unique identifier variable. Details about this linking process are available elsewhere [13]. Crucially, the IDI houses the GSS data, allowing linkage with the country’s Census data which the majority of the nation's population completes (given it is a legal requirement to do so).
The Census is a comprehensive nationwide survey conducted once every five years in New Zealand, with the primary aim of officially counting individuals and households in the country [14]. It also provides a snapshot of various aspects of life, including demographic information, educational qualifications, employment status, and more. Additionally, the Census gathers data on addresses for each household, which are then aggregated at the meshblock level for reporting purposes. A meshblock represents the smallest administrative geographical unit, typically encompassing about 30 to 60 households [15]. Environmental data, such as the extent of green spaces, are also available at the meshblock level and can therefore be linked to the Census data. One notable example is the Healthy Location Index, which captures accessibility to health-promoting elements (e.g., green spaces, physical activity facilities) and health-constraining elements (e.g., alcohol outlets, fast-food shops) [16]. The ability to link such key environmental information to the Census is crucial, given the established links between the environment and wellbeing [17, 18].
The aim of this study is to predict GSS-derived wellbeing measures from Census-based sociodemographic information and meshblock-level environmental indicators. If successful, such a model could be used to extrapolate these predicted wellbeing scores to the entire IDI population, thereby creating a population-level estimate of subjective wellbeing. This could yield transformative benefits by facilitating the integration of wellbeing metrics into policy analysis. It also holds the potential to significantly enhance our understanding of how the political, social, and economic landscape impacts the wellbeing and overall functioning of individuals in New Zealand. This would further empower decision-makers to formulate more informed, targeted, and effective policies that address the genuine needs and concerns of New Zealanders.