Wellbeing and Quality of Life in Rural Europe

| 0 OCTOBER 2025 WELLBEING AND QUALITY OF LIFE IN RURAL EUROPE D4.4 Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Executive Agency. Neither the European Union nor the granting authority can be held responsible for them. UK participants in the GRANULAR project are supported by UKRIGrant numbers 10039965 (James Hutton Institute) and 10041831 (University of Southampton). | 1 D4.4 WELLBEING AND QUALITY OF LIFE IN RURAL EUROPE Project name GRANULAR: Giving Rural Actors Novel data and re-Useable tools to Lead public Action in Rural areas Project No Horizon Europe Grant Number (101061068); UKRI Grant Numbers James Hutton Institute (10039965) and University of Southampton (10041831) Type of funding scheme Horizon Europe Research and Innovation Action (RIA)- UK Research & Innovation Grant Call ID & topic HORIZON-CL6-2021-COMMUNITIES-01-01 Website www.ruralgranular.eu Document type Deliverable Status Final version Dissemination level Public Authors Gundi Knies (Thünen-Institute of Rural Studies), Jonathan Hopkins (James Hutton Institute) Work Package Leader Nordregio (NOR) Project coordinator Mediterranean Agronomic Institute of Montpellier (IAMM) Citation: Knies, G., & Hopkins, J. (2025). Wellbeing and Quality of Life in Rural Europe. GRANULAR. https://doi.org/10.5281/zenodo.17864862 This license allows users to distribute, remix, adapt, and build upon the material in any medium or format for noncommercial purposes only, and only so long as attribution is given to the creator. Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Executive Agency. Neither the European Union nor the granting authority can be held responsible for them. UK participants in the GRANULAR project are supported by UKRIGrant numbers 10039965 (James Hutton Institute) and 10041831 (University of Southampton). Funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Executive Agency. Neither the European Union nor the granting | 2 Table of Contents Table of Contents ................................................................................................................................................ 2 Table of Figures .................................................................................................................................................. 4 Table of Tables .................................................................................................................................................... 5 Table of Appendices ........................................................................................................................................... 6 Executive summary ............................................................................................................................................ 7 1. Introduction ..................................................................................................................................................... 8 2. Subjective Wellbeing in Rural and Urban Contexts: Is there a rural wellbeing advantage in Europe? 11 2.1. Introduction ............................................................................................................................................. 11 2.1. Searching and reviewing the empirical literature – how it was done ...................................................... 12 2.2. Results .................................................................................................................................................... 15 2.2.1. Overview of the empirical evidence database ..................................................................... 15 2.2.2. Is there a rural advantage in subjective wellbeing? ............................................................. 18 2.2.3. What factors shape the rural advantage in subjective wellbeing? ....................................... 18 2.3. Discussion .............................................................................................................................................. 28 2.4. Conclusion and outlook .......................................................................................................................... 29 3. Rural wellbeing: Subjective wellbeing across European Countries and Settlement Types .................. 30 3.1. Introduction ............................................................................................................................................. 30 3.1.1. Theories and debates on rural-urban differences in wellbeing ............................................ 30 3.1.2. Wellbeing as subjective experience .................................................................................... 32 3.1.3. Empirical research on correlates of SWB ............................................................................ 33 3.2. Data and methods .................................................................................................................................. 35 3.3. Results .................................................................................................................................................... 39 3.3.1. Settlement structure and subjective wellbeing across Europe ............................................ 40 3.3.2. Bivariate associations between settlement status and subjective wellbeing ....................... 41 3.3.3. Multivariate associations between settlement status and subjective wellbeing ............ 44 3.3.4. Country-specific variation in associations ............................................................................ 47 3.4. Discussion .............................................................................................................................................. 56 4. Is rural = rural? An empirical analysis of how findings on the rural life satisfaction advantage changes when we swap the definition of rural .............................................................................................................. 58 4.1. Introduction ............................................................................................................................................. 58 4.1.1. Hypotheses .......................................................................................................................... 59 4.2. Data and methods .................................................................................................................................. 60 | 3 4.2.1. Empirical strategy ................................................................................................................ 64 4.3. Results .................................................................................................................................................... 64 4.3.1. Evolution of life satisfaction over time by rurality ................................................................. 64 4.3.2. Factors explaining the rural wellbeing advantage ............................................................... 66 4.4. Discussion .............................................................................................................................................. 73 5. Leveraging POI Data on Public Infrastructure: Wind turbines and Quality of Life in Germany ............ 75 5.1. Introduction ............................................................................................................................................. 75 5.2. Data and methods .................................................................................................................................. 76 5.2.1. German Socio-economic Panel Study (SOEP) ................................................................... 76 5.2.2. Core Energy Market Registry (MaStR) ................................................................................ 76 5.2.3. Treatment, treatment radius and ban radius specification ................................................... 77 5.2.4. Allocation to quasi-experimental groups .............................................................................. 78 5.2.5. Empirical model ................................................................................................................... 79 5.3. Results .................................................................................................................................................... 80 5.4. Discussion .............................................................................................................................................. 86 5.5. Conclusions ............................................................................................................................................ 87 6. Overall conclusions ...................................................................................................................................... 88 7. Acknowledgements ...................................................................................................................................... 89 8. References ..................................................................................................................................................... 90 9. Appendix ........................................................................................................................................................ 98 | 4 Table of Figures Figure 2-1: Literature search criteria and outcomes ................................................................................................ 14 Figure 2-2: Proportion rural advantage in subjective wellbeing, by time period studied .......................................... 19 Figure 2-3: Proportion rural advantage in subjective wellbeing, by age of the population studied .......................... 20 Figure 2-4: Proportion rural advantage in subjective wellbeing, by (groups of) countries studied ........................... 20 Figure 2-5: Proportion rural advantage in subjective wellbeing, by wellbeing outcome studied .............................. 21 Figure 2-6: Proportion rural advantage in subjective wellbeing, by wellbeing instruments studied ......................... 22 Figure 2-7: Proportion rural advantage in subjective wellbeing, by geographic scale of effects studied ................. 24 Figure 2-8: Proportion rural advantage in subjective wellbeing, by area classification utilised ................................ 24 Figure 2-9: Proportion rural advantage in subjective wellbeing, by rural base category studied ............................. 27 Figure 2-10: Proportion rural advantage in subjective wellbeing, by urban (and less rural) comparison category .. 28 Figure 3-1: Conceptual framework for dimensions of subjective wellbeing. ............................................................ 32 Figure 3-2: Settlement structure across Europe and European welfare states (% of the population) ..................... 40 Figure 3-3: Spider plots of subjective wellbeing by settlement type and welfare state group .................................. 43 Figure 3-4: Multivariate predictions of subjective wellbeing on settlement type. Unadjusted and adjusted means. 45 Figure 3-5: Settlement types with the highest life satisfaction in the EU ................................................................. 48 Figure 3-6: Settlement types with the highest affective wellbeing in the EU ............................................................ 49 Figure 3-7: Settlement types with the highest basic psychological needs satisfaction in the EU ............................ 50 Figure 3-8: Settlement types with the highest mental resources in the EU ............................................................. 51 Figure 3-9: Settlement types with the highest social wellbeing in the EU ................................................................ 52 Figure 3-10: Settlement types with the highest evaluative wellbeing in the EU ....................................................... 53 Figure 4-1: Mean life satisfaction by rurality (with 95% CI), Great Britain 2009-2023 ............................................. 64 Figure 4-2: Mean life satisfaction by rural area type (DEGURBA2, with 95% CI), Great Britain 2009-2023 ........... 65 Figure 5-1: Illustration of allocation to treatment and control group based on placement of wind turbines in the treatment, ban and match radii ..................................................................................................................... 78 Figure 5-2: Expansion of wind turbines in Germany from 1983 to end of 2020 ....................................................... 80 Figure 5-3: Share of households located within 6 km of wind turbines in Germany 2002-2020, by type of rural area ...................................................................................................................................................................... 81 | 5 Table of Tables Table 2-1: Search Term Selection (Scopus format) ................................................................................................. 13 Table 2-2: Literature search string development ..................................................................................................... 13 Table 2-3: Study dimensions extracted from selected research papers .................................................................. 15 Table 2-4: Number of eligible studies, rural area type and rural context effects ...................................................... 15 Table 2-5: Focus of the research providing rural area type effects .......................................................................... 16 Table 2-6: Time period of effects ............................................................................................................................. 16 Table 2-7: Characteristics of the population for which rural area type effects have been reported ......................... 17 Table 2-8: Geographic dispersion of rural area type effects. ................................................................................... 17 Table 2-9: Rural advantage in subjective wellbeing ................................................................................................. 18 Table 2-10: Rural base categories by geographic scale of the administrative area classification ........................... 26 Table 3-1: Mapping of the ESS Round 6 subjective wellbeing items onto the Round 12 subjective wellbeing framework ..................................................................................................................................................... 37 Table 3-2: Overview of countries and welfare state models .................................................................................... 38 Table 3-3: Descriptive statistics of key variables used in the analysis (EU overall) ................................................. 39 Table 3-4: Subjective wellbeing across Europe and European welfare states (Average Mazziotta-Pareto Indices, MPI) ............................................................................................................................................................... 41 Table 3-5: Average subjective wellbeing by settlement structure across Europe (mean MPI) ................................ 42 Table 3-6: Settlement structures with the highest average score in six dimensions of subjective wellbeing ........... 47 Table 4-1: Overview of data sources used in the analysis ....................................................................................... 60 Table 4-2: Overview of variables used in this analysis ............................................................................................ 64 Table 4-3: Random-effects generalised least squares (GLS) regression of life satisfaction on individual blocks of context variables: Rural-urban coefficient ..................................................................................................... 66 Table 4-4: Random-effects generalised least squares (GLS) regression of life satisfaction on rurality, area and individual characteristics by rural-urban indicator (geodemographic classification = OAC 2011), GB 20092023 .............................................................................................................................................................. 67 Table 4-5: Random-effects generalised least squares (GLS) regression of life satisfaction on rurality, area and individual characteristics by rural-urban indicator (geodemographic classification = ACORN 2015), GB 20092023 .............................................................................................................................................................. 68 Table 4-6: Random-effects generalised least squares (GLS) regression of life satisfaction on area and individual characteristics by rural-urban indicator (geodemographic classification = OAC 2011), GB 2009-2023 ....... 70 Table 4-7: Random-effects generalised least squares (GLS) regression of life satisfaction on area and individual characteristics by rural-urban indicator (geodemographic classification = ACORN 2015), GB 2009-2023 .. 71 Table 4-8: Summary of hypotheses and results ...................................................................................................... 72 Table 5-1: Wind turbines in Germany, hub height and capacity .............................................................................. 77 Table 5-2: Fixed effects regressions of health-related quality of life (SF-12) on the presence of wind turbines in a radius of 1 km to 6 km of residential homes, 2002-2022 .............................................................................. 82 Table 5-3: Fixed effects difference-in-differences panel regressions of health-related quality of life (SF-12) on the presence of wind turbines in a radius of 1.5 km to 6 km of residential homes (quasi-experimental sample) 83 Table 5-4: Fixed effects difference-in-differences panel regressions of health-related quality of life (SF-12) on number of wind turbines (quasi-experimental sample) ............................................................................................... 84 Table 5-5: Fixed effects DiD panel regressions of health-related quality of life (SF-12) on inverse distance to wind turbines (quasi-experimental sample) ........................................................................................................... 85 | 6 Table of Appendices Appendix 1 (A2-1): List of publications included in the systematic review of rural area effects on subjective wellbeing and rural subjective wellbeing advantage in Europe in the empirical literature 2000-2023 .......................... 98 Appendix 2 (A2-2): Rural area effects on subjective wellbeing and rural subjective wellbeing advantage in Europe in the empirical literature 2000-2023 ................................................................................................................. 98 Appendix 3 (A2-3): Rural area effects on subjective wellbeing and rural subjective wellbeing advantage in Europe in the empirical literature 2000-2023: Rural-urban classifications by geographical scale ................................. 98 Appendix 4 (A3-1): Descriptive statistics of analytical sample - whole sample followed by welfare state groupand country-specific statistics .............................................................................................................................. 98 Appendix 5 (A3-2): Population estimates - all countries followed by welfare state groupand country-specific statistics ........................................................................................................................................................ 98 Appendix 6 (A3-3): Average subjective wellbeing scores (MPI) .............................................................................. 98 Appendix 7 (A3-4): Average subjective wellbeing scores (MPI) by settlement type ................................................ 98 Appendix 8 (A3-5_*): Multivariate regressions on life satisfaction (LS), Affective wellbeing (AW), Basic Psychological Needs Satisfaction (NS), Mental Resources (MR), Social wellbeing (SW), and Evaluative wellbeing (EW). 98 Appendix 9 (A3-6_*): Country-specific multivariate regressions on subjective wellbeing (UK, NL, PL, ES, FR, D1, D2, PT) .......................................................................................................................................................... 98 Appendix 10 (A4-1): Descriptive statistics of all variables used in the empirical analysis ....................................... 98 Appendix 11 (A4-2): Random Effects Generalised Least Squares Regressions of life satisfaction on rurality, area and individual characteristics, GB 2009-2023 ............................................................................................... 98 Appendix 12 (A4-3): Random Effects Generalised Least Squares Regressions of life satisfaction on area and individual characteristics by rurality, GB 2009-2023 ..................................................................................... 98 Appendix 13 (A5-1): Description of analytical samples used in wind turbine analysis ............................................. 98 Appendix 14 (A5-2): Propensity score matching statistics for all treatment groups ................................................. 98 Appendix 15 (A5-3): Fixed effects regressions of SF12 (MCS and PCS) on the presence of wind turbines, the number of wind turbines, and the inverse distance to wind turbines in a radius of 1 km to 6 km of residential homes. Germany 2002-2022 ..................................................................................................................................... 98 Appendix 16 (A5-4): Fixed effects difference-in-differences regressions of SF-12 (MCS and PCS) on the presence of wind turbines, the number of wind turbines, and the inverse distance to wind turbines within a radius of 1 km to 6 km of residential homes. Germany 2002-2022 ................................................................................ 98 | 7 Executive summary A society in which more people experience a higher quality of life—essentially, a greater wellbeing—is an important policy goal. This ambition is increasingly recognised across Europe, as reflected, for example, in the European Commission’s Long-Term Vision for Rural Areas (European Commission 2021), which places wellbeing, fairness and quality of life at the centre of rural development through to 2040. Yet the drivers of wellbeing remain complex, going well beyond traditional measures such as GDP. This report contributes to the evidence base by exploring how subjective wellbeing (SWB) varies across rural and urban contexts in Europe. It builds on new opportunities afforded by large-scale, high-quality survey data and their geographical linkages. Using three major datasets—the European Social Survey (ESS), the UK Household Longitudinal Study (UKHLS/Understanding Society), and the German Socio-Economic Panel (SOEP)—we examine how settlement type, socio-ecological context, and environmental exposures shape different dimensions of wellbeing. The analysis proceeds in four main steps: ▪ Systematic review (Chapter 2): Across the literature, a consistent but nuanced rural subjective wellbeing advantage emerges, particularly for evaluative life satisfaction measures. However, results vary strongly depending on how “rural” and “urban” are defined. Respondent-based definitions (e.g., “farm/countryside”) yield clearer rural advantages than administrative or density-based definitions. ▪ Wellbeing dimensions (Chapter 3): Rural life is associated with higher social wellbeing and mental resources (e.g., resilience, optimism), while advantages in life satisfaction and affective wellbeing often diminish once individual and country characteristics are controlled for. Differences also vary by welfare state grouping: Nordic, Central European, Mediterranean and Eastern European settlement types each show distinct wellbeing profiles. ▪ Neighbourhood and context effects (Chapter 4): Rurality remains positively linked to life satisfaction, but the effect weakens once neighbourhood characteristics are included. Area deprivation has a consistently negative impact in all settings. Contextual influences, such as access to health services and environmental quality, explain some—but not all—of the observed rural advantage in life satisfaction. ▪ Environmental exposures (Chapter 5): A case study of onshore wind turbines in Germany illustrates the complexity of local environmental impacts. Fixed-effects models suggest that proximity has a negative effect on physical health-related quality of life (as measured by the SF-12). At the same time, quasi-experimental designs reveal that larger turbines can be associated with improved physical wellbeing—likely reflecting community benefits. However, turbine density undermines mental wellbeing, highlighting the importance of careful siting and engagement strategies. Policy implications: ▪ Acknowledge the heterogeneity of rural experiences by recognising that rural living influences different dimensions of wellbeing in diverse ways across contexts and populations. ▪ Move beyond simplistic rural–urban binaries by adopting more nuanced, fine-grained definitions of place. ▪ Mitigate contextual risks, including socioeconomic deprivation, limited-service provision, and cumulative environmental pressures. ▪ Embed wellbeing considerations into rural policy frameworks, ensuring that initiatives—such as renewable energy deployment and other forms of development—are responsive to residents’ lived experiences and perceptions. Taken together, the analysis demonstrates that subjective wellbeing offers a valuable perspective for understanding rural–urban differences. By integrating conceptual, empirical, and methodological insights, the report identifies both opportunities and challenges in advancing wellbeing in rural Europe, thereby providing evidence in support of the European Commission’s aim to foster thriving, resilient, and inclusive rural areas by 2040. | 8 1. Introduction A society where more people experience a better and more positive quality of life—essentially, a greater wellbeing—is a worthy goal. Policymakers around the world are increasingly focused on creating conditions that support this aspiration, which also aligns with the European Commission’s Long-Term Vision for Rural Areas (European Commission 2021). The Vision explicitly adopts wellbeing, fairness and quality of life as central goals for rural Europe through to 2040. The traditional emphasis on economic growth as the primary means to enhance wellbeing is being questioned. There is a growing need to understand what contributes to wellbeing beyond just measuring GDP. Data on wellbeing is crucial in this context. Only by systematically gathering detailed information about people's experiences in various aspects of wellbeing, alongside broader insights into different areas of their lives, can we develop a comprehensive understanding of what drives wellbeing. This understanding is essential for creating effective policies aimed at maximising overall wellbeing. The EU Vision similarly emphasises strengthening data collection and monitoring for rural areas (including a rural observatory and rural proofing of policies) to better capture quality of life beyond economic indicators. Over the past two decades, researchers across economics, psychology, sociology, and geography have honed in on subjective wellbeing (SWB) — broadly comprising both evaluative dimensions (e.g., life satisfaction) and affective dimensions (e.g., momentary happiness, psychological functioning) — to understand how individuals perceive their quality of life and how social, economic, and environmental contexts shape these perceptions. A fascinating finding in the literature examining the relationship between place and subjective wellbeing (SWB) is the somewhat counterintuitive finding that rural residents frequently report levels of SWB that match or exceed those of urban populations, even though rural areas tend to score lower on many objective indicators of wellbeing, such as income, service provision, and employment opportunities. This “urban happiness paradox” challenges conventional wisdom about the benefits of urban agglomeration and highlights the need to understand the complex interplay between place, personal circumstances, and wellbeing. This report seeks to advance our understanding of rural–urban differences in SWB by addressing two interrelated sets of puzzles. First, why do some studies document a clear “rural wellbeing advantage” while others find no such effect? We argue that heterogeneity in measurement — both of SWB itself and of what constitutes “rurality” — underlies many of these divergent findings. Second, beyond simple comparative statements, what specific socioecological characteristics of rural contexts contribute to the wellbeing of their inhabitants? To answer these questions, we employ a multi-chapter approach that combines systematic literature review, theoretical extension to multiple dimensions of SWB, and novel empirical analyses leveraging fine-grained geographic linkages. Europe has seen a significant investment in new data on subjective wellbeing since the early 2000s – for example, the European Quality of Life Survey (EQLS) and the focused module on “Personal and Social Wellbeing” in the third and sixth rounds of the European Social Survey (ESS). In parallel, many data custodians also expanded the range of geodemographics and other socio-ecological context data, as well as facilitating an increased array of geographical data linkages. For example, the German Socio-economic Panel (SOEP) began making the postcode of the survey participants' addresses available for analysis in a secure setting in 2004 (although only going back to 2000, rather than to the start of the longitudinal household panel study, which started in 1986). Since 2010, it has been possible to draw on the exact geolocation of SOEP respondents’ addresses, again in a secure setting (Goebel, Wurm and Wagner 2010). The British sister study of SOEP, the British Household Panel Study (BHPS), also began to make fine-grained data from the UK Census 1991 available for analysis in secure settings around the turn of the Millennium. In 2012, the ESS released, for the first time, a linked geodata file with its cross-national survey data. By What matters for well-being? | 15 Table 2-3: Study dimensions extracted from selected research papers Study dimension Detail recorded Study information Author(s), year, title, abstract, journal, data sources Wellbeing outcome Detailed description, and categorised into four broad categories: Life Satisfaction (LS), Happiness (HP), Subjective Quality of Life (sQoL), and Other Subjective Wellbeing (Other SWB). Evaluative or affective wellbeing dimension, and the name of the survey instrument. Rural indicator Detailed information about the information used to capture the rural effect, including the type of spatial data used (if any). Rural effect Details about the size and direction of the rural effect, comparing more rural with less rural area type (i.e., negative numbers / Odd ratios below 1 indicating urban disadvantage and vice versa), and information about the population or subpopulation for which the effect was measured. 2.2. Results 2.2.1. Overview of the empirical evidence database To set the scene, we begin with a brief description of the literature review database. Each row of the database represents a unique rural–urban comparison in a specific SWB outcome (hereafter “effects”). Studies appear multiple times if they report results for different regions, age groups, periods, SWB measures, or model specifications. Table 2-4: Number of eligible studies, rural area type and rural context effects # of studies # of effects Eligible studies 97 652 No relevant effects 12 0 Final number of studies included 85 630 Studies with rural context effect(s) 8 27 Studies with area type effect(s) 80 603 Source: Own analysis of literature database. In total, the database contains 97 eligible studies with 652 observations (Table 2-4). Twelve studies did not report empirical results on rural–urban SWB differences (“rural area type effects”) or on contextual interactions (“rural context effects”). We were also interested in identifying the geospatial factors that contribute to SWB in rural settings. The literature search yielded only eight studies (27 effects) that examined how specific aspects of the rural context affect subjective wellbeing in rural populations. The following analysis, therefore, focuses on 80 studies that provide 603 “rural area type effects”. 2 As shown in Table 2-5, the database contains both studies explicitly focused on rural–urban wellbeing (N = 24, 217 effects) and studies in which area type is only one among many controls (N = 54, 382 effects). The range of topics of the “control-only” studies is broad, e.g., food consumption and lifestyle (Gschwandtner, Jewell and Kambhampati 2022), neighbourhood deprivation (Knies, Melo and Zhang 2020), the impact of tourism (Ivlevs 2017) and young people’s sexual attraction and drug abuse (Kuyper, de Roos et al. 2016). Only two studies are explicitly centred on 6 We direct the Reader to a recent systematic review of rural well-being studies by Veréb, Marques et al. (2024). Although it does not focus specifically on Europe or current issues, it offers valuable insights into the factors considered to affect rural well-being. However, these factors are not categorised by individual and local levels, and the literature database remains unpublished. | 16 rural wellbeing 3 (McGrath, Brennan et al. 2009, Jones, Malesios et al. 2020). Overall, the proportion of statistically significant effects among all effects is approximately 40% for the “control-only” and “rural-urban focus” studies. Table 2-5: Focus of the research providing rural area type effects All effects Statistically significant effects # of studies # of effects # of studies # of effects Rural focus 2 4 1 2 Rural-urban focus 24 217 18 87 Control only 54 382 38 162 Total 80 603 57 251 Source: Own analysis of literature database. Overall, the database covers research published since 2000, with the majority of effects concentrated between 2000 and 2019 (Table 2-6). The period 2000-2009 had a particularly high share of statistically significant effects (159 effects of which 83, or 52.2%, were statistically significant) compared to the pre-2000 period (68 effects of which 30, or 44.1%, were statistically significant), the 2010 to 2019 period (265 effects out of which 102, or 38.5%, were statistically significant) or the period since 2020 (29 effects out of which 11, or 37.9%, were statistically significant). Table 2-6: Time period of effects Notes: 1 Effects do not align with the discrete periods mentioned above, e.g., because they are based on surveys whose fieldwork spans two calendar years (e.g., the 2009–2010 and 2019–2020 waves of the Understanding Society study). 2 Studies that report effects for multiple periods appear multiple times in this count. Source: Own analysis of literature database. Most effects are based on adult populations (aged 15+), with fewer focusing on young people or older cohorts (Table 2-7). For example, 12 studies provided a total of 32 effects relating to the population aged over 49 (e.g., Efklides, Kalaitzidou and Chankin 2003, Mollenkopf, Kaspar et al. 2004, Lang, Bachinger and Welechovszky 2013, Schmitz and Brandt 2022). Seven studies (32 effects) relate to the population aged under 20 (e.g., Eriksson, Hochwälder 3 Other studies have a rural focus but provide only rural context effects (Oswald, Wahl et al. 2003, Mollenkopf and Kaspar 2005, Brereton, Bullock et al. 2011, Podgorelec, Gregurović and Bogadi 2015). No rural area type or rural context effects are reported in the otherwise relevant studies by Triadó, Villar et al. (2009), Beck and Elkeles (2012), Michalska-Żyła and Marks-Krzyszkowska (2018), Vaznonienė and Wojewódzka-Wiewiórska (2021), Dercan, Zivkovic et al. (2022). All effects Statistically significant effects # of studies # of effects # of studies # of effects before 2000 9 68 5 30 2000-2009 30 159 22 83 2010-2019 33 265 22 102 since 2020 6 29 4 11 Other1 9 82 7 25 Total2 87 603 60 251 | 17 and Sellström 2011, Labiscsak-Erdelyi, Veres-Balajti et al. 2022, Róbert, Geszler and Nagy 2022). 4 Only four studies reported differences in rural-urban wellbeing for males and females (i.e., Ek, Koiranen et al. 2008, Carta, Aguglia et al. 2012, González-Carrasco, Casas et al. 2019, Keller, Groot et al. 2022). Table 2-7: Characteristics of the population for which rural area type effects have been reported Source: Own analysis of literature database Grouping countries according to the welfare state typology developed by Lauzadyte-Tutliene, Balezentis and Goculenko (2018) 5 shows that coverage is uneven: some countries are well-represented, while others are underrepresented (Table 2-8). Specifically, our review database does not include rural area type effects for any countries that fall into the welfare state groups “small European” (i.e., Luxembourg and Malta) or “Eastern European” (i.e., Bulgaria, Estonia, Latvia, Lithuania, Romania). We note, however, that Eastern European countries (and, to a lesser extent, Luxembourg and Malta) are well represented among the studies that pool cases from multiple countries. By contrast, we find that the database includes a large number of effects for the Mediterranean countries. This is partly explained by the high prevalence of regionally stratified studies published for Spain and Italy. For example, Navarro, D'Agostino and Neri (2020) provide wellbeing figures for all regions of Italy and Spain. For further details on the literature database, including statistics for individual countries for which “rural area type” effects on subjective wellbeing have been reported, see Appendix 2 (A2-2). Table 2-8: Geographic dispersion of rural area type effects. 4 Many child well-being studies classify schools by rurality, but we only included effects tied to children’s home addresses, as the home–school distinction matters (González-Carrasco, Casas et al. 2019). Our database thus covers 22 effects by age/sex and home urbanisation, excluding school-location results. 5 We modified groupings by separating Nordic from non-Nordic welfare states, adding a Balkan category (Albania, Kosovo, Macedonia, Serbia), and reallocating Iceland, Ireland, East Germany, and Slovakia based on historical and institutional contexts. Esping-Andersen (2001)’s welfare typology is an alternative but excludes Central and Eastern Europe. All effects Statistically significant effects # of studies # of effects # of studies # of effects Age of the study population Aged 15 or older (no upper bound) 57 502 45 220 Aged 50 or older 12 32 7 12 Aged under 20 7 36 2 8 Aged 18-65 (working age) 4 20 3 7 Aged 15-25 (young people) 4 13 3 4 Total 84 603 60 251 Sex of the study population Both sexes pooled 78 561 54 235 Females 4 21 3 8 Males 4 21 4 8 Total 86 603 61 251 Welfare state group of countries All effects Statistically significant effects | 18 Source: Own analysis of literature database. 2.2.2. Is there a rural advantage in subjective wellbeing? Table 2-9 reports the proportion of rural advantage for all effects recorded in the database 6 ; the right panel focuses on only the rural-urban differences (effects) that, according to the authors of the publication, were statistically significant. In our analysis, we correct all confidence intervals for clustering at the study level. Specifically, CIs and p-values are based on a wild cluster bootstrap-t procedure rather than the usual asymptotic F–test (Cameron, Gelbach and Miller 2008). By using 5000 bootstrap replications with Webb weights, we relax the large-sample and equal-cluster-size assumptions that underlie the standard F-distribution approximation. In particular, our bootstrap approach remains valid when the number of clusters is small, cluster sizes vary markedly, and the error distribution may deviate from normality, all of which can render cluster-robust standard errors and F-tests unreliable. Table 2-9: Rural advantage in subjective wellbeing All effects Statistically significant effects only Mean1 95% CI # of effects Mean1 95% CI # of effects All 0.635 0.547 0.727 603 0.685 0.562 0.813 251 Notes: 1 The mean of a dummy variable with values 0 and 1 multiplied by 100 is the percentage value. We show the means in the graphs, seeing as values between 0 and 1 scale better in visualisations. Source: Own analysis of literature review database. It can be seen that six in ten effects indicate that the wellbeing of the population living in rural areas is higher than that of the population living in urban (or less rural) areas (mean: 63.5%, 95% CI [54.7, 72.7]). Among the 251 statistically significant effects, this figure amounts to seven in ten effects (mean: 68.5%, 95% CI [56.2, 81.3]). 2.2.3. What factors shape the rural advantage in subjective wellbeing? 6 Note that some studies reported effects using an urban base category; in such cases, we reversed the base (and the effect direction in the rural wellbeing advantage analysis). For example, if “cities” were the reference and comparisons included “farms”, “villages”, “small towns”, etc., we only extracted coefficients for “farms” and “villages” vs. cities, reversing the sign, and ignoring the “small towns” vs. cities comparison. If “farm” was the base, we extracted all comparisons without changes. If “village” was the base, we reversed the farm-to-village coefficient and kept all others as reported. # of studies # of effects # of studies # of effects Eastern European 0 0 0 0 Small European 0 0 0 0 Central European 14 67 7 27 Balkan states 2 3 1 1 Mediterranean 12 129 10 39 Old Europeannot Nordic 30 141 18 54 Old EuropeanNordic 7 34 3 7 Multiple countries pooled 26 229 23 123 Total 91 603 62 251 | 19 2.2.3.1. Time, geography and study population Next, we report the proportion of effects that indicate rural advantage by several characteristics, which may underpin the rural advantage in subjective wellbeing. For results in tabular format, see the right-hand side panels of Appendix 2 (A2-2). A subgroup is considered significantly different from the overall mean if its 95% CI lies entirely above or entirely below the relevant overall mean (see Schenker and Gentleman 2001). For ease of interpretation, we visualise the results in plots, adding the 95% confidence interval around the mean values and differentiating between all effects (yellow) and statistically significant effects (green). The added vertical lines represent the overall mean (yellow dashed line) and the overall mean among statistically significant rural advantage effects (green dasheddotted line). Temporal variation, as shown in Figure 2-2, suggests that the rural advantage was more frequently reported before 2000 and after 2020. However, the confidence intervals are wide and only the statistically significant rural advantages reported pre-2000 reach statistical significance (Msig = 0.976, 95% CI [0.777, 1.001]). Figure 2-2: Proportion rural advantage in subjective wellbeing, by time period studied Source: Own analysis of literature review database. Demographic breakdowns, as shown in Figure 2-3, suggest that rural advantage is less evident among those aged 49 or older (Mall = 0.313, 95% CI [0.145, 0443]) and young people (aged 15–25; Mall = 0.231, 95% CI [-0.515, 0.579]). However, the number of subgroup-specific studies is limited, and estimates are imprecise. Geographic patterns, as shown in Figure 2-4, show that single-country studies from Central Europe less often report a rural advantage (Msig = 0.222, 95% CI [-0.018, 0.370]), while multi-country studies tend to do so more frequently. However, this difference just failed to reach statistical significance (Msig = 0.846, 95% CI [0.680, 0.992]). Given the lack of single-country effects reported for countries in the “Eastern European”, “Balkan” 7 , and “small European countries” welfare state groups, we also tagged whether the multiple-country studies included countries from these welfare state groups. The results suggest that the inclusion or exclusion of any country within the tagged group does not impact the overall mean. 7 While we observe two studies that report rural area type effects for North Macedonia and Montenegro, respectively, wild bootstrap CIs cannot be reliably computed for sample sizes of less than N=3 or when there is no variation in the outcome variable. In this case, neither study reported a rural advantage. We refer the Reader to Appendix 2-1 for further details on why specific CIs are not reported. | 20 Figure 2-3: Proportion rural advantage in subjective wellbeing, by age of the population studied Notes: CIs capped at -.05 and 1.5 for ease of display. See Appendix 2 (A2-2) for values of the computed wild bootstrap CIs. Source: Own analysis of literature review database. Figure 2-4: Proportion rural advantage in subjective wellbeing, by (groups of) countries studied Notes: CIs capped at -.05 and 1.5 for ease of display. See Appendix 2 (A2-2) for values of the computed wild bootstrap CIs. Source: Own analysis of literature review database. | 21 2.2.3.2. Measurement of subjective wellbeing The research on rural-urban differences in subjective wellbeing is dominated by the use of evaluative wellbeing measures, with 70 studies contributing 549 out of a total of 603 effects. Eighteen studies with 54 effects provided empirical results on how an affective wellbeing measure varies between rural and urban (or between different types of rural) areas. Life satisfaction measures are by far the most prevalent measures, with 51 studies contributing 426 effects. Next up are “Other subjective wellbeing” outcomes (23 studies with 111 effects). 8 Figure 2-5 shows that the evaluative wellbeing outcomes have a slightly higher proportion of effects indicating rural advantage than the affective wellbeing outcomes. But the difference to the overall mean fails to reach statistical significance (evaluative wellbeing: Mall = 0.654, 95% CI [0.561, 0.753]; Msig = 0.715, 95% CI [0.581, 0.855]; affective wellbeing: Mall = 0.444, 95% CI [0.206, 0.735], Msig = 0.391, 95% CI [-0.006, 0.922]). The higher proportion of effects indicating rural advantage in the evaluative wellbeing dimension is driven by studies focusing on life satisfaction. Among the statistically significant effects based on life satisfaction as an outcome, rural advantage is found in 82.1% of effects (Msig = 0.821, 95% CI [0.677, 0.957]). The group of studies examining "Other subjective wellbeing" evaluatively has a particularly low rate of effects indicating rural advantage in wellbeing (Mall = 0.441, 95% CI [0.290, 0.631]; Msig = 0.387, 95% CI [0.104, 0.625]). Overall, therefore, the urban happiness paradox is essentially a paradox of urban life satisfaction. Figure 2-5: Proportion rural advantage in subjective wellbeing, by wellbeing outcome studied Notes: CIs capped at -.05 and 1.5 for ease of display. See Appendix 2 (A2-2) for values of the computed wild bootstrap CIs. Source: Own analysis of literature review database. Figure 2-6 highlights the instruments measuring wellbeing and their connection to the rural wellbeing advantage. The 1-item life satisfaction scale (0 to 10) is the most frequently used evaluative wellbeing instrument, used in 27 8 We added to this group instruments that combine life satisfaction and happiness measures, or incorporate other components of well-being, such as meaning or purpose in life. | 22 studies with 285 effects (e.g., Mollenkopf, Kaspar et al. 2004, Knies 2012, Ivlevs 2017). 9 Other 1-item scales (1-10, 1-7, 1-5, 1-4) and shorter versions focused on the most or least satisfied groups were used in fewer studies. Multiitem scales such as Diener’s Satisfaction with Life Scale and Huebner’s Students’ Life Satisfaction Scale appeared in only five studies with 36 effects (e.g., Hooghe and Vanhoutte 2011, Podgorelec, Gregurović and Bogadi 2015, Róbert, Geszler and Nagy 2022). Global measures of happiness (0-10 and 1-10 scales) appeared in seven studies with 37 effects (e.g., Shucksmith, Cameron et al. 2009, Rodríguez-Pose and von Berlepsch 2014, Requena 2016). Subjective Quality of Life (QoL) measures were less frequently studied, with eight studies reporting 19 effects, the Cantril ladder being the most frequently used (e.g., Brereton, Clinch and Ferreira 2008, Eriksson, Hochwälder and Sellström 2011, Graham and Nikolova 2015). 10 Other measures fall into the miscellaneous category, and there is no single, widely used instrument for assessing affective wellbeing: 1-item happiness measures with response formats ranging from 0–10, 1–5, and 1–6 (condensed) each appeared in only one study. The Positive and Negative Affect Schedule (PANAS) and the General Health Questionnaire (GHQ-12) were used in two studies, the GHQ-12 yielding one of the highest instance counts (10 effects). 11 Seven studies (25 effects) used other affective wellbeing measures, reflecting the diversity of alternative measures used in wellbeing research. Figure 2-6: Proportion rural advantage in subjective wellbeing, by wellbeing instruments studied Notes: CIs capped at -.05 and 1.5 for ease of display. See Appendix 2 (A2-2) for values of the computed wild bootstrap CIs. Source: Own analysis of literature review database. 9 The wording of life satisfaction questions varies slightly across studies (e.g., “nowadays”, “currently”, or “overall”, and using scale endpoints like “complete”, “total” or “extreme”), but we treat these as equivalent. Surveys such as the European Social Survey adjust wording across languages for conceptual comparability. We assume all Eurobarometer studies used the 10-point life satisfaction scale unless otherwise specified (cf. GESIS-Leibniz Institute for the Social Sciences n.d.). 10 The Cantril Self-Anchoring Striving Scale (Cantril 1965) is often presented as a life satisfaction instrument. Respondents are asked to imagine a ladder with steps numbered from zero at the bottom, representing the worst possible life, to 10 at the top, representing the best possible life. They then pick the number which represents the step of the ladder where they stand at this time. We consider it a measure of subjective Quality of Life. Also see Glatzer and Gulyas (2014). 11 GHQ-12 is a widely used screening tool for assessing mental health and psychological distress. We did not specify the measure as one of the general subjective well-being outcomes of interest that we specifically set out to include in the review. A future study may consider increasing efforts to incorporate GHQ-12-based studies into this literature review database. | 23 As shown in Figure 2-6, among the evaluative wellbeing instruments, studies that use short 1-item life satisfaction scales have a particularly high rate of effects indicating a rural advantage. All effects based on studies using a 1-7 or a 1-4 life satisfaction scale indicate a rural wellbeing advantage (as there is no variation, there are no confidence intervals). Almost all statistically significant effects using "other 1-item life satisfaction scales" (which include condensed versions of the longer 10or 11-point scales) report rural advantage (Msig = 0.909, 95% CI [0.604, 1.655]). Conversely, none of the five studies that investigated differences in subjective Quality of Life using the Cantril ladder (both the longer 11-point scale and the shorter 7-point scale) reported a rural wellbeing advantage. While studies using 1-item happiness scales with 11-point scales report a high share of rural advantage (Mall = 0.742, 95% CI [0.598, 1.060]), those using shorter 1-10 scales report a lower share of rural advantage (Mall = 0.167, 95% CI [- 2343612, 97651]), albeit the CIs are very large as this instrument is not used in many studies. There is no clear pattern regarding the role of affective wellbeing instruments in the rural wellbeing advantage; the number of studies reporting statistically significant effects is low (N = 18), as shown in Appendix 2 (A2-2). 2.2.3.3. Operationalisation of Rural and Urban Areas When recording the definitions of rural-urban areas used in the research, we followed Knies and Melo (2019) and distinguished between administrative definitions of rural-urban areas and survey-based definitions. We differentiate between classifications that describe the local context at geographical scales below the level of municipalities (e.g., postcodes or 1 km grids), at the scale of municipalities (e.g., LAU or NUTS5), and scales larger than municipalities (e.g., counties, NUTS3 or NUTS2). Additionally, some administrative classifications relate to built-up areas (e.g., aggregations of 100-metre grids or small-scale Census output areas, such as the UK rural-urban classification of output areas). Of the 80 studies in our review, 53 studies used administrative classifications (338 effects of which 126, or 37.3%, were statistically significant), and 31 studies relied on respondentor interviewer-perceived classifications (265 effects of which 125, or 47.2%, were statistically significant). Within the administrative definitions, markedly higher shares of statistically significant differences in subjective wellbeing between rural and urban populations were reported in studies that used built-up areas (45 out of 98 = 45.9%) or below municipality-level (10 out of 20 effects = 50%) classifications. In comparison, the share is markedly lower for municipality-level (19 out of 65 = 29.2%) and above municipality-level (52 out of 155 effects = 33.6%) classifications. There is considerable heterogeneity in the area definitions used at these scales. For detailed results, see Appendix 3 (A2-3). For example, among the six studies that measured rurality below municipality-level, one study used Eurostat’s Degree of Urbanisation (DEGURBA) 12 , two studies used population density, one used population-size bins, and a further two studies compared subjective wellbeing in specific small rural places with wellbeing in specific urban areas. 13 At the municipality level, researchers employed a wide range of measures—from continuous population density and nuanced population-size binning strategies to more complex definitions combining size or density with land use, economic structure, and proximity to urban centres. Interestingly, all studies that use population size bins or population density bins condense the classifications, most frequently creating a rural-urban binary. Above the municipality level, DEGURBA dominates—six studies applied it at the NUTS3 level (one condensing it into a rural-urban binary), yielding 103 separate effects, alongside population-based metrics (density or size bins), though no single measure prevails. Classifications of built-up areas tend to rely on settlement size bins, often condensed into binary rural-urban, with no specific measure standing out as providing a particularly high share of statistically significant results on rural-urban wellbeing differentials. Additionally, survey questions ask respondents (or interviewers) to classify the respondents' place of residence in terms of settlement type (e.g., open countryside, village, small town, mid-sized town or big city) and/or size (e.g., less than 10k inhabitants, more than 500k inhabitants). Settlement type, condensed to a simple rural-urban dichotomy, is the most frequently used instrument in this “respondent/interviewer-perceived” category, with 15 studies and 89 reported effects (of which 56, 12 The research uses the original DEGURBA. It combines fine-grained population size and population density thresholds to establish three mutually exclusive classes (cities, towns and suburbs, and rural areas. For further information, see https://ec.europa.eu/eurostat/web/degree-of-urbanisation/information-data. 13 While it is common to compare well-being in different rural localities, due to data protection legislation, the research does not usually provide detailed information about population size or other parameters we need for this review. Hence, our review database does contain very few studies of this type. | 24 or 62.9%, are statistically significant). This suggests a strong preference for binary or condensed classifications when relying on respondents' assessments. A reason for this is that general population studies typically do not include large enough samples of individuals from all types of settlements. Figure 2-7: Proportion rural advantage in subjective wellbeing, by geographic scale of effects studied Source: Own analysis of literature review database. Figure 2-8: Proportion rural advantage in subjective wellbeing, by area classification utilised Notes: CIs capped at -.05 and 1.5 for ease of display. See Appendix 2 (A2-2) for values of the computed wild bootstrap CIs. Source: Own analysis of literature review database. | 31 his childhood home in rural Kirkcaldy after years of teaching in Glasgow) serenated the beauty, pleasures, and tranquillity of life in the countryside (Berry and Okulicz-Kozaryn 2009). Happiness researchers suggest that there is a socio-biological component to human preferences for living in smaller communities and closer to nature (Veenhoven 1994). Notwithstanding an apparent over-idealisation of country life (Veenhoven 1994, Shucksmith 2018), empirical studies have found support for the positive impacts of access to green space on stress exposure and coping (van den Berg, Maas et al. 2010), as well as physical health and psychological wellbeing (Wolch, Byrne and Newell 2014). Similarly, Sorensen (2014) finds that access to natural amenities is a significant predictor of rural wellbeing advantages and is also an important contributor to variation within urban areas (also see Ambrey 2016). A parallel to these longstanding theories emerged during the COVID-19 pandemic, when many urban residents relocated to scenic, nature-rich regions—an implicit demonstration of the same inherent pull toward restorative rural environments posited by 18th and 19th-century thinkers (González-Leonardo, López-Gay et al. 2022). On the other hand, the eroding effects of urbanism on social life and social trust were outlined (Tönnies 1887). They became, for example, codified as a general mechanism of social ill in the social disorganisation theory of the Chicago School (e.g., Wirth 1938) and the Shefky and Bell models of social areas (e.g., Bell and Force 1956). Moreover, epidemiological research in the USA up to the 1970s supported the assumption of rural wellbeing and urban malaise when measured using mental health indicators (Armstrong 1991). Research by Burger, Morrison et al. (2020) finds that in Northern Europe, Western Europe and the Anglo-Saxon countries, one of the factors explaining higher rural happiness is community attachment. 15 In all likelihood, therefore, we would expect rural-urban differences to be more pronounced in the dimensions of social wellbeing and psychological functioning than in evaluative wellbeing. While economic theory has stressed the positive role that cities can play in peoples’ lives - diverse kinds of people meet, new ideas emerge, cultural events are plenty, educational institutions are prevalent, businesses strive, employment opportunities are varied, etc. (e.g., Glaeser 2011) - more generalised theories argue that an overall opportunity structure for human flourishing needs to consider economic opportunities, public and private services, opportunities for community and civic engagement, and the accessibility of the natural and built environment (Bernard, Steinführer et al. 2022), i.e., suggesting a focus on a greater range of spaces and a move away from the urban rural dichotomy, which we still find in much of the empirical literature (see, e.g., chapter 2 in this report). In this view, individual preferences for small-town and rural living are balanced with the need for employment and access to other amenities of urban life, suggesting a potential advantage of suburban or, especially, rural and smalltown living in commuting distance from urban centres (Fuguitt and Brown 1990, Biagi and Meleddu 2023, Finnemann, Huth et al. 2024). This position corresponds closely with migration flows away from more remote rural regions and urban centres, leading to population accumulations in accessible rural areas (Shucksmith 2012). Indeed, theory and some empirical research point towards a “sweet spot” of rural-urban balance (Lenzi and Perucca 2020, Hoogerbrugge, Burger and Van Oort 2022). Finnemann, Huth et al. (2024) find that while urban areas in the United Kingdom show lower degrees of life satisfaction, there are optimal distances – defined as the radius in which population average life satisfaction is highest on average and shows the least variation - which “tend to cluster in the hinterlands of cities”. The potential of suburban or small-town and small-city advantages is also supported by Biagi and Meleddu (2023), focusing on Italy. The authors conclude that “it is not living in rural areas per se that guarantees a higher life satisfaction but actually living in rural areas where labour, leisure, social relations and access to amenities are more elevated in terms of availability and quality compared to other areas” (cf. Weckroth and Kemppainen 2021, who find no support for the "Sweet spot" hypothesis using data for a greater range of EU countries). In conclusion, several studies and theories suggest that research on rural-urban differences in wellbeing should consider non-binary rural-urban classifications, ideally differentiating between different types of rural settlements and using nuanced definitions of wellbeing, as this may help uncover pathways to improving wellbeing in diverse contexts. 15 However, researchers have questioned whether lower levels of SWB in cities are really due to weaker social ties and trust or simply due to concentration of negative structural factors such as poverty (Fischer 1973; Okulicz-Kozaryn & Mazelis, 2018). | 32 3.1.2. Wellbeing as subjective experience Subjective wellbeing is a complex concept that relates to optimal experience and functioning. There are two main theoretical perspectives on subjective wellbeing. The first is the hedonic approach, which emphasises happiness and defines wellbeing in terms of the pursuit of pleasure and the avoidance of pain (e.g., Waterman 2008). Subjective evaluations of life, such as life satisfaction and subjective quality of life, are prominent outcomes studied under the umbrella of the hedonic wellbeing approach. The second is the eudaimonic approach, which emphasises meaning and self-realisation and defines wellbeing by the extent to which a person is fulfilling their potential and identifying meaningful pursuits in life (e.g., Waterman 2007). The two approaches are not straightforwardly integrated, and research findings are heterogeneous, in some cases conflicting (Ryan and Deci 2001). In their proposal to repeat a revised comprehensive module on “Personal and social wellbeing” on Round 12 of the European Social Survey (ESS), Martela, Delle-Fave et al. (2023) provide an integrative wellbeing framework that builds on self-determination theory (Ryan and Deci 2017) and the previous module developed by Huppert, Marks et al. (2010). In the framework, personal wellbeing comprises psychological functioning and experienced wellbeing and is complemented by social wellbeing (i.e., the quality of the social environment in which individuals operate), see Figure 3-1. The scholars stress the importance of satisfaction of basic psychological needs - like autonomy (feeling in control of one’s life), competence (feeling capable), and relatedness (having supportive relationships) - and mental resources related to human flourishing -such as resilience (staying mentally healthy during tough times), mindfulness (noticing and understanding one’s emotions), optimism (having positive expectations for the future), and prosocial behaviours (helping and supporting others) – for personal wellbeing, and, via the individual’s interaction with their environment, for social wellbeing. Figure 3-1: Conceptual framework for dimensions of subjective wellbeing. Source: Figure adapted from Martela, Delle-Fave et al. (2023). As regards the factors leading to optimal experience and functioning, three perspectives have been identified (Brief, Butcher et al. 1993). The bottom-up approach builds on the philosophical assumption that there are universal needs that must be met for people to be happy. The settlement of needs is thought to be dependent upon external factors, and people who find themselves in a ‘good situation’ for the fulfilment of needs are happy, while those who find themselves in a ‘bad situation’ are unhappy (see, e.g., Diener, Suh et al. 1999). The context in which individuals can be happy is implicitly thought to be the same for everybody, and people will be happier the more moments of happiness they experience. It follows from this that the independent variables employed in bottom-up models are | 33 objective life circumstances (providing a theoretical link also to objective wellbeing factors frequently used to monitor population wellbeing, see, e.g., the GRANULAR rural wellbeing compass). As regards the factors leading to optimal experience and functioning, three perspectives have been identified (Brief, Butcher et al. 1993). The bottom-up approach builds on the philosophical assumption that there are universal needs which have to be met for people to be happy. The settlement of needs is thought to be dependent upon external factors, and people who find themselves in a ‘good situation’ for the fulfilment of needs are happy, while those who find themselves in a ‘bad situation’ are unhappy (see, e.g., Diener, Suh et al. 1999). The context in which individuals can be happy is implicitly thought to be the same for everybody, and people will be happier the more moments of happiness they experience. It follows from this that the independent variables employed in bottom-up models are objective life circumstances. An alternative philosophical theory to understanding happiness is the so-called topdown approach. These models assume that subjective wellbeing is influenced by characteristics that are internal to the individual. Among the factors these models include are personality traits such as determination, optimism, and self-confidence. Moreover, a person–environment fit perspective suggests that the congruence between an individual’s attributes and their environmental demands also shapes subjective wellbeing: when personal traits “fit” the context, people report higher happiness, whereas misfit can undermine wellbeing (Götz, Ebert and Rentfrow 2018, Morrison and Weckroth 2018, Hanell 2022). While some characteristics, such as employment, health, and marital status, will play a role in the top-down models, the focus will be on the subjective evaluations of these states rather than the objective states. Last but not least, the third type of models acknowledges the multiple interactions between the internal and external context in which individuals operate. The so-called interactionist models recognise, for instance, that married people are happier than non-married people, but also that more optimistic and happier people are more likely to get married. 3.1.3. Empirical research on correlates of SWB The bulk of the quantitative empirical research has focused on life satisfaction, as also seen in the research on ruralurban differences in the wellbeing of the population living in European rural areas (see Chapter 2). Research on life satisfaction has ascertained a number of interesting and consistent relationships between individual characteristics and life satisfaction (Layard 2005, Dolan, Peasgood and White 2008, Kaiser and Oswald 2022). First, life satisfaction exhibits a U-shaped pattern with age, with life satisfaction typically being at its lowest in midlife (e.g., Blanchflower and Oswald 2008, Kaiser, Otterbach and Sousa-Poza 2022). Second, unemployment (Clark and Oswald 1994) and a lower level of financial wellbeing (see, e.g., Easterlin 1974, Frijters, Haisken-DeNew and Shields 2004) are associated with lower life satisfaction. Third, people who are married are more satisfied with life than never-married singles, divorcees (including those living in separation) and widowers (see, e.g., Shapiro and Keyes 2008). A further consistent finding is that individuals who belong to a religion tend to be more satisfied with their lives (Lim and Putnam 2010). Lastly, it has been found that markers of poor health are significant factors in explaining lower selfreported levels of life satisfaction (Brief, Butcher et al. 1993, Diener, Suh et al. 1999). Diener and Chan (2011) demonstrate that happier individuals also tend to live healthier and longer lives. Findings concerning other individual characteristics typically included in life satisfaction models (such as sex, education, and the number of children in the household) are mixed (Frijters, Haisken-DeNew and Shields 2004). Research examining various dimensions of subjective wellbeing is currently sparse, and systematic reviews on this topic are not yet available. The few available studies that examined multiple dimensions of subjective wellbeing in a comparative framework have shown the usefulness of these conceptualisations in capturing domain-specific differences in SWB across countries and regions (Huppert and So 2013, Ruggeri, Garcia-Garzon et al. 2020, Martela, Lehmus-Sun et al. 2023). A consistent finding from these studies is that countries that score similarly on life satisfaction can differ significantly on other dimensions of subjective wellbeing (Huppert and So 2013, Ruggeri, Garcia-Garzon et al. 2020). For example, Huppert and So (2013) found that Germany, middling among 23 countries on the overall flourishing scale, ranked second and third place on the (mental resources) dimensions of self-esteem and emotional stability, yet 17th and 18th place on positive relationships, engagement and meaning. Michaelson, Abdallah et al. (2009), using the same ESS data as Huppert and So (2013) (and which we, too, will use), also pointed out that many countries that score high on general markers of wellbeing show remarkable differences across wellbeing dimensions. For example, Hungary shows a “very low score for emotional (wellbeing) – absence of | 34 negative feelings and a considerably above-average score for trust and belonging.” Sánchez-Sellero, García-Carro and Sánchez-Sellero (2021), using the 2016 European Quality of Life Survey (EQLS), identified five different dimensions of evaluative wellbeing – in their terminology ‘components of subjective quality of life’: (1) satisfaction with governance (i.e., trust in legal system, parliament and government and satisfaction with democracy and economy), (2) subjective quality of public services (i.e., health services, public transport, education and pension system), (3) satisfaction with the immediate environment (i.e., satisfaction with accommodation, local area and family life), (4) general satisfaction with life and happiness, and (5) satisfaction with socioeconomic situation (i.e., satisfaction with education and standard of living, no difficulties making ends meet) and found sizeable consistency of ratings across these dimensions of subjective quality of life, meaning countries that scored overall high on general life satisfaction also scored high on the other dimensions of subjective quality of life. 16 Regarding the correlates of different dimensions of subjective wellbeing, Ruggeri, Garcia-Garzon et al. (2020) reported lower psychological wellbeing scores for women compared to men, albeit these differences are only significant for the overall sample and four out of 27 countries (the Netherlands, Belgium, Cyprus, and Portugal). They also reported significantly higher scores for younger age groups compared to older age groups, as well as for those with higher education and employed respondents compared to unemployed respondents. Martela, LehmusSun et al. (2023), using data from 27 countries participating in the ESS Round 6, examined how different dimensions of personal wellbeing significantly predict happiness, life satisfaction, and depressive symptoms. Specifically, the authors demonstrate that satisfaction of basic psychological needs, as well as its subdomains of autonomy, competence, and relatedness, are strong predictors of life satisfaction, happiness, and depression. The associations are less pronounced with several socioeconomic and demographic factors, indicating the independent effects of these psychological measures of wellbeing. Consistent with past research, they also find country-specific differences in the associations. Some countries exhibit stronger associations with happiness, others with life satisfaction or depressive symptoms, although the authors do not identify a clear pattern in these associations (Martela, LehmusSun et al. 2023). An increasing amount of research has investigated the effects of mindfulness-based practices, such as meditation and yoga, on subjective wellbeing. These studies indicate that such practices can positively influence mood, reduce stress levels, and enhance overall wellbeing (Brown and Ryan 2003). Furthermore, cognitive-behavioural therapies, including cognitive-behavioural therapy, have proven effective in fostering wellbeing, particularly among individuals experiencing depression and anxiety (Cuijpers, Berking et al. 2013). From the perspective of policymaking, a focus on the basic psychological need satisfaction and mental resources dimensions of subjective wellbeing, therefore, seems promising in the pursuit of achieving greater happiness for the greatest number of people (a commonly set goal for policymaking and individuals). To our knowledge, no study has investigated the differences in subjective wellbeing across multiple dimensions of subjective wellbeing along the rural-urban continuum. The only exception – if we consider different dimensions of evaluative wellbeing to be different dimensions of subjective wellbeing – is Sánchez-Sellero, García-Carro and Sánchez-Sellero (2021), who used Eurostat’s trichotomous degree of urbanisation (degurba) indicator, which differentiates between cities (densely populated areas, at least 50% of the population lives in urban centres), suburban cities (areas of intermediate density), and rural areas (sparsely populated areas, more than 50% of the population lives in rural areas). The study finds no significant differences in the overall measure of subjective quality of life, nor in the satisfactions most linked to the personal and socio-economic environment. By contrast, there are significant differences in those satisfactions related to public services and the government across different levels of urbanisation. As past studies have noted, the underlying reasons for the differences in SWB between rural and urban areas are not yet well understood (Weckroth and Kemppainen 2021). One key factor limiting our understanding of rural-urban differences might be the one-sided focus on the evaluative dimensions of SWB. Studies using multidimensional 16 There are some countries for which the rank across the different dimensions varies by up to three positions only (UK, Bulgaria, Finland and Latvia), while two countries occupy positions at different poles of the league tables in some dimensions: Cyprus ranks in the bottom third place concerning quality of services but in a top 10 position (20th) regarding the immediate environment while Belgium ranks sixth concerning environment but 21st concerning quality of services. Fourteen countries have a difference of seven or less in rank positions across dimensions of subjective quality of life. | 35 measures of SWB have shown the immense usefulness in comparing SWB profiles across countries (Michaelson, Abdallah et al. 2009, Huppert and So 2013, European Social Survey 2015, Ruggeri, Garcia-Garzon et al. 2020, Sánchez-Sellero, García-Carro and Sánchez-Sellero 2021, Martela, Lehmus-Sun et al. 2023). Moreover, the multidisciplinary approach to SWB has refocused the social dimension as a dimension of human wellbeing (Martela, Delle-Fave et al. 2023). For instance, research into the social determinants of health has highlighted that negative social experiences, including generalised social distrust and anxiety, as well as experiences of discrimination, can impact health and wellbeing through stress responses and other mechanisms (Braveman, Egerter and Mockenhaupt 2011). Aspects of social wellbeing, such as social cohesion, trust, respect, discrimination, safety, and loneliness, can impact overall SWB (Martela, Delle-Fave et al. 2023). Especially, loneliness is currently experiencing a renaissance of interest. While already central to several analyses of modern and urban life in Western countries (Putnam 1995), current research underscores the health risks associated with loneliness, including premature death. Consequently, several countries have reacted by appointing special ministers for loneliness (Wurm, Ehrlich et al. 2023), and development approaches which are sensitive to place attachment and belonging have been recommended for left-behind places (MacKinnon, Kempton et al. 2021). In summary, while research into SWB is still heavily dominated by a focus on the evaluative dimension of SWB, the outlined research on the role of social ties for wellbeing and current concerns about the role of loneliness and, especially, the proposed close connection of urban anonymity and lack of social connection might make this dimension highly relevant to assess differences in SWB across the rural-urban continuum. 3.2. Data and methods We use data from Round 6 (2012/13) of the European Social Survey (ESS), an academically driven cross-national survey conducted biannually across Europe since 2002/3 (Schnaudt, Weinhardt et al. 2014). The aim of the ESS is to provide high-quality, harmonised data to monitor the attitudes, beliefs, wellbeing, and behaviour patterns of diverse populations in Europe. ESS data is available free of charge and can be downloaded from the study website, which also provides a comprehensive study documentation (https://ess.sikt.no). The study adopts a complex survey design (Kaminska and Lynn 2017). Sampling files containing information about primary sampling and stratification units, as well as weights for early rounds of the study, are deposited with the country-level study documentation. Additionally, all survey instruments are included in the respective countries’ languages. The questionnaire in each round of the ESS consists of a core and two in-depth (repeated) modules. Round 6 of the ESS (European Social Survey European Research Infrastructure (ESS ERIC) 2023), which was fielded in 29 countries in 2012/13, was chosen for this analysis because it contains the latest in-depth module on “Personal and social wellbeing” (Huppert, Marks et al. 2010) and allows us to measure different dimensions of subjective wellbeing simultaneously in a large representative sample of the population living in Europe (subject to accounting for the complex cross-national survey design). Overall, approximately 54,600 people aged 15 or older living in private households in participating countries participated in the personal interviews, which were conducted either face-toface or by phone (Jeffrey, Abdallah and Quick 2015). Operationalising dimensions of subjective wellbeing We had shortlisted the ESS Round 6 questionnaire as a tool that we could share in different languages with local actors across rural areas in Europe to collect information on and monitor the subjective wellbeing of the rural population. The “Personal and social wellbeing” module within the questionnaire comprised a total of 42 items. It was complemented by questions, which are included in each round of the ESS (so-called core questions), such as the general life satisfaction and happiness questions. The Round 6 questionnaire is, therefore, quite cumbersome. Moreover, through our involvement in the ESS Round 12 questionnaire development team 17 we anticipated that the module being prepared for the ESS 2025/26 (Round 12) would include considerably fewer items (30 overall, of which 10 items would be new) than the Round 6 module. We bore this in mind when setting up our analyses. Specifically, after mapping the ESS Round 6 wellbeing items onto the conceptual framework presented by Martela, Delle-Fave 17 GRANULAR team member Gundi Knies of Thünen-Institute (Germany) is one of the scientific leads of the ESS Round 12 “Personal and social well-being” module proposal. | 36 et al. (2023), we picked that subset of items that (we knew from the cognitive and quantitative testing of the Round 12 module 18 ) would be repeated in the Round 12 module, or would be the most suitable substitutes for any Round 6 items that would be dropped. The alignment of survey questions and SWB dimensions, as well as the question wording, is listed in Table 3-1. We compute the Mazziotta-Pareto Index (MPI, Mazziotta and Pareto 2012), a composite index 19 to measure multidimensional phenomena (here: dimensions of subjective wellbeing). We reverse-coded all negatively worded questions so that higher values represent higher wellbeing. Considering the matrix 𝑌 = 𝑦𝑖𝑗 with n rows (statistical units, here: individuals living in EEA member states that participated in ESS Round 6) and m columns (individual subjective wellbeing items within an SWB dimension), we first calculate the normalized matrix 𝑍 = 𝑧𝑖𝑗 as follows: 𝑧𝑖𝑗 =100 ±(𝑦𝑖𝑗 − 𝑀𝑦𝑗) 𝑆𝑦𝑗 10 where 𝑀𝑦𝑗 and 𝑆𝑦𝑗 are, respectively, the mean and standard deviation of the indicator 𝑗 in the population 20 and the sign ± is the 'polarity' of the indicator 𝑗. The sign indicates the relation between the indicator 𝑗and the phenomenon to be measured. We use + seeing as all our indicators represent a dimension of a positive good (wellbeing). Denoting with 𝑀𝑧𝑖 , 𝑆𝑧𝑖 , and 𝑐𝑣𝑧𝑖 the mean, standard deviation, and coefficient of variation of the normalised values for the unit 𝑖, respectively, the composite index is given by 𝑀𝑃𝐼𝑖 ±= 𝑀𝑧𝑖± 𝑆𝑧𝑖∗𝑐𝑣𝑧𝑖 where the sign depends on the kind of phenomenon to be measured. When the composite index is 'increasing' or 'positive', i.e., increasing values of the index correspond to positive variations of the phenomenon, it is appropriate to use . The final MPI makes it possible to make relative comparisons: Values above 100 indicate above-average subjective wellbeing for the unit 𝑖 compared to other people living in Europe, and values below 100 indicate values below the average. Operationalisation of rural. To contribute to the outlined debates on rural-urban differences in SWB, we use the ESS settlement type variable as our main independent variable. The ESS routinely asks participants to report “(w)hich (…) best describes the area where (they) live?”. The response options range from the largest, and most urbanised, settlement type [1] “A big city” (dubbed here: cities), [2] “the suburbs or outskirts of big city” (dubbed here: suburbs), [3] “a town or small city” (dubbed here: towns), [4] “a country village” (dubbed here: villages), to the smallest (and least populated) settlement type [5] “farm or home in countryside” (dubbed here: countryside). Other independent variables. We include basic socio-demographic and socio-economic characteristics linked to life satisfaction. These are age group (15-24, 25-54, 55-64, 65 or older), sex, citizenship in the country of residence, marital status (married, separated/divorced, widowed, never married), living with children in the household, health limitations (no, slight, severe), Highest level of education (lower secondary or less, upper secondary, advanced vocational or sub-degree, degree or higher), main economic activity status in the last fortnight (employed, education, unemployed, permanently sick or disabled, retired, housework, looking after children, others, other), and net household income (deciles). 18 Unfortunately, it is not possible to share the results of this experimental work outside the ESS development team. 19 Composite indicators are frequently used to combine a great deal of information in a simplified dashboard-style statistic, see Saisana and Tarantola (2002). Notwithstanding their usefulness, there is continued debate about the appropriateness of composite (or synthetic) indicators. The approach relies on a series of decisions regarding the aggregation and weighting of the constituent items in the index. Studies have shown that the method of index construction and weighting procedures can severely affect evaluations of outcomes, see Greco, Ishizaka et al. (2019). 20 Given that the individuals participating in the ESS are not a completely random subset of the population, we apply the population weights and compute standard errors adjusted for clustering and stratification. | 37 Table 3-1: Mapping of the ESS Round 6 subjective wellbeing items onto the Round 12 subjective wellbeing framework Dimension Construct Question wording Variable name Basic psychological needs satisfaction Autonomy1 To what extent do you make time to do the things you really want to do? tmdotwa I feel I am free to decide for myself how to live my life dclvlf Competence In my daily life I get very little chance to show how capable I am. lchshcp Most days I feel a sense of accomplishment from what I do accdng2 There are lots of things I feel I am good at. lotsgot2 Relatedness3 How often socially meet with friends, relatives or colleagues sclmeet Take part in social activities compared to others of same age sclact How many people with whom you can discuss intimate and personal matters inprdsc How close do you feel to the people in your local area? flclpla Mental resources6 Resilience4 How difficult or easy do you find it to deal with important problems that come up in your life? deaimpp When things go wrong in my life, it generally takes me a long time to get back to normal. wrbknrm Mindfulness5 On a typical day, how often do you take notice of and appreciate your surroundings? tnapsur Optimism I’m always optimistic about my future optftr Prosociality7 To what extent do you provide help and support to people you are close to when they need it? prhlppl In the past 12 months, how often did you get involved in work for voluntary or charitable organisations? wkvlorg Affective wellbeing Positive affect How much of the time during the past week were you happy? wrhpp How much of the time during the past week did you have a lot of energy? enrglot How much of the time during the past week did you enjoy life? enjlf How much of the time during the past week did you feel calm and peaceful? fltpcfl Negative affect How much of the time during the past week did you feel sad? fltsd How much of the time during the past week did you feel depressed? fltdpr How much of the time during the past week was your sleep restless? slprl How much of the time during the past week could you not get going? cldgng How much of the time during the past week did you feel anxious? fltanx Evaluative wellbeing8 Life satisfaction How satisfied with life as a whole stflife Happiness How happy are you happy Meaning in life I generally feel that what I do in my life is valuable and worthwhile. dngval Purpose in life To what extent do you feel that you have a sense of direction in your life? sedirlf Social wellbeing9 Local support To what extent do you feel that people in your local area help one another? pplahlp Social trust Most people can be trusted or you can't be too careful ppltrst Most people try to take advantage of you, or try to be fair pplfair Most of the time people helpful or mostly looking out for themselves pplhlp Reciprocity in social exchange To what extent do you receive help and support from people you are close to when you need? rehlppl Respect To what extent do you feel that people treat you with respect? trtrsp Discrimination Member of a group discriminated against in this country dscrgrp Safety Feeling of safety of walking alone in local area after dark aesfdrk Loneliness How much of the time during the past week did you feel lonely? fltlnl Notes: 1 To be replaced in R12 by: To what extent are you doing the things you really want and value in your life? (dgwtva) 2 To be replaced in R12 by: To what extent are you able to achieve your goals? (achgoal) 3 R12 core additions: How often do you chat to your neighbours, more than hello? (chatnb); How strongly do you feel you belong to your neighbourhood? (stblgnb); How close and connected do you feel to other people? (clcnppl). 4 To be replaced in R12 by: In difficult periods, I can usually find something good that helps me change for the better. (dprfgd) 5 Addition in R12: In difficult situations, how often are you able to take a pause without immediately reacting? (dstpwir) 6 The R6 “mental resources” “Engagement” and “Self-esteem” will not be repeated in R12. New concept: Self-compassion: When I’m going through a very hard time, I give myself the care and kindness I need. (kndhtm) 7 R12 will additionally measure compassion: When you hear about an acquaintance going through a difficult time, how much compassion do you usually feel for them? (acqcomp) 8 R12 new concept: “Inner harmony”: To what extent do you feel there is harmony in your life? (hrmlife). 9 R12 new concept: “Social harmony”: To what extent do you feel there is harmony among the people who live in [country]? (hrmcnt) and “General safety”: To what extent do you feel safe and secure in your life nowadays? (flsfsec) | 38 Sample restrictions. Given the project focus on wellbeing in European rural areas, we focus the analysis on participating countries from the European Economic Area (EEA). These are Belgium (BE), Bulgaria (BG), Cyprus (CY), Czechia (CZ), Germany (DE), Denmark (DK), Estonia (EE), Spain (ES), Finland (FI), France (FR), United Kingdom (GB) 21 , Hungary (HU), Ireland (IE), Iceland (IS), Italy (IT), Lithuania (LT), Netherlands (NL), Norway (NO), Poland (PL), Portugal (PT), Sweden (SE), Slovenia (SI), Slovakia (SK). 22 In addition to looking at these countries as a whole (dubbed here: EU overall), we will group countries according to the welfare state typology developed by Lauzadyte-Tutliene, Balezentis and Goculenko (2018) 23 , see Table 3-2. Table 3-2: Overview of countries and welfare state models Eastern Europe Central Europe Mediterranean Old European – not Nordic Old European – Nordic Bulgaria, Estonia, Lithuania Czechia, Poland, Slovakia, Slovenia, Hungary, East Germany1) Spain, Italy, Cyprus, Portugal Belgium, United Kingdom, Ireland2), Netherlands, France, West Germany1) Denmark, Norway, Finland, Sweden, Iceland2) Notes: 1) We differentiate between East and West Germany, treating East Germany as “Central European” (based on its socialist history and geographical proximity to the other Central EU countries) and West Germany as “Old European”. The settlement structures differ, such that we do, e.g., only observe seven respondents in East Germany who live in the countryside, while in West Germany, this figure amounts to ~7%. Moreover, life satisfaction is significantly lower in East Germany. 2) Ireland and Iceland were not included in the analysis that informed the construction of the typology (Lauzadyte-Tutliene, Balezentis and Goculenko, 2018). We added the countries to the “Old EU” group, based on the shared history and geographical proximity with Northern Ireland and Denmark, respectively. Analyses are restricted to those cases that have no missing information on any variables used in the empirical analysis. Table 3-3 provides an overview of the key dependent and independent variables used in this study, along with descriptive statistics. Our analysis proceeds as follows. To set the scene, we present some basic descriptions of the population living in Europe (specifically in rural settlements) and describe the patterns of subjective wellbeing. First, we focus on Europe as a whole, but we will then explore the usefulness of grouping countries by welfare state regime, before examining individual countries and comparing the wellbeing of the population in rural settlements to that in non-rural settlements, as well as the key correlates of subjective wellbeing. In more technical terms, we first present univariate and bivariate population statistics and then see if the associations hold when we account for additional heterogeneity in population characteristics using multivariate linear regression models. All analyses are conducted using the statistical data analysis software Stata (StataCorp 2021), and we use the svy suite of commands to consider the complex survey design of the ESS appropriately. 21 The United Kingdom was part of the EEA until 31st December 2019. 22 Put differently, we drop cases from the non-European countries Australia, Israel and Russia, as well as cases from Switzerland, Albania and Kosovo. This focus also addresses some analytical challenges: No sampling files have been supplied for Israel, and Russia includes so many cases that our results would be skewed towards the experiences of the Russian population. A survey implementation error means that we could not produce indicators for all well-being dimensions for Albania. 23 Esping-Andersen (2001)’s more prominent typology was not useful in our case, as the typology excludes the new Eastern and Central European welfare states that we observe in the ESS. We did, however, split off the Nordic states, identified by EspingAnderson as one group, from the otherwise very large “Old EU” group identified by Lauzadyte-Tutliene, Balezentis and Goculenko (2018). | 39 Table 3-3: Descriptive statistics of key variables used in the analysis (EU overall) Mean SD CV Min Max Wellbeing measure (MPI) Life satisfaction 99.6 9.499 0.095 70.3 112.2 Affective wellbeing 99.5 6.327 0.064 73.9 109.7 Basic psychological needs satisfaction 99.0 4.980 0.050 75.3 115.2 Mental resources 99.2 5.538 0.056 70.2 114.2 Social wellbeing 99.6 5.417 0.054 74.4 112.3 Evaluative wellbeing 99.2 7.514 0.076 65.2 112.9 Settlement type Cities 0.205 0.404 1.970 0 1 Suburbs 0.119 0.324 2.720 0 1 Towns 0.319 0.466 1.460 0 1 Villages 0.288 0.453 1.573 0 1 Countryside 0.069 0.253 3.680 0 1 Welfare state group Eastern European 0.146 0.353 2.415 0 1 Central European 0.198 0.398 2.014 0 1 Mediterranean 0.111 0.315 2.823 0 1 Old European - not Nordic 0.321 0.467 1.455 0 1 Old European – Nordic 0.224 0.417 1.864 0 1 Number of observations 29913 Notes: For descriptive statistics of all variables used in the analysis, see Appendix 4 (A3-1). Source: ESS Round 6 (2012/13, ESS6e02_6) with sample file supplement. 3.3. Results In 2012/2013, over a third of the EU population lived in a rural settlement, either in a village (32%) or on a farm or house in the countryside (4%), see Figure 3-2. The most prevalent settlement type is living in small or mid-sized towns (34%). There is, however, considerable heterogeneity across European countries. In Mediterranean countries (here: Italy, Portugal, Spain, and Cyprus), the rural population accounts for 50%, while in Eastern European countries (here: Bulgaria, Estonia, and Lithuania), the proportion living in cities is relatively high, at 40% (across all of Europe, the figure is 17%). In Nordic countries (i.e., Denmark, Norway, Finland, Sweden, and Iceland), the proportion of people living on farms or in the countryside was three times the European average, at 15% compared to 4%. The Old European – not Nordic versus Nordic states have the same proportions of the population living in urban and rural settlements (67% and 33%, respectively). Within the urban category, however, a significantly larger population share in the non-Nordic states compared to the Nordic states lives in towns (37% compared to 30%), whereas a considerably larger share of the population in the Nordic states compared to the non-Nordic states lives in suburbs (20% compared to 16%). Within the rural category, a significantly larger population share in the non-Nordic states compared to the Nordic states lives in villages (28% compared to 18%), whereas a considerably larger share of the population in the Nordic states compared to the non-Nordic states lives in the countryside (15% compared to 5%). | 40 Figure 3-2: Settlement structure across Europe and European welfare states (% of the population) Notes: Population estimates for participating countries in the European Economic Area. Source: ESS Round 6 (2012/13, ESS6e02_6) with sample file supplement. 3.3.1. Settlement structure and subjective wellbeing across Europe Table 3-4 reports the average subjective wellbeing scores for all countries and each welfare state group. By design, the average wellbeing scores in the EU amount to around 100 (i.e., the population average in the EU). Differentiation by welfare state shows that the Old European -not Nordic group, the most populous group in our analysis, most closely resembles the European average. The population in the Nordic countries show slightly above-average wellbeing scores across the board. By contrast, the population in the “Eastern European” countries scored significantly lower on life satisfaction than on any other subjective wellbeing dimension; the score is the lowest among all measured subjective wellbeing dimensions and welfare model groupings. 17% 12% 34% 32% 5% All countries 40% 2% 26% 31% 1% Eastern European welfare states 21% 6% 36% 36% 1% Central European welfare states 16% 6% 28% 46% 4% Mediterranean welfare states 16% 16% 37% 27% 4% Old EU welfare states (not Nordic) 17% 20% 30% 18% 15% Old EU welfare states (Nordic) | 47 [0.71, 2.09]). No other differences between city dwellers and those living in other settlement types are statistically significant in the fully adjusted models. 3.3.4. Country-specific variation in associations 3.3.4.1. Settlement types and wellbeing Overall, the results of the multivariate regressions show considerable heterogeneity in subjective wellbeing for the people living in different types of settlements. While the grouping of countries by welfare state group captures some cultural differences, we can still see a few statistically significant country effects (even in the models that group countries by welfare state). This suggests that this type of wellbeing analysis should ideally be undertaken also at a national level (and probably with a boost of cases in the least populated rural areas to draw more powerful conclusions). As a first stab at exploring whether there are country-specific differences in the dimensions of subjective wellbeing, we present in which settlement type a country’s average wellbeing score is the highest, separately for each dimension of subjective wellbeing (see Table 3-6). If there were a clear rural subjective wellbeing advantage—as suggested by the urban happiness paradox—we would expect the top half of the table to have no or significantly fewer entries than the bottom half of the table. Table 3-6 shows that the average wellbeing of rural dwellers (i.e., those living in villages or the countryside) is highest in about 50% of the countries and highest among urban dwellers in the other ~50% of countries. The only notable exception is social wellbeing, the highest among urban dwellers in only five out of 24 countries. There is a substantial overlap of results for life satisfaction and the broader evaluative wellbeing in all but six countries: Cyprus, Denmark, and Iceland have the highest evaluative wellbeing among city dwellers but the highest life satisfaction among town (Cyprus) and village residents (Denmark and Iceland); Hungary switches from highest life satisfaction in suburbs to highest evaluative wellbeing in villages, while Finland switches from highest life satisfaction in the countryside to highest evaluative wellbeing in towns and West Germany from highest life satisfaction in villages to highest evaluative wellbeing in the countryside. Table 3-6: Settlement structures with the highest average score in six dimensions of subjective wellbeing Most well-off settlement Life Satisfaction Affective wellbeing Basic psychol needs satisfaction Mental resources Social wellbeing Evaluative wellbeing Cities BG, LT, CZ, ES, IT BG, LT, D2, ES LT, D2, DK, FI, SE, IS BG, LT, PL, FI IT BG, LT, CZ, ES, IT, CY, DK, IS Suburbs EE, SK, SI, HU, D2 EE, CZ, SK, SI, HU, D1 EE, SK, PT EE, CZ, SK, HU, ES, CY, PT CZ, SK, ES, CY EE, SK, SI, D2 Towns CY CY SI, CY SI FI Villages PL, PT, D1, NO, DK, SE, IS PL, IT, PT, DK, SE BG, CZ, PL, HU, ES, IT, UK, NO D2, IT, DK, IS BG, LT, PL, SI, HU, D2, PT, D1, NO, DK PL, HU, PT, NO, SE Countryside BE, FR, UK, IE, NL, FI BE, FR, UK, IE, NL, NO, FI, IS BE, FR, D1, IE, NL BE, FR, D1, UK, IE, NL, NO, SE EE, BE, FR, UK, IE, NL, FI, SE, IS BE, FR, D1, UK, IE, NL Source: ESS Round 6 (2012/13, ESS6e02_6) with sample file supplement. For detailed results, see Appendix 7 (A3-4). Regarding the other dimensions of subjective wellbeing, we find considerable differences in which settlement type has the highest wellbeing in the 24 countries. For example, of the seven countries where life satisfaction is highest in villages, Poland, Portugal, Denmark and Sweden also have the highest affective wellbeing, only Poland and | 48 Norway have the highest basic psychological needs satisfaction, only Denmark and Iceland have the highest mental resources and Poland, Portugal, West Germany, Norway and Denmark the highest social wellbeing in this settlement type. By contrast, of the six countries where life satisfaction was highest in the countryside, in four countries (Belgium, France, the Netherlands and Ireland), all other dimensions of wellbeing were highest in this settlement type. For the UK, affective wellbeing was highest in villages and for Finland, basic psychological needs satisfaction and mental resources were highest in cities. These associations between settlement type and wellbeing are also visualised in Figure 3-5 to Figure 3-10. Countries are coloured according to their membership in welfare state groups; the coloured icons show which settlement type has the highest wellbeing in the respective dimension. (See also Appendix 3-3a to 3-3f for detailed results, including confidence intervals and settlement type rankings.) Figure 3-5: Settlement types with the highest life satisfaction in the EU Notes: Icons produced by Ricardo Ferrand. Maps produced by Anna Augstein (Thünen-Institute). Source: ESS Round 6 (2012/13, ESS6e02_6) with sample file supplement. | 49 Figure 3-6: Settlement types with the highest affective wellbeing in the EU Notes: Icons produced by Ricardo Ferrand. Maps produced by Anna Augstein (Thünen-Institute). Source: ESS Round 6 (2012/13, ESS6e02_6) with sample file supplement. | 50 Figure 3-7: Settlement types with the highest basic psychological needs satisfaction in the EU Notes: Icons produced by Ricardo Ferrand. Maps produced by Anna Augstein (Thünen-Institute). Source: ESS Round 6 (2012/13, ESS6e02_6) with sample file supplement. | 51 Figure 3-8: Settlement types with the highest mental resources in the EU Notes: Icons produced by Ricardo Ferrand. Maps produced by Anna Augstein (Thünen-Institute). Source: ESS Round 6 (2012/13, ESS6e02_6) with sample file supplement. | 52 Figure 3-9: Settlement types with the highest social wellbeing in the EU Notes: Icons produced by Ricardo Ferrand. Maps produced by Anna Augstein (Thünen-Institute). Source: ESS Round 6 (2012/13, ESS6e02_6) with sample file supplement. | 53 Figure 3-10: Settlement types with the highest evaluative wellbeing in the EU Notes: Icons produced by Ricardo Ferrand. Maps produced by Anna Augstein (Thünen-Institute). Source: ESS Round 6 (2012/13, ESS6e02_6) with sample file supplement. 3.3.4.2. Country-level wellbeing profiles For countries with sufficient cases in rural settlements, it is also possible to create country profiles of rural wellbeing. In the following, we present such a profile for two countries participating in the GRANULAR project, which have Living Labs with a specific focus on “wellbeing”: the UK and the Netherlands. We can see that the proportions of the population living in rural areas, as well as their demographics, vary across the two countries. We additionally regressed demographic characteristics on each dimension of wellbeing to see which factors are contributing the most to higher and lower wellbeing in these countries (marked by the three largest positive and the three largest negative b-coefficients observed in the multivariate regressions. For complete results, see Appendix 9 (A3-6_UK and A3-6_NL). For policymakers, these factors may be used to identify key groups of the population at risk of facing low wellbeing in rural areas. | 54 United Kingdom About one in five people in the UK (19%) live in rural settlements, with villages being the second most frequent settlement type following small or mid-sized towns, which are home to 49% of the people in the UK. Living in UK rural areas is associated with being separated or divorced (reference category: married) and with not being in education, not being permanently sick or disabled nor having a non-listed main economic activity status (such as being in the army or not working as in living off one’s wealth; reference category: in employment). Subjective wellbeing is considerably higher in the countryside and in villages than in urban settlements in all but the basic psychological needs satisfaction dimension. Subjective wellbeing by settlement type Being older, having good health, and a higher income stand out as the factors underpinning high rural wellbeing in the UK across all dimensions of subjective wellbeing. For basic psychological needs satisfaction, social wellbeing and evaluative wellbeing, having a degree or higher is beneficial. A key challenge to rural wellbeing across all dimensions of wellbeing is being unemployed, in education, looking after others, or being permanently sick or disabled. Those with children in the household experience reduced basic psychological needs satisfaction. Life satisfaction Affective wellbeing Psychological needs satisfaction Mental resources Social wellbeing Evaluative wellbeing + + + Aged 65 or older (ref. 15-24) No health limitations No health limitations Aged 55 to 64 (ref. 15-24) Having a degree or higher No health limitations + + No health limitations Household income: Top quintile (ref. Q1) Highest level of education: degree or higher Household income: Top quintile (ref. Q1) Aged 65 or older (ref. 15-24) Aged 65 or older (ref. 15-24) + Aged 55 to 64 (ref. 15-24) Household income: 4th quintile (ref. Q1) Household income: 3rd quintile (ref. Q1) No health limitations No health limitations Having a degree or higher - Economic activity: permanently sick or disabled (ref. employed) Economic activity: education (ref. employed) Economic activity: unemployed (ref. employed) Economic activity: unemployed (ref. employed) Economic activity: unemployed (ref. employed) Economic activity: other (ref. employed) - - Economic activity: unemployed (ref. employed) Economic activity: unemployed (ref. employed) Children in the household Economic activity: other (ref. employed) Economic activity: housework, looking after children, others (ref. employed) Economic activity: unemployed (ref. employed) - - - Economic activity: other (ref. employed) Economic activity: other (ref. employed) Economic activity: housework, looking after children, others (ref. employed) Economic activity: education (ref. employed) Economic activity: other (ref. employed) Economic activity: housework, looking after children, others (ref. employed) | 55 The Netherlands More than one in three people in the Netherlands (38%) live in rural settlements; one in four people on a farm or house in the countryside. Only 13% live, more than double that in small or mid-sized towns, giving the Netherlands an overall rural and small-town flair. Living in Dutch rural areas is associated with not being aged 25-54 (reference: aged 15-24), not being never married (reference category: married) and not having a degree. Subjective wellbeing is considerably higher in the countryside and in villages than in urban settlements in all but the basic psychological needs satisfaction dimension. While life satisfaction is relatively high, affective wellbeing and basic psychological needs satisfaction is relatively low. Having a higher income stands out as the single most important factor underpinning higher rural wellbeing in the Netherlands. Affective wellbeing is positively impacted by good health. Key challenges to rural wellbeing differ by wellbeing dimension. The youngest (aged 15-24) have higher life satisfaction and evaluative wellbeing than the other age groups. Those looking after others or being permanently sick or disabled have lower affective wellbeing, basic psychological needs satisfaction and social wellbeing. Widowhood and being in education are negatively associated with basic psychological needs satisfaction and social wellbeing in rural settlements. Life satisfaction Affective wellbeing Psychological needs satisfaction Mental resources Social wellbeing Evaluative wellbeing +++ Household income: Top quintile (ref. Q1) Household income: Top quintile (ref. Q1) Household income: Top quintile (ref. Q1) Household income: Top quintile (ref. Q1) Household income: Top quintile (ref. Q1) Household income: Top quintile (ref. Q1) ++ Household income: 4th quintile (ref. Q1) Household income: 4th quintile (ref. Q1) Household income: 4th quintile (ref. Q1) Household income: 4th quintile (ref. Q1) Household income: 4th quintile (ref. Q1) Household income: 4th quintile (ref. Q1) + Household income: 3rd quintile (ref. Q1) No health limitations Household income: 2nd quintile (ref. Q1) Household income: 2nd quintile (ref. Q1) Household income: 3rd quintile (ref. Q1) Household income: 2nd quintile (ref. Q1) - Aged 55 to 64 (ref. 15-24) Economic activity: housework, looking after children, others (ref. employed) Economic activity: housework, looking after children, others (ref. employed) Marital status: never married (ref. married) Economic activity: permanently sick or disabled (ref. employed) Aged 55 to 64 (ref. 15-24) - - Aged 25 to 54 (ref. 15-24) Marital status: widowed (ref. married) Economic activity: permanently sick or disabled (ref. employed) Aged 25 to 54 (ref. 15-24) Marital status: widowed (ref. married) Aged 65 or older (ref. 15-24) - - - Aged 65 or older (ref. 15-24) Economic activity: permanently sick or disabled (ref. employed) Economic activity: education (ref. employed) Marital status: widowed (ref. married) Economic activity: education (ref. employed) Aged 25 to 54 (ref. 15-24) Subjective wellbeing by settlement type | 56 3.4. Discussion A sizeable proportion of the EU population resides in rural settlements, with 32% living in villages and 4% on farms or in the countryside. The most common settlement type is small or mid-sized towns (34%), the middling category between cities and suburbs of cities and rural settlements. Rural living is particularly high in Mediterranean countries (50%) compared to Eastern European countries (40% urban population). In Nordic countries, the rural population living on farms or in the countryside is significantly higher than the EU average (15% vs. 4%). When we investigate how the subjective wellbeing of people living in different types of settlements differs, typically, we are left with drawing on relatively simple bivariate associations. Doing this to explore whether there are specific wellbeing profiles for different types of settlements, we found considerable heterogeneity in the subjective wellbeing profiles across the different countries and settlement types. In the Nordic countries, in all settlements, life satisfaction is higher than all other dimensions of subjective wellbeing. Social wellbeing and evaluative wellbeing are also higher than affective wellbeing, basic psychological needs satisfaction, and mental resources in all settlement types. In Eastern European countries, in villages, towns and cities, life satisfaction is lower than all other dimensions of subjective wellbeing. Evaluative wellbeing, too, is lower in these settlements than affective wellbeing, basic psychological needs satisfaction, mental resources and social wellbeing. In Central European and old European - not Nordic countries, the subjective wellbeing profiles differ only for those living in suburbs: Life satisfaction is higher than basic psychological needs satisfaction and social wellbeing in Central European suburbs, while the same is true also for mental resources in old European - not Nordic countries. No statistically significant differences were found in the different dimensions of subjective wellbeing across the Mediterranean countries. Bivariate associations may give a misleading impression, i.e., when unobserved characteristics that are linked to subjective wellbeing are also linked to where people live. To alleviate these concerns, we estimated a number of multivariate models that adjust for a comprehensive set of individual socio-economic and socio-demographic characteristics, as well as absorb country-level heterogeneity. We found that: ▪ Those living in suburbs or in the countryside show life satisfaction advantages over city dwellers, which are, however, explained by the characteristics of the population living in this settlement type. ▪ Countryside residents report higher affective wellbeing than city dwellers; this advantage is, however, explained by the characteristics of the population living in this settlement type. ▪ No significant differences were found overall in basic psychological needs satisfaction. ▪ Individual characteristics explain suburban residents' wellbeing advantage, while village and countryside residents have significantly higher mental resources than those in cities. ▪ Social wellbeing advantages exist for residents of all settlement types compared to city dwellers; the village and countryside advantage is robust to the inclusion of the characteristics of the population living in the respective settlement type. ▪ Both suburbanites and countryside residents report advantages over city dwellers in baseline models, but only the countryside advantage remains significant when controlling for observed population characteristics. In summary, living in villages and the countryside is generally associated with higher wellbeing, particularly in terms of mental resources and social wellbeing. However, many advantages observed in the unadjusted results diminish when accounting for individual and country-level characteristics. In some cases, however, the wellbeing advantages only come to the fore when we consider that different types of people tend to live in different types of settlements. In line with past research, we find that these advantages are most pronounced on the social wellbeing dimension as suggested by theories of strong social ties in rural communities. It remains to be seen with more recent data whether this social wellbeing advantage persists after the pandemic, which has led to increased concerns over isolation and loneliness (e.g., O’Sullivan, Burns et al. 2021, Ernst, Niederer et al. 2022). Moreover, mental resources appear to show a unique association with living on a farm or in the countryside. On the one hand, this could be related to scenic beauty and the mental wellbeing impact that these landscapes might offer. In the GRANULAR project, in Task 4.5, we explore models to predict scenic beauty of rural areas across the EU and such model-derived indicators can be used in conjunction with individual panel data such as SOEP (Germany) and Understanding Society (UK) to explore the different pathways through which scenic beauty might impact | 63 with different socio-economic conditions, housing markets, and consumer behaviours. This provides clear comparisons with our rural reference type. To absorb further unobserved heterogeneity in rural and urban neighbourhood contexts, we include the Townsend Area Deprivation Score and indicators from the index of Access to Healthy Assets and Hazards (AHAH). These are linked to UKHLS using the LSOA look-up file (University of Essex and Institute for Social and Economic Research 2024). The Townsend Score of Area Deprivation (2001, 2011, 2021) is a composite measure of material deprivation which was developed in the UK in the late 1980s (Townsend, Phillimore and Beattie 1988) following earlier conceptualisation of the link between deprivation and resource availability (Townsend 1979). Material deprivation is “a direct measure of poverty derived from the lack of items and activities deemed to be necessary for a minimum standard of living” (McKnight, Bucelli et al. 2024). Townsend Scores have been calculated for LSOA units based on a relatively simple analysis involving four indicators, which are calculable from Census datasets: the proportions of residents who are unemployed, and three indicators based on percentages of households which are a) overcrowded, b) not owner-occupied, and c) without access to a car. Standardised values of these indicators (following log transformation of the values of unemployment and overcrowding) are summed to form the Townsend Score, with higher values representing greater deprivation (Yousaf and Bonsall 2017). In this analysis, we calculated Townsend Scores for 2011 LSOAs in Great Britain (England, Scotland, Wales) for 2001, 2011 and 2021 using Census data for these three years and data from the 2022 Scottish Census. Changes in the boundaries and numbers of LSOAs between Censuses present a considerable challenge in calculating comparable variables over time. In our analysis, we used the distribution of postcodes (as of 2024) as a proxy for the population distribution, to re-allocate Census data from 2001 and 2021-22 to the 2011 LSOAs on a proportional basis. By doing this, we enhance the consistency and comparability of deprivation measurement. The LSOA look-up file was used to link the Townsend Scores to individual responses in the UKHLS datasets. Scores for 2001 were used for responses in 2010 or earlier; 2011 scores were used from 2011 to 2020, and 2021 scores were used from that year onwards. Calculations of the DEGURBA2 and Townsend Deprivation scores, described above, were carried out using R (R Core Team 2024) and the packages “sf” (Pebesma 2018, Pebesma and Bivand 2023), “dplyr” (Wickham, François et al. 2023), “terra” (Hijmans 2025), “exactextractr” (Baston 2023) and “mapview” (Appelhans, Detsch et al. 2023). The Access to Health Assets and Hazards (AHAH) (2017, 2019, 2022) is a multi-dimensional index developed by the CDRC for Great Britain, measuring how ‘healthy’ neighbourhoods are. The AHAH index combines indicators under four different domains of accessibility: ▪ Retail environment (access to fast food outlets, pubs, off-licences, tobacconists, gambling outlets), ▪ Health services (access to GPs, hospitals, pharmacies, dentists, leisure services), ▪ Physical environment (Blue Space, Green Space - Passive, Green Space - active), and ▪ Air quality (Nitrogen Dioxide, Particulate Matter 10, Sulphur Dioxide). We use the individual components of the index, not the composite index, seeing as we want to work out how these factors influence wellbeing differently in different types of rural and urban areas. Indicators for 2017 are used as proxies for the surveys conducted in 2009-2017, indicators for 2019 for the surveys conducted in 2018 and 2019, and indicators for 2022 for surveys conducted in 2020-2022. Due to changes in the availability of variables over time, our analysis does not include all variables in the full AHAH list given above; more specifically, access to tobacconists, off-licenses, hospitals, and active green space are not included in our analysis. | 64 For the Reader’s convenience, Table 4-2 provides an overview of the variables used in this analysis. Descriptive statistics of all variables used in the analysis are provided in Appendix 10 (A4-1). Table 4-2: Overview of variables used in this analysis Blocks of variables included in the regressions Rural-urban definition (R/U) RUC2011: Urban = greater or equal 10k population Rural = less than 10k population DEGURBA 2: Urban = Cities, Towns, Suburbs [1-3] Rural = Villages, Dispersed Rural Areas, Mostly uninhabited areas Area characteristics - Deprivation (D) Townsend Deprivation Score Individual and household characteristics (I) Age, Age2, Sex, Born in UK, Ethnicity [reference: White British], Employment status [reference: employed], Real equivalent household income, Highest qualification [reference: degree or higher], Number of children in household, Marital status [reference: single], Housing tenure [reference: owner-occupied] (other) Area characteristics (A) Access to the retail environment (three variables) Access to health services (four variables) Access to the physical environment (two variables) Air quality (three variables) Geodemographic characteristics (G) Output Area Classification 2011 [reference: Agricultural Communities] ACORN 2015 [reference: Farms and Cottages] 4.2.1. Empirical strategy We adopt a standard micro-economic life satisfaction model that assumes that exogenous individual and neighbourhood characteristics have a direct impact on life satisfaction: 𝑌𝑖𝑡 =∝ +𝛽′𝑋𝑖𝑡 + 𝛾′𝑁𝑗𝑖𝑡 + 𝜀𝑖𝑡 (1) where 𝑖 denotes individuals, 𝑗 neighbourhoods, and 𝑡 time. Individual wellbeing (𝑌𝑖𝑡) is a function of individual characteristics (𝑋𝑖𝑡) and neighbourhood characteristics (𝑁𝑗𝑖𝑡) that have been shown to influence wellbeing, and the error (𝜀𝑖𝑡). This model is implemented using the pooled ordinary least squares (OLS) estimator. As we are working with panel data, all standard errors are adjusted for clustering on individuals and for heteroscedasticity. All models are estimated using the “xtreg” command in Stata 15. In the first stage, we include only the rural-urban dummy to see how large the coefficient on the rural wellbeing advantage is. Next, we add blocks of controls that may explain the rural wellbeing advantage separately. This allows us to see which (block of) variable(s) attenuate the rural advantage the most. Is it individual and household characteristics or the characteristics of places and the local community? All models are nested so that we can compare the same cases across specifications. Last but not least, we include the rural dummy, geodemographic classification, Townsend Score and Access to Healthy Assets and Hazards variables simultaneously to see which factors stand out. 4.3. Results 4.3.1. Evolution of life satisfaction over time by rurality To set the scene, we will explore how life satisfaction changed from 2009 to 2023, focusing on the differences between rural and urban areas. Figure 4-1 illustrates that life satisfaction is consistently higher in rural areas than in urban areas across various rural–urban classifications. This rural advantage in wellbeing continues over time, although its extent varies. Both rural and urban areas display a recurring cyclical pattern in life satisfaction between 2009 and 2023, rather than following a linear trend. These patterns demonstrate simultaneous rises and declines in both classifications. Figure 4-1: Mean life satisfaction by rurality (with 95% CI), Great Britain 2009-2023 A) Rural-urban Classification (RUC 2011) B) Degree of Urbanisation (DEGURBA2) | 65 Source: Understanding Society (v20) linked with special licence geodata (v14). Weighted population estimates. Figure 4-2: Mean life satisfaction by rural area type (DEGURBA2, with 95% CI), Great Britain 2009-2023 Panel A) Rural clusters Panel B) Low-density rural areas C) Very low-density areas Source: Understanding Society (v20) linked with special licence geodata (v14). Weighted population estimates. While rural wellbeing remains higher overall, urban areas exhibit greater volatility, especially after 2020, which may be linked to the effects of COVID-19. Both rural-urban classification systems convey the same overall narrative: from 2009 to 2023, the population residing in rural areas of Great Britain reports higher wellbeing compared to those in urban areas. Despite fluctuations from year to year, this pattern remains constant, with urban wellbeing consistently | 66 lagging behind that of rural regions. The RUC 2011 measure demonstrates a slightly clearer and more persistent rural wellbeing advantage than the DEGURBA2 classification. For the DEGURBA2 classification specifically, we can examine whether there are any variations in patterns over time among different rural area types – i.e., rural clusters, low-density rural areas, and very low-density rural areas. Figure 4-2 shows that, across all rural categories, life satisfaction is higher in rural areas than in urban areas. The rural wellbeing advantage is most stable in rural clusters and low-density areas. In low-density rural areas, after 2015, the rural advantage becomes less consistent, but on average, wellbeing remains higher in low-density areas than in urban areas. By contrast, results for very low-density areas are more variable due to wider confidence intervals (due to low sample sizes). Both rural and urban wellbeing exhibit cyclical fluctuations, with notable dips in the mid-2010s and following the 2020 period. 4.3.2. Factors explaining the rural wellbeing advantage Next, we explore which characteristics of rural (and urban) areas might explain the rural wellbeing advantage. Table 4-3 shows the coefficient of the rural definition (based on the RUC and DEGURBA classifications) within a series of 12 regressions of life satisfaction. The two regressions on the top of the table include the rural definition alone, and the remainder include the rural definition plus one other block of variables (described in Table 4-2). Table 4-3: Random-effects generalised least squares (GLS) regression of life satisfaction on individual blocks of context variables: Rural-urban coefficient RUC 2011 DEGURBA2 R/U only b-coef. 0.148 *** b-coef. 0.144 *** S.E. 0.009 S.E. 0.011 R2 0.003 R2 0.002 +OAC 2011 b-coef. 0.075 *** b-coef. 0.062 *** S.E. 0.011 S.E. 0.014 R2 0.016 R2 0.016 + ACORN 2015 b-coef. 0.058 *** b-coef. 0.039 ** S.E. 0.009 S.E. 0.013 R2 0.02 R2 0.02 + Area deprivation b-coef. 0.054 *** b-coef. 0.051 *** S.E. 0.009 S.E. 0.012 R2 0.016 R2 0.016 + AHAH b-coef. 0.088 *** b-coef. 0.072 *** S.E. 0.011 S.E. 0.013 R2 0.004 R2 0.004 + individual characteristics b-coef. 0.087 *** b-coef. 0.088 *** S.E. 0.009 S.E. 0.011 R2 0.067 R2 0.067 Notes: For complete model results, see Appendix 11 (A4-2). Source: Understanding Society (v20) linked with special licence geodata (v14). Supporting the observations of a rural advantage in Section 4.3.1 above, we find that rurality is significantly and positively associated with life satisfaction in all analyses (models with rural/urban definition only included: βRUC = 0.148, p<0.001; βDEGURBA = 0.144, p<0.001). However, this association is reduced when individualand householdlevel characteristics (βRUC = 0.087, p<0.001; βDEGURBA = 0.088, p<0.001) or contextual characteristics are introduced, with the largest attenuation of the rurality coefficient found when area deprivation (RUC: βRUC = 0.054, p<0.001) and the ACORN classification (DEGURBA: βDEGURBA = 0.039, p<0.005) are included. Comparing the impact of adding the OAC or the ACORN classification to the model, we find that the ACORN typology seems to include more of the factors linked to the rural wellbeing advantage (thus, attenuating the rural-urban coefficient more) than the OAC. Furthermore, it is noteworthy that adding in the accessibility and air pollution data explains about as much of the rural wellbeing advantage (i.e., reducing the coefficient to βRUC = 0.088, p<0.001 and βDEGURBA = 0.072, p<0.001, respectively) as when we include individual-level controls (albeit, the latter models explain more of the overall variance in life satisfaction). | 67 Next, we assessed four regressions of life satisfaction on variables in all five blocks of variables, substituting both definitions of rurality (RUC, see Table 4-4; DEGURBA, see Table 4-5) and the two geodemographic classifications 25 . Table 4-4: Random-effects generalised least squares (GLS) regression of life satisfaction on rurality, area and individual characteristics by rural-urban indicator (geodemographic classification = OAC 2011), GB 2009-2023 RUC 2011 DEGURBA b-coef. b-coef. Rural-urban indicator (rural=1) 0.055 *** 0.055 *** OAC 2011 (ref. Agricultural communities) 1A1: Rural Workers and Families -0.04 -0.039 1A2: Established Farming Communities -0.045 -0.039 1A4: Older Farming Communities -0.015 -0.008 1B1: Rural Life -0.084 * -0.078 * 1B2: Rural White-Collar Workers -0.043 -0.037 1B3: Ageing Rural Flat Tenants -0.1 * -0.098 * 1C1: Rural Employment and Retirees -0.01 -0.007 1C2: Renting Rural Retirement -0.066 -0.059 1C3: Detached Rural Retirement 0.063 0.073 2: Cosmopolitans -0.002 -0.002 3: Ethnicity Central -0.034 -0.032 4: Multicultural Metropolitans -0.118 ** -0.118 ** 5: Urbanites -0.058 -0.056 6: Suburbanites -0.048 -0.046 7: Constrained City Dwellers -0.132 *** -0.132 *** 8: Hard-Pressed Living -0.143 *** -0.141 *** Other area characteristics: Townsend Score of Area Deprivation Distance to … -0.04 -0.039 … general practitioners -0.045 -0.039 … dentists -0.015 -0.008 … pharmacies -0.084 * -0.078 * … gambling sites -0.043 -0.037 … fast food restaurants -0.1 * -0.098 * … pubs -0.01 -0.007 … leisure facilities -0.066 -0.059 … blue space 0.063 0.073 … green space -0.002 -0.002 Air pollution -0.034 -0.032 … NO2 -0.118 ** -0.118 ** … PM10 -0.058 -0.056 … SO2 -0.048 -0.046 Individual and household characteristics Yes Yes Constant 6.139 *** 6.139 *** Number of observations 428880 428880 Number of individuals 76908 76908 Rho 0 0.381 R-squared (within) 0.004 0.004 R-squared (between) 0.093 0.093 R-squared (overall) 0.069 0.069 Notes: For complete model results, see Appendix 11 (A4-2). Source: Understanding Society (v20) linked with special licence geodata (v14). 25 The geodemographic classifications could not be analysed simultaneously, due to special licence data access restrictions. | 68 Table 4-5: Random-effects generalised least squares (GLS) regression of life satisfaction on rurality, area and individual characteristics by rural-urban indicator (geodemographic classification = ACORN 2015), GB 2009-2023 RUC 2011 DEGURBA b-coef. b-coef. Rural-urban indicator (rural=1) 0.046 *** 0.045 *** ACORN 2015 (ref. Farms and cottages) 1: Affluent Achievers -0.058 * -0.054 * 2: Rising Prosperity -0.122 *** -0.118 *** 3F22: Larger families in rural areas -0.114 ** -0.109 ** 3F23: Owner occupiers in small towns and villages -0.11 *** -0.103 ** 3G: Successful Suburbs -0.132 *** -0.128 *** 3H: Steady Neighbourhoods -0.162 *** -0.161 *** 3I: Comfortable Seniors -0.15 *** -0.147 *** 3J: Starting Out -0.125 *** -0.125 *** 4: Financially Stretched -0.184 *** -0.181 *** 5: Urban Adversity -0.209 *** -0.207 *** 6: Not Private Households -0.054 -0.051 Other area characteristics: Townsend Score of Area Deprivation -0.008 *** -0.008 *** Distance to … … general practitioners -0.005 ** -0.005 ** … dentists -0.004 ** -0.004 ** … pharmacies 0.002 0.002 … gambling sites 0.000 0.000 … fast food restaurants 0.002 * 0.003 * … pubs -0.001 -0.001 … leisure facilities 0.000 0.000 … blue space -0.002 * -0.002 * … green space -0.002 -0.002 Air pollution … NO2 0.001 0.000 … PM10 -0.005 * -0.005 * … SO2 -0.031 ** -0.031 ** Individual and household characteristics Yes Yes Constant 6.219 *** 6.219 *** Number of observations 428880 428880 Number of individuals 76908 76908 Rho 0.38 0.381 R-squared (within) 0.004 0.004 R-squared (between) 0.093 0.093 R-squared (overall) 0.069 0.069 Notes: For complete model results, see Appendix 11 (A4-2). Source: Understanding Society (v20) linked with special licence geodata (v14). Firstly, we note that the rural wellbeing advantage reduces to about 32-38% of its original size (OAC: βRUC and βDEGURBA = 0.055, p< 0.001; ACORN βRUC = 0.046, p< 0.001; βDEGURBA = 0.045, p< 0.005) when we include all five blocks of individual and neighbourhood context variables simultaneously. The “black box” rural advantage in life satisfaction remains evident in these models, controlling for a vast range of contextual factors. Two-thirds of the raw rural advantage in life satisfaction can be attributed to individual and neighbourhood characteristics, which are also associated with life satisfaction themselves. Specifically, this analysis found evidence of significant associations between neighbourhood characteristics - measured using geodemographic typologies - and life satisfaction. Considering firstly the analyses which include the OAC typology (see Table 4-4), we find that reported life satisfaction in the “Agricultural Communities” reference category was relatively higher than in most neighbourhood types, although few of these differences (five out of 16 categories for both rural-urban classification measures) were significant at the 95% confidence level (in support of H4). The results for the different rural-urban | 69 classifications vary insignificantly; for parsimony, we will refer to only the results using the RUC (which is the more frequently used classification for GB). Residents within “Hard-Pressed Living” neighbourhoods reported the lowest life satisfaction compared with those in “Agricultural Communities” (RUC: βHard-Pressed Living = -0.143, p < 0.001), which is in support of H7. Individuals living in “Constrained City Dwellers” neighbourhoods, which are also associated with urban economic disadvantage, also reported significantly lower life satisfaction than the reference (RUC: βConstrained City Dwellers = -0.132, p < 0.001). Meanwhile, there was no significant difference in life satisfaction between the “Urbanites” and “Cosmopolitans” neighbourhood types and the “Agricultural Communities” areas (H5), or between “Suburbanites” and “Agricultural Communities” (H6). Separately, it is notable that residents in “Multicultural Metropolitans” neighbourhoods, reflecting lower-income areas associated with overseas migrants and mixed ethnic backgrounds (Office for National Statistics 2015), show significantly lower life satisfaction compared with “Agricultural Communities” (RUC: βMulticultural Metropolitans = -0.118, p < 0.01). While residents in all rural neighbourhood types (OAC group 1) reported lower life satisfaction than those in agricultural communities, only the differences with “Rural Life” and “Ageing Rural Tenants” reach statistical significance. Considering the ACORN classification (see Table 4-5), we find that living in areas classified as “Farms and cottages” was positively associated with life satisfaction (H1), as coefficients for all other neighbourhood types were negative, and statistically significant except for “Not Private Households” (signifying communal dwellings). Again, the results for the different rural-urban classifications vary insignificantly and we will refer to only the results using the RUC definition. The “Affluent Achievers” areas, which signify wealthy suburban neighbourhoods, report marginally (but significantly) lower life satisfaction compared with the “Farms and cottages” reference category (RUC: βAffluent Achievers = -0.058, p < 0.05). Other categories reflecting suburban affluence (“Rising Prosperity”, “Successful Suburbs”), too, report significantly lower life satisfaction, albeit at a higher level of statistical confidence (e.g., βRising Prosperity = -0.122, p < 0.001) (H2). There is a much larger difference in life satisfaction between areas of lower urban incomes and deprivation (“Financially Stretched” and “Urban Adversity”) and the “Farms and cottages”, shown by the relatively large negative coefficients (e.g., RUC βUrban Adversity = -0.209, p < 0.001) (H3). Residents in both rural neighbourhood types (“Larger families in rural areas” and “Owner-occupiers in small towns and villages”) have lower life satisfaction than those on farms and in cottages, with both differences reaching statistical significance. Evaluation of the effects of other area characteristics revealed further associations between the characteristics of local areas and life satisfaction. Increasing area deprivation, measured using the Townsend Deprivation Score, was significantly negatively associated with life satisfaction in all four regressions (H8). Nuanced contextual effects emerge when access to retail and health services are considered. Access to GP practices and dentists were significantly negatively associated with life satisfaction, but this was not the case for pharmacists (no significant associations found) (H9). Distance to leisure facilities, gambling establishments and pubs were not significantly associated with life satisfaction, but there was a consistent positive association between fast food outlets and life satisfaction (H10). Similarly, some – but not all – types of air pollution (particulate matter and Sulphur dioxide, but not nitrogen dioxide) were significantly, negatively associated with life satisfaction (H11). Surprisingly, proximity to (passive) green and blue space were not associated with greater life satisfaction (H12). Following this, we compare the contextual predictors of life satisfaction in the rural and urban samples, to identify potential drivers of the rural advantage in wellbeing. Results for the OAC are shown in Table 4-6; results for the models using ACORN are shown in Table 4-7. Unlike the models based on the whole sample, we can see that the associations of neighbourhood characteristics with life satisfaction vary considerably across models, both within rural areas when a different rural-urban classification is used and when comparing rural and urban residents (i.e., holding the rural-urban definition constant). 26 26 Differences in results across the two rural-urban classifications may be driven by differences in which sample members are classified as rural (and urban). The rural sample is considerably smaller when we use DEGURBA2 (N = 9801) than when we use RUC (N = 17843), meaning we have more statistical power to detect rural area effects using the RUC. Differences in results across the two geodemographic classifications are trickier to interpret. The OAC geodemographic classification follows more closely the rural-urban definition of the RUC: it considers rurality as one of the key stratifying variables, putting almost all areas with a | 70 Table 4-6: Random-effects generalised least squares (GLS) regression of life satisfaction on area and individual characteristics by rural-urban indicator (geodemographic classification = OAC 2011), GB 2009-2023 RUC 2011 DEGURBA2 Rural area Urban area Rural area Urban area b-coef. b-coef. b-coef. b-coef. OAC 2011 (ref. Agricultural communities) 1A1: Rural Workers and Families -0.046 0.547 * -0.029 -0.181 1A2: Established Farming Communities -0.016 -0.061 -0.027 -0.144 1A4: Older Farming Communities 0.022 0.239 0.025 -0.154 1B1: Rural Life -0.073 0.280 -0.050 -0.263 ** 1B2: Rural White-Collar Workers -0.029 0.292 -0.007 -0.185 * 1B3: Ageing Rural Flat Tenants -0.102 * 0.225 -0.137 ** -0.204 * 1C1: Rural Employment and Retirees 0.027 0.556 * 0.059 -0.270 1C2: Renting Rural Retirement -0.101 0.289 -0.098 -0.200 * 1C3: Detached Rural Retirement 0.127 * 0.304 0.171 * -0.149 2: Cosmopolitans -0.471 0.294 1.118 ** -0.185 * 3: Ethnicity Central 1.276 *** 0.258 0.000 -0.217 ** 4: Multicultural Metropolitans -0.239 0.182 -0.005 -0.300 *** 5: Urbanites -0.042 0.253 -0.021 -0.236 ** 6: Suburbanites -0.043 0.268 -0.009 -0.222 ** 7: Constrained City Dwellers -0.033 0.159 0.073 -0.317 *** 8: Hard-Pressed Living -0.137 *** 0.163 -0.129 ** -0.321 *** Other area characteristics: Townsend Score of Area Deprivation -0.008 -0.012 *** -0.002 -0.013 *** Distance to … … general practitioners -0.001 -0.008 * 0.000 -0.006 * … dentists -0.003 * -0.010 ** -0.003 * -0.008 ** … pharmacies 0.001 0.005 0.001 0.004 … gambling sites 0.000 0.000 0.000 0.000 … fast food restaurants 0.001 0.006 * 0.001 0.004 * … pubs -0.001 -0.005 0.000 -0.006 ** … leisure facilities 0.000 -0.002 * 0.000 0.000 … blue space 0.000 0.000 0.000 -0.001 … green space -0.020 0.000 -0.010 -0.001 Air pollution … NO2 -0.009 * 0.001 -0.012 0.000 … PM10 -0.002 -0.005 -0.007 -0.004 … SO2 0.062 * -0.034 ** 0.074 -0.031 ** Individual and household characteristics Yes Yes Yes Yes Constant 5.776 *** 5.879 *** 5.784 *** 6.330 *** Number of observations 98885 329995 51952 376928 Number of individuals 17843 62486 9801 69443 Rho 0.368 0.385 0.372 0.382 R-squared (within) 0.002 0.004 0.003 0.004 R-squared (between) 0.084 0.091 0.079 0.092 R-squared (overall) 0.056 0.071 0.053 0.07 Notes: For complete model results, see Appendix 12 (A4-3). Source: Understanding Society (v20) linked with special licence geodata (v14). population of less than 10k into the “Countryside communities” group and then differentiating between community types withingroup. Thus, the rural RUC sample will have very few urban cases and vice versa. The ACORN typology, by contrast, is not stratified by degree of urbanisation and cases can straddle more easily both rural-urban definitions. Due to resource constraints, we could not experiment with recoding geodemographic types straddling the rural-urban categories (yet) and unfortunately, we cannot compare the two geodemographics simultaneously due to restrictions on the number and types of data that can be linked under special licence arrangement. | 71 It is notable that area deprivation has significant negative associations with life satisfaction for urban residents, only, and that this pattern is present for all four rural-urban comparisons (i.e., the combinations of the RUC and DEGURBA2 definitions of rurality, and the OAC and ACORN geodemographic typologies). In terms of access to health facilities, greater distance to GPs is significantly negatively associated with life satisfaction in all urban regressions, but this is not the case for any of the rural analyses. By contrast, greater distance to dentists significantly reduces life satisfaction in all models: rural and urban. There is also evidence of a greater sensitivity of life satisfaction to service access within urban areas, as significant associations for distance to fast food restaurants, pubs and leisure facilities are only found in the urban sample. The direction of these urban associations was always positive for fast food restaurants and always negative for pubs and leisure facilities. Table 4-7: Random-effects generalised least squares (GLS) regression of life satisfaction on area and individual characteristics by rural-urban indicator (geodemographic classification = ACORN 2015), GB 2009-2023 RUC 2011 DEGURBA2 Rural area Urban area Rural area Urban area b-coef. b-coef. b-coef. b-coef. ACORN 2015 (ref. Farms and cottages) 1: Affluent Achievers -0.049 -0.028 -0.035 -0.121 2: Rising Prosperity -0.123 ** -0.093 -0.156 * -0.185 ** 3F22: Larger families in rural areas -0.122 ** -0.102 -0.106 * -0.177 * 3F23: Owner occupiers in small towns and villages -0.116 *** -0.067 -0.106 ** -0.150 * 3G: Successful Suburbs -0.103 ** -0.107 -0.067 -0.202 ** 3H: Steady Neighbourhoods -0.146 ** -0.135 -0.219 ** -0.229 ** 3I: Comfortable Seniors -0.146 ** -0.116 -0.194 -0.213 ** 3J: Starting Out -0.232 ** -0.096 0.161 -0.196 ** 4: Financially Stretched -0.196 *** -0.156 -0.195 *** -0.251 *** 5: Urban Adversity -0.213 *** -0.178 -0.262 ** -0.273 *** 6: Not Private Households 0.067 -0.087 0.145 -0.145 Other area characteristics: Townsend Score of Area Deprivation -0.004 -0.010 *** 0.004 -0.009 *** Distance to … … general practitioners -0.001 -0.008 * 0.000 -0.006 * … dentists -0.003 * -0.010 ** -0.003 * -0.008 ** … pharmacies 0.001 0.005 0.000 0.004 … gambling sites 0.000 0.000 0.000 0.000 … fast food restaurants 0.001 0.006 0.001 0.004 * … pubs 0.000 -0.006 0.000 -0.006 ** … leisure facilities 0.000 -0.002 * 0.000 0.000 … blue space -0.001 0.000 0.000 -0.001 … green space -0.017 -0.002 -0.001 -0.003 Air pollution … NO2 -0.009 * 0.002 -0.010 0.001 … PM10 -0.002 -0.005 -0.007 -0.004 … SO2 0.061 * -0.036 *** 0.067 -0.034 *** Individual and household characteristics Yes Yes Yes Yes Constant 5.849 *** 6.236 *** 5.859 *** 6.300 *** Number of observations 98885 329995 51952 376928 Number of individuals 17843 62486 9801 69443 Rho 0.368 0.385 0.373 0.382 R-squared (within) 0.002 0.004 0.003 0.004 R-squared (between) 0.085 0.091 0.078 0.093 R-squared (overall) 0.056 0.071 0.053 0.07 Notes: For complete model results, see Appendix 12 (A4-3). Source: Understanding Society (v20) linked with special licence geodata (v14). | 72 We also observe distinct results for variables reflecting forms of air pollution: increased sulphur dioxide (SO2) pollution significantly reduces life satisfaction in all urban analyses, but this is not the case in any of the rural samples. In fact, unusually, SO2 has a significant positive effect on life satisfaction in two rural models. As higher SO2 levels in rural areas are linked to biomass or coal burning, agriculture, and small-scale industries, combined with transport of pollutants from elsewhere, this could indicate living in rural areas with greater economic activity. Conversely, nitrogen dioxide (NO2) pollution is negatively associated with life satisfaction in all rural samples (twice significantly), but this is not the case in any urban samples. In rural areas, background air is usually cleaner, so when NO₂ levels rise, the effects are more noticeable (smell, haze, irritation) and may have a stronger impact on wellbeing. In urban areas, by contrast, high NO2 is normal. Therefore, our analysis finds support for H13, but not H14. Table 4-8: Summary of hypotheses and results Lastly, due to our selection of a rural reference category (reflecting farming communities) in the two geodemographic classifications, we restrict our interpretation of associations between the geodemographic neighbourhood types and life satisfaction to rural respondents only. Considering the OAC classification, we find that places associated with “Ageing Rural Flat Tenants” report significantly lower life satisfaction than “Agricultural communities” for both rural/urban definitions (RUC: βAgeing Rural Flat Tenants = -0.102, p < 0.05; DEGURBA 2: βAgeing Rural Flat Tenants = -0.137, p < 0.05), while “Detached Rural Retirement” has relatively higher life satisfaction (RUC: βDetached Rural Retirement = 0.127, p Hypotheses Hypothesis supported? ACORN 2015 related: H1: Residents in “Farms and Cottages” areas report higher life satisfaction than those in most other ACORN categories, due to greater environmental quality and social cohesion. yes H2: Affluent suburban categories (e.g., “Affluent Achievers”) may report life satisfaction levels similar to or higher than “Farms and Cottages”, reflecting economic security and service access. no H3: Deprived urban categories (e.g., “Urban Adversity”) are expected to report significantly lower life satisfaction than “Farms and Cottages.” yes OAC 2011 related: H4: Residents in “Agricultural Communities” report higher life satisfaction than those in most other OAC groups, reflecting strong community ties and environmental advantages. yes H5: Residents in “Urbanities” and “Cosmopolitans” are expected to report lower life satisfaction compared to “Agricultural Communities” due to higher exposure to deprivation and pollution. no H6: Residents in “Suburbanites” may show similar or higher life satisfaction compared to “Agricultural Communities,” given their socioeconomic advantages and good service access. no H7: Residents in “Hard-Pressed Living” areas are expected to report the lowest life satisfaction relative to “Agricultural Communities.” yes Other neighbourhood-related: H8: Higher area deprivation (Townsend score) is associated with lower life satisfaction. yes H9: Greater distances to health services are associated with lower life satisfaction. partly (not pharmacies) H10: Proximity to leisure facilities is associated with higher life satisfaction, while accessibility to unhealthy outlets (fast food outlets, pubs and gambling outlets) is associated with lower life satisfaction. no H11: Higher levels of air pollution (NO₂, PM₁₀, SO₂) are associated with lower life satisfaction. partly (not NO2) H12: Proximity to blue space and passive green space is associated with higher life satisfaction. no For rural and urban areas specifically, we hypothesise: H13: In rural areas, deprivation, access to services and unhealthy outlets have weaker associations with life satisfaction than they do in urban areas, reflecting adaptation to lower accessibility and lower prevalence. yes H14: In both rural and urban areas, access to leisure facilities, green and blue space, and lower air pollution are positively associated with life satisfaction. no | 79 observed. Individuals who are in any treatment group at a later stage in the survey are excluded from the control group at earlier stages in the study. See Appendix 13 (A5-1) for a description of all samples with respect to the respondents’ socio-demographic and socio-economic characteristics (age, gender, education, employment status, marital status, personal earned income, net household income) and characteristics of their living environment (type of residential area, type of building, degree of rurality, unemployment rate in the district (BBSR 2022), proportion of road traffic area in the total area of the municipality (Leibniz Institute of Ecological Urban and Regional Development 2023), which have an explanatory power for differences in health-related quality of life, see e.g. Ellert und Kurth (2013). 5.2.5. Empirical model To examine whether there is a causal relationship between health-related quality of life and the presence of wind turbines in the immediate residential area, we use a so-called difference-in-differences (DiD) design. In this quasiexperimental design, the difference in the health-related quality of life of people in the treatment and control groups is compared before (pre-treatment) and after the introduction of the intervention (post-treatment). Suppose the introduction of wind turbines in the residential environment has an effect. In that case, there is a difference in wellbeing between the treatment and control groups after the introduction of the intervention, which differs from the difference between the control and treatment groups before the introduction of wind turbines in the residential environment. The DiD estimator 𝛿 can be summarised as follows: 𝛿 = (𝑦𝑇1− 𝑦𝐶1)−(𝑦𝑇0− 𝑦𝐶0) (1) with 𝑦𝑇1 the wellbeing of treatment group T after the intervention (t1), 𝑦𝑇0 the wellbeing of treatment group T before the intervention (t0), 𝑦𝐶1 the wellbeing of the control group C after the intervention and 𝑦𝐶0 the wellbeing of control group C before the intervention. The design requires that the treatment and control groups differ only in the treatment. Some of our analytical samples have statistically significant differences regarding the local area characteristics (mostly at smaller radii), however. We, therefore, use the propensity score matching method, matching treatment and control subjects based on the observed characteristics at time t0 (i.e., before treatment). The covariates were selected to explain differences in health-related quality of life between the treatment and control groups and do not influence the probability that a wind turbine will be built within a 6 km radius. As individual-level controls, we use sex, age group, whether the respondent had any overnight stays at the hospital in the previous calendar year, and the number of doctoral appointments attended in the three calendar months preceding the interview at the last wave. As area-level controls, we use the degree of rurality, whether the respondent lives in a residential area, and the proportion of the area (at municipality scale) used for road traffic. We use the statistical analysis software Stata v18 for the analysis. Propensity score matching is implemented as 1:1 nearest neighbour matching using the psmatch2 command. Balance on covariates is assessed using Rubin's B (the absolute standardised difference of the means of the linear index of the propensity score in the treated and (matched) non-treated group) and Rubin's R (the ratio of treated to (matched) non-treated variances of the propensity score index). Rubin (2001) recommends that B be less than 25 and that R be between 0.5 and 2 for the samples to be considered sufficiently balanced. Except when the treatment is wind turbines with at least 100 m hub height in a 2 km radius, all matched samples are balanced on covariates (see Appendix 14 (A5-2)). For estimation, we use a linear model and the fixed effects estimator. Regression equation (2) is as follows: 𝑦𝑖𝑡 = 𝛼𝑖+ 𝜆𝑡 + 𝛽1𝑊𝑇𝑖𝑡 +𝜀𝑖𝑡 (2) where 𝑦 is the health-related quality of life of respondent i at survey time t, 𝛼𝑖 captures all time-invariant individual traits (i.e., the individual fixed effect), and 𝜆𝑡 is the period effect. 𝛽1 is the difference-in-differences (average treatment) effect, and 𝜀 is the idiosyncratic error term. Note that we focus our analysis on individuals who did not move houses between t0 and t1; hence, 𝛼𝑖 also captures any neighbourhood fixed effects (such as the relatively stable distances to local amenities such as hospitals). 𝑊𝑇𝑖𝑡 is a dummy variable that takes the value 1 in period 𝑡 if a wind turbine is present within the treatment radius 𝑟 around the household of individual 𝑖 , else nought. | 80 In further specifications, we consider measures of the intensity of the treatment. For this, we multiply the treatment dummy by a) the number of installed wind turbines and b) the inverse distance to the closest installed wind turbine in the respective treatment radius. 5.3. Results Figure 5-2 illustrates the expansion of wind turbines in Germany from 1983 to the end of 2020 using kernel density plots. The black spots in the map represent the exact location of wind turbines present at any time during the depicted period; the heat map indicates where more wind turbines have been built compared to the respective previous period. Figure 5-2: Expansion of wind turbines in Germany from 1983 to end of 2020 Source: Core Energy Market Data Register (MaStR, version of 31.03.2023) with own corrections. The first 3,345 onshore wind turbines were constructed between 1983 and 1999, predominantly along the shorelines and within the federal states of Lower Saxony, Schleswig-Holstein, and Mecklenburg-Vorpommern in northern Germany. The second wave of onshore wind turbine construction, in the period 2000 to 2009, saw the construction of 12,931 wind turbines, this time spreading out into the interior of the aforementioned states as well as the Eastern German states of Thuringia and Saxony-Anhalt. In the third period of wind turbine construction, from 2010 to 2020, a total of 11,701 new wind turbines were constructed across all German states (except Berlin), with the regional focus being in Hesse and Rhineland-Palatinate, in the South West of Germany. Figure 5-3 presents the proportion of households located within 6 km to 1.5 km of a wind turbine for two periods— 2002–2010 and 2012–2020—disaggregated using the Thünen typology of rural areas. The Thünen Typology of Rural Areas is a classification system developed by the Thünen Institute to distinguish regions in Germany based on their degree of rurality and socio-economic characteristics. It categorises counties and county-regions (similar to NUTS 3 regions) according to factors such as population density, settlement structure, and economic performance (e.g., employment and income levels). The typology includes four main rural area types, and a fifth category, not rural, captures more urbanised regions. 1983 to end of 1999 (new: 3,345) 2000 to end of 2009 (new: 12,931) 2010 to end of 2020 (new: 11,701) | 81 Figure 5-3: Share of households located within 6 km of wind turbines in Germany 2002-2020, by type of rural area Notes: Population estimates using the cross-sectional household population weights provided in SOEP. Source: SOEP (v39) linked with Core Energy Market Data Register (MaSTR, version of April 1, 2023, with own corrections). The results show a clear trend: the proportion of households living near wind turbines increased across all area types over time, including in non-rural areas. In the 2002–2010 period, 19% of households in socio-economically weak rural regions had at least one wind turbine within 6 km of their residence. This figure rose to 23% in the 2012– 2020 period. A similar pattern was observed in very rural, socio-economically weak areas, where the share increased from 14% to 19%. In socio-economically not weak rural areas, proximity also grew significantly—from 11% of households in 2002–2010 to 19% in 2012–2020. These trends indicate that wind turbine development is not limited to remote or disadvantaged rural areas. Notably, even in non-rural areas and rural areas with stronger socioeconomic profiles, exposure to wind turbines has risen. In these areas, the share of households within 6 km of a wind turbine grew from 9% to 14% over the two periods. Contrary to the common perception that wind turbine placement predominantly affects very rural and socio-economically weaker regions, the data reveal that wind turbine presence has expanded across the rural–urban spectrum. Table 5-2 presents the results of standard fixed effects regression of HRQoL on the presence of wind turbines using the whole SOEP sample (i.e., without identifying treatment and control cases and matching them on observables in t0). These models control for unobserved time-invariant individual characteristics but do not consider any aspects that change over time and that might (a) affect health‐related quality of life and (b) be correlated with the introduction (or growth) of wind turbines. The sample is also not restricted to non-movers (thus not accounting for unobserved time-invariant place factors) or people who participated more than once in the panel study (thus, individual fixed effects are not accounted for). 0,00 0,05 0,10 0,15 0,20 0,25 6 km 5 km 4 km 3 km 2 km 1.5 km 6 km 5 km 4 km 3 km 2 km 1.5 km Share of households 2002-2010 2012-2020 Very rural, socio-economically weak Very rural, socio-economically not weak Rural, socio-economically not weak Rural, socio-economically weak Not rural | 82 Table 5-2: Fixed effects regressions of health-related quality of life (SF-12) on the presence of wind turbines in a radius of 1 km to 6 km of residential homes, 2002-2022 Any wind turbine Mental health (MCS) Physical health (PCS) N β-coef. SE p-value β-coef. SE p-value 6 km 0.043 0.151 0.775 -1.468 *** 0.123 0.000 209273 5 km -0.012 0.173 0.945 -1.579 *** 0.138 0.000 209273 4 km -0.130 0.187 0.486 -1.703 *** 0.158 0.000 209273 3 km -0.190 0.243 0.435 -1.771 *** 0.199 0.000 209273 2 km -0.349 0.348 0.316 -1.673 *** 0.273 0.000 209273 1 km 0.348 0.922 0.705 -1.527 *** 0.590 0.010 209273 GE 500kW capacity 6 km -0.009 0.156 0.954 -1.479 *** 0.127 0.000 209273 5 km -0.060 0.179 0.740 -1.560 *** 0.145 0.000 209273 4 km -0.155 0.197 0.433 -1.664 *** 0.167 0.000 209273 3 km -0.160 0.254 0.529 -1.665 *** 0.209 0.000 209273 2 km -0.349 0.364 0.338 -1.627 *** 0.293 0.000 209273 1 km 0.022 0.966 0.982 -1.691 *** 0.616 0.006 209273 Hub GE 50m 6 km -0.013 0.161 0.934 -1.449 *** 0.130 0.000 209273 5 km -0.095 0.182 0.603 -1.567 *** 0.146 0.000 209273 4 km -0.198 0.201 0.325 -1.698 *** 0.170 0.000 209273 3 km -0.193 0.258 0.455 -1.732 *** 0.212 0.000 209273 2 km -0.379 0.370 0.306 -1.750 *** 0.294 0.000 209273 1 km -0.037 1.037 0.972 -1.934 *** 0.635 0.002 209273 Hub GE 100m 6 km 0.103 0.182 0.573 -1.989 *** 0.150 0.000 209273 5 km 0.128 0.215 0.551 -2.020 *** 0.179 0.000 209273 4 km 0.092 0.261 0.725 -2.135 *** 0.216 0.000 209273 3 km 0.093 0.351 0.792 -2.309 *** 0.266 0.000 209273 2 km 0.008 0.556 0.988 -2.338 *** 0.401 0.000 209273 1 km 0.857 1.379 0.534 -2.559 *** 0.804 0.001 209273 Notes: Robust standard errors. Statistical significance at 95%-level: * p<0.05, ** p<0.005, *** p<0.001. See Appendix 15 (A5-3) for additional results on associations with the number of wind turbines and with the inverse distance to wind turbines. Results with additional socio-economic controls are also reported. Source: SOEP (v39) linked with Core Energy Market Data Register (MaSTR, version of April 1, 2023, with own corrections). We present findings on both components of health-related quality of life (HRQoL). For the mental health component, we found no statistically significant differences in Mental Component Summary (MCS) scores between individuals living near wind turbines and those who do not. Although the coefficients suggest a negative association across all groups of wind turbines—whether considering any wind turbines or larger ones with at least 500 kW capacity and varying hub heights (>=50 m and >=100 m)—these associations do not reach conventional levels of statistical significance. In contrast, all associations between physical HRQoL (PCS) and wind turbines within radii ranging from 1 km to 6 km are negative and statistically significant. We observe a general trend of increasing negative associations as the radius is tightened. The size of the coefficients increases slightly when excluding the smallest wind turbines, which are typically located closer to homes. In Table 5-3, we present the findings from more rigorous difference-in-differences fixed effects panel estimations conducted on matched samples of non-movers. These samples include individuals who experienced the | 83 construction of a wind turbine within a specified radius around their residences (treatment group) and those who did not encounter such construction but had a wind turbine located between 10 km and 15 km from their home (control group). The analysis compares two measurement occasions of health-related quality of life (HRQoL). Table 5-3: Fixed effects difference-in-differences panel regressions of health-related quality of life (SF-12) on the presence of wind turbines in a radius of 1.5 km to 6 km of residential homes (quasi-experimental sample) Any wind turbines Mental health (MCS) Physical health (PCS) β-coef. SE p-value β-coef. SE p-value N 6 km 0.216 0.626 0.730 0.531 0.491 0.279 1764 5 km -0.187 0.750 0.803 0.504 0.600 0.401 1164 4 km 0.007 0.929 0.994 0.669 0.747 0.371 792 3 km -0.555 1.421 0.697 2.243 1.261 0.077 316 2 km 1.455 2.034 0.477 2.783 1.782 0.123 128 1.5 km 1.047 2.294 0.651 0.611 2.284 0.791 76 GE 500kW capacity 6 km 0.227 0.634 0.721 -0.043 0.495 0.930 1724 5 km 0.185 0.751 0.805 -0.194 0.606 0.749 1124 4 km 0.028 0.937 0.976 0.054 0.745 0.942 756 3 km -0.349 1.382 0.801 0.645 1.104 0.560 328 2 km -1.680 2.074 0.421 2.374 1.766 0.184 124 1.5 km 0.429 2.603 0.870 -0.097 2.317 0.967 72 hub GE 50m 6 km -0.066 0.650 0.919 0.673 0.499 0.178 1692 5 km -0.576 0.784 0.463 0.357 0.604 0.555 1124 4 km -0.450 0.965 0.641 0.678 0.702 0.335 780 3 km -2.443 1.394 0.082 2.299 * 1.129 0.043 328 2 km -2.840 2.078 0.177 3.577 * 1.772 0.048 128 1.5 km -0.622 2.368 0.794 0.253 2.238 0.911 76 hub GE 100m 6 km -0.227 0.518 0.672 0.896 * 0.405 0.027 2668 5 km 0.345 0.653 0.815 0.524 0.494 0.289 1756 4 km -0.140 0.793 0.798 0.378 0.638 0.554 1076 3 km -0.857 1.125 0.633 0.398 0.976 0.684 516 2 km -1.632 1.879 0.493 2.966 2.198 0.183 112 1.5 km -0.559 3.655 0.860 -4.916 4.732 0.315 32 Notes: Robust standard errors. Statistical significance at 95%-level: * p<0.05, ** p<0.005, *** p<0.001. See Appendix 16 (A5-4). Source: SOEP (v39) linked with Core Energy Market Data Register (MaSTR, version of April 1, 2023, with own corrections). The results for the quasi-experimental sample generally indicate that there are no statistically significant associations between the presence of wind turbines within a 6 km radius of individuals' residences and their physical or mental HRQoL. However, we find several statistically significant associations when focusing on larger wind turbines. Specifically, individuals living within a radius of 2 km and 3 km of wind turbines with a hub height of at least 50 m have, on average, increased their physical HRQoL (+3.577, p<0.05, and +2.299, p<0.05, respectively) compared to the control group. Additionally, for individuals near turbines with a hub height of at least 100 m within 6 km, the mean difference in the PCS amounts to +0.896 (p<0.05). The latter two improvements in mean PSC scores in the treatment | 84 group compared to the control group are robust to the inclusion of a range of individual and area-level controls (+0.9, p<0.05, and +2.550, p<0.05, respectively), see Appendix 15 (A5-3). Table 5-4: Fixed effects difference-in-differences panel regressions of health-related quality of life (SF-12) on number of wind turbines (quasi-experimental sample) All wind turbines Mental health (MCS) Physical health (PCS) N β-coef. SE p-value β-coef. SE p-value 6 km -0.133 0.216 0.539 0.077 0.205 0.707 1764 5 km -0.341 0.261 0.191 0.121 0.269 0.654 1164 4 km -0.609 0.314 0.053 0.266 0.344 0.440 792 3 km -0.817 * 0.331 0.015 0.334 0.487 0.494 316 2 km 0.800 1.981 0.688 2.388 1.648 0.152 128 1.5 km 1.047 2.294 0.651 0.611 2.284 0.791 76 GE 500kW capacity 6 km -0.142 0.209 0.498 -0.032 0.188 0.865 1724 5 km -0.281 0.259 0.279 -0.009 0.260 0.973 1124 4 km -0.654 * 0.315 0.038 0.115 0.331 0.728 756 3 km -0.739 * 0.324 0.024 0.026 0.429 0.953 328 2 km -1.950 1.877 0.303 1.974 1.625 0.229 124 1.5 km 0.429 2.603 0.870 -0.097 2.317 0.967 72 hub GE 50m 6 km -0.166 0.208 0.425 0.121 0.197 0.538 1692 5 km -0.430 0.258 0.096 0.145 0.258 0.575 1124 4 km -0.667 * 0.307 0.030 0.314 0.315 0.320 780 3 km -1.161 ** 0.350 0.001 0.359 0.478 0.453 328 2 km -3.015 1.873 0.112 3.094 1.659 0.067 128 1.5 km -0.622 2.368 0.794 0.253 2.238 0.911 76 hub GE 100m 6 km 0.052 0.088 0.554 0.080 0.062 0.197 2668 5 km 0.121 0.094 0.199 0.051 0.067 0.447 1756 4 km 0.217 0.136 0.112 0.051 0.115 0.658 1076 3 km 0.150 0.215 0.487 0.088 0.149 0.555 516 2 km -0.814 0.818 0.324 1.487 1.006 0.145 112 1.5 km -0.559 3.655 0.881 -4.916 4.732 0.315 32 Notes: Robust standard errors. Statistical significance at 95%-level: * p<0.05, ** p<0.005, *** p<0.001. See Appendix 16 (A5-4). Source: SOEP (v39) linked with Core Energy Market Data Register (MaSTR, version of April 1, 2023, with own corrections). Wind turbines, particularly the larger ones operated by commercial providers, are often put up in wind farms. This can also be observed in our quasi-experimental samples. 32 The experience of having many large wind turbines placed in one's neighbourhood may be more stressful than having one small wind turbine put up. Thus, rather than just examining whether the presence of a wind turbine matters for HRQoL, next we check whether the number of wind turbines matters, see Table 5-4. We find empirical support for the hypothesis that the intensity of treatment matters in some of our samples. Regarding mental health, the results suggest that each additional wind turbine within a radius of 3 km is associated with a reduction of -0.817 (p<0.05) in the MCS; if the wind turbine has a hub height of at least 50 m, the reduction is more marked, amounting to -1.161 (p<0.01). Within 4 km, wind turbines with 32 The individuals in our treatment samples had between 1 and 16 (Mean: 1.8) wind turbines built within 6 km of their home; for the largest wind turbines (hub heights of at least 100 m), the respective average amounts to 2.876, see Appendix 5-1. | 85 500 kW or more are associated with a -0.654 (p<0.05) in MCS; for wind turbines >= 50m, the effect is in the same ballpark (-0.667, p<0.05). Wind turbines do not impact physical health. Table 5-5: Fixed effects DiD panel regressions of health-related quality of life (SF-12) on inverse distance to wind turbines (quasi-experimental sample) All wind turbines Mental health (MCS) Physical health (PCS) N β-coef. SE p-value β-coef. SE p-value 6 km -0.013 0.015 0.367 -0.010 0.011 0.399 1764 5 km -0.003 0.021 0.886 -0.012 0.017 0.476 1164 4 km -0.007 0.031 0.818 -0.013 0.025 0.613 792 3 km 0.014 0.068 0.837 -0.109 0.058 0.061 316 2 km -0.114 0.147 0.442 -0.192 0.115 0.100 128 1.5 km -0.117 0.182 0.523 0.002 0.192 0.993 76 GE 500kW capacity 6 km -0.014 0.015 0.354 0.002 0.011 0.866 1724 5 km -0.016 0.021 0.442 0.009 0.017 0.584 1124 4 km -0.018 0.032 0.559 0.009 0.025 0.711 756 3 km -0.010 0.067 0.883 -0.034 0.052 0.517 328 2 km 0.174 0.147 0.239 -0.174 0.114 0.131 124 1.5 km -0.070 0.205 0.733 0.062 0.192 0.750 72 hub GE 50m 6 km -0.005 0.015 0.720 -0.014 0.011 0.210 1692 5 km 0.004 0.022 0.868 -0.005 0.018 0.760 1124 4 km -0.004 0.032 0.894 -0.011 0.024 0.643 780 3 km 0.079 0.067 0.245 -0.104 0.053 0.051 328 2 km 0.252 0.147 0.092 -0.253 * 0.115 0.032 128 1.5 km 0.017 0.189 0.928 0.030 0.188 0.873 76 hub GE 100m 6 km -0.001 0.012 0.914 -0.020 * 0.010 0.035 2668 5 km -0.027 0.020 0.170 -0.015 0.014 0.287 1756 4 km -0.013 0.029 0.654 -0.014 0.022 0.532 1076 3 km 0.024 0.051 0.632 -0.013 0.045 0.776 516 2 km 0.116 0.115 0.321 -0.218 0.134 0.110 112 1.5 km 0.040 0.272 0.886 0.360 0.352 0.323 32 Notes: Robust standard errors. Statistical significance at 95%-level: * p<0.05, ** p<0.005, *** p<0.001. See Appendix 16 (A5-4). Source: SOEP (v39) linked with Core Energy Market Data Register (MaSTR, version of April 1, 2023, with own corrections). Last but not least, we examine whether the proximity to wind turbines matters for HRQoL. The results, presented in Table 5-5, suggest that greater proximity to wind turbines is associated with reductions in HRQoL – the majority of the coefficients are negative – but only the effects on physical HRQoL of distances to wind turbines with at least 50 m hub height in a radius of 2 km (-0.253, p<0.05) and wind turbines with at least 100m hub height in a radius of 6 km (-0.020, p<0.05) reach conventional levels of statistical significance. | 86 5.4. Discussion Our analysis yields a nuanced picture of how wind turbines relate to health-related quality of life (HRQoL). The standard fixed effects regressions on the full SOEP sample (Table 5-2) suggest that, while mental HRQoL (MCS) shows no systematic relationship with proximity to wind turbines, physical HRQoL (PCS) declines significantly among individuals living closer to turbines—an effect that intensifies as the radius is tightened and when smaller turbines are excluded. This pattern is consistent with prior cross-sectional findings linking wind turbine proximity to sleep disturbance, annoyance, and somatic complaints (Shepherd, McBride et al. 2011, Bakker, Pedersen et al. 2012), which may manifest as reduced physical wellbeing even if mental health remains unaffected. However, when leveraging the quasi-experimental, difference-in-differences design on matched non-movers (Table 5-3), we observed no adverse effects on either HRQoL component for turbines within a 6 km radius. Intriguingly, larger turbines (hub height ≥ 50 m and ≥ 100 m) are associated with modest improvements in physical HRQoL (+3.58 PCS points at 2 km; +2.30 at 3 km; +0.90 at 6 km), all statistically significant at p < 0.05. One plausible interpretation is that larger, commercially operated turbines are often erected as part of wind farms, which may bring local economic benefits—such as employment opportunities, enhanced infrastructure, or community funds (Gavard, Göbel and Schoch 2025)—that offset any direct nuisances and translate into perceived physical wellbeing gains. Our intensity analyses (Table 5-4) further refine this story by showing that the number of turbines within 3–4 km is negatively associated with mental HRQoL: each additional turbine reduces MCS by 0.65–1.16 points, depending on turbine size. This finding implies that cumulative exposure may exacerbate annoyance or stress even if proximity alone or a single new turbine does not. Notably, physical HRQoL remains unaffected by turbine count, indicating that the mechanisms through which turbine density influences wellbeing are primarily psychological or perceptual. Finally, examining inverse distance measures (Table 5-5) confirms that closer proximity to larger turbines can modestly depress physical HRQoL (−0.253 PCS points per 1/distance unit for ≥ 50 m turbines within 2 km; −0.020 for ≥ 100 m turbines within 6 km). These small but statistically significant effects underscore that even in our more rigorous designs, some detrimental impacts on physical wellbeing persist, particularly for the tallest turbines. How large are these negative effects on mental health? Cleland, Kearns et al. (2016) reported several statistically significant associations between HRQoL and various life events. Specifically, MCS is reduced by a relationship breakdown (−7.3), a health event (−6.0), being the victim of a crime (−5.4), bereavement (−2.3) and a house move (−2.1). MCS is increased by a new job/promotion (+4.9) and parenthood (+3.6). Despite the strengths of panel fixed effects and matched quasi-experimental designs, several limitations warrant caution. First, our models cannot entirely rule out time-varying confounders (e.g., evolving local labour markets or concurrent infrastructure developments) that correlate with turbine installation. Second, restricting the quasiexperimental sample to non-movers may limit generalisability to more mobile populations. However, from this perspective, our results may be regarded as lower-bound estimates, as those most affected by wind turbines will Key messages • Component-specific effects: Physical and mental HRQoL respond differently to turbine exposure. While physical wellbeing appears most sensitive to proximity and turbine size, mental wellbeing is influenced by overall turbine density. • Importance of research design: Effects apparent in uncontrolled fixed effects models can attenuate or reverse when leveraging quasi-experimental matching, suggesting that unobserved confounders (e.g., pre-existing neighbourhood amenities or individual health trends) may bias simpler specifications. • Heterogeneity by turbine characteristics: Larger turbines—often sited in commercial wind farms—may confer indirect benefits that improve physical HRQoL, even as increased density heightens mental strain. • Treatment intensity matters: More turbines correlate with worsened mental health, highlighting the need to consider cumulative effects in policy and planning. | 87 likely move to areas without wind turbines. Third, HRQoL measures, while validated, may not capture specific wind turbine–related annoyances (e.g., shadow flicker, infrasound perception), suggesting the value of incorporating specialised noise and annoyance metrics. Finally, the average follow-up duration between preand post-installation HRQoL measures may be too short to observe long-term adaptation or chronic health impacts. Future studies should aim to integrate objective noise measurements – possibly by adding further wind turbine typespecific information from providers - and longitudinal follow-up over extended periods, as well as broader wellbeing indicators (e.g., community cohesion, economic satisfaction). Qualitative research exploring residents’ subjective experiences could also illuminate the pathways linking turbine exposure to components of HRQoL. Framing this research neutrally – i.e., without biasing participants in a particular direction - and recruiting neutral participants - i.e., those who are not already determined wind energy supporters or opponents - is challenging in areas already earmarked for wind energy. Germany represents a unique case for analysing the wellbeing impacts of wind turbine exposure, owing to the combination of a comprehensive, open-access wind turbine register with precise coordinates and installation dates and the availability of detailed, geocoded longitudinal survey data from the SOEP. Replicating such research in other European countries remains challenging. While several countries maintain national databases of wind turbine locations—such as France with the Base OREOL (Ministry for the Ecological Transition and Territorial Cohesion 2025) and Sweden with the Vindbrukskollen (Länsstyrelserna gemensamt 2025)—few possess long-running, largescale panel datasets with georeferenced residential information and subjective wellbeing measures that would allow for comparable micro-level linkage. The United Kingdom offers high-quality longitudinal data through Understanding Society and wind turbine information via the Renewable Energy Planning Database (Department for Energy Security and Net Zero 2025). However, both data sources are less granular than their German counterparts: the wind turbine datasets lack exact installation dates and individual turbine coordinates, and the panel study does not provide precise residential geocodes for research use. 33 Moreover, strict planning regulations have thus far resulted in relatively low residential exposure to onshore wind turbines in the UK, limiting statistical power to detect wellbeing effects. Consequently, Germany remains a uniquely data-rich context for studying the causal impacts of renewable energy infrastructure on subjective wellbeing. While the German data exploited here is unique, findings may be generalisable because the underlying mechanisms linking wind turbine exposure to wellbeing—such as visual, acoustic, and social perceptions of renewable infrastructure—are not unique to Germany. Similar planning frameworks, settlement patterns, and attitudes toward wind energy exist across much of Europe. Moreover, the analytical design, which exploits within-individual variation and controls for both personal and neighbourhood fixed effects, identifies causal pathways likely to operate similarly in other contexts. While contextual differences in regulation, turbine density, and community engagement may affect the magnitude of effects, the fundamental processes underlying the German results are expected to extend to other European settings. 5.5. Conclusions This study leverages a large, nationally representative panel and quasi-experimental matching to untangle the complex relationships between wind turbine exposure and HRQoL. While simple fixed-effects analyses suggest that proximity to turbines diminishes physical wellbeing, rigorous difference-in-differences estimations reveal that larger turbines may paradoxically enhance physical HRQoL—potentially via community benefits—though higher turbine density consistently undermines mental health. These findings highlight the need for strategies that reduce clustering of wind turbines near residences, integrate community benefit schemes, and proactively address any noise and visual impact concerns local residents may have. Policymakers and developers should also consider the mental health and wellbeing of the local population, not only physical health impacts and impacts on objective wellbeing metrics. It is important to ensure that renewable energy goals align with the subjective wellbeing of local populations. Future work that harnesses objective environmental data, incorporates longer follow-up periods, and utilises granular annoyance measures will be crucial to elucidate the health implications of wind power deployment fully. 33 In secure data settings, only the coordinates of postcode centroids—not individual addresses—are available for analysis. Because postcode areas are much larger in rural areas than in urban ones, where a postcode may correspond to a single dwelling, distance-based measures of exposure are considerably less precise in rural contexts. | 88 6. Overall conclusions This report has sought to advance understanding of subjective wellbeing (SWB) across rural (and urban) Europe, combining systematic review, conceptual extension, and novel empirical analysis. The findings collectively confirm that rural contexts often foster wellbeing advantages, but these are not uniform, nor are they guaranteed. They depend on how “rurality” is defined, which wellbeing dimensions are considered, and which contextual features are taken into account. Across the chapters, several consistent themes emerge: ▪ Rural advantage in subjective wellbeing exists, but is heterogeneous. Life satisfaction is, on average, higher in rural areas, yet advantages in affective wellbeing and basic psychological needs are less robust. Social wellbeing and mental resources (e.g., resilience) are the most consistently higher among rural residents. ▪ Measurement matters. Studies using self-reported settlement type reveal stronger rural advantages than those using administrative or density-based classifications. This highlights the need for greater precision and transparency in defining “rural” and “urban.” ▪ Contextual conditions are critical. Area deprivation consistently erodes life satisfaction, particularly in urban areas. Local service access, environmental quality, and neighbourhood composition moderate wellbeing outcomes, demonstrating that rural advantages cannot be reduced to population density alone. ▪ Environmental trade-offs are complex. Wind turbines exemplify the nuanced relationship between rural development and wellbeing. While simple models suggest that proximity harms physical health, quasiexperimental evidence indicates potential wellbeing gains from larger turbines, linked to community benefits, alongside mental health costs associated with high turbine density. Taken together, these results underscore that rural wellbeing cannot be understood in isolation from context. The “rural advantage” is real but contingent—shaped by social ties, mental resources, service access, environmental quality, and the ways rural development is planned and implemented. Policy lessons: ▪ Rural policies should explicitly incorporate subjective wellbeing alongside economic indicators, recognising the multidimensional nature of quality of life. ▪ Fine-grained measurement and monitoring are essential—moving beyond simple rural–urban divides to capture the diversity of rural contexts. ▪ Investment in local services, reducing deprivation, and strengthening social cohesion can reinforce rural wellbeing advantages. ▪ Renewable energy and other rural developments should be pursued with community engagement and benefit-sharing mechanisms to align environmental goals with residents’ quality of life. Future research priorities: ▪ Broaden geographic coverage, particularly in underrepresented regions of Europe. ▪ Standardise measurement instruments, integrating evaluative and affective dimensions of wellbeing. ▪ Exploit longitudinal and geocoded data to unpack causal mechanisms and track change over time. ▪ Link wellbeing outcomes more directly to policy interventions to assess their real-world impact. In conclusion, the report affirms the importance of rural contexts for wellbeing while highlighting the complexity of the pathways involved. By deepening our understanding of these dynamics, policymakers and researchers can better support thriving rural communities and ensure that the pursuit of economic, social, and environmental goals is grounded in the lived experiences of Europe’s citizens. | 95 Michalska-Żyła, A. and M. Marks-Krzyszkowska (2018). "Quality of Life and Quality of Living in Rural Communes in Poland." European Countryside 10(2): 280-299. Ministry for the Ecological Transition and Territorial Cohesion (2025). OREOL database – Onshore wind turbines. . Géorisques. https://www.georisques.gouv.fr/donnees/bases-de-donnees/eoliennes-terrestres. Mollenkopf, H. and R. Kaspar (2005). "Ageing in rural areas of East and West Germany: Increasing similarities and remaining differences." European Journal of Ageing 2(2): 120-130. Mollenkopf, H., R. Kaspar, F. Marcellini, I. Ruoppila, Z. SzÉMan, M. Tacken and H.-W. Wahl (2004). "Quality of life in urban and rural areas of five European countries: Similarities and differences." Hallym International Journal of Aging 6(1): 1-36. Morrison, M., L. Tay and E. Diener (2011). "Subjective Well-Being and National Satisfaction:Findings From a Worldwide Survey." Psychological Science 22(2): 166-171. Morrison, P. S. (2021). "Wellbeing and the Region." Handbook of Regional Science: Second and Extended Edition: With 238 Figures and 78 Tables: 779-798. Morrison, P. S. and M. Weckroth (2018). "Human values, subjective well-being and the metropolitan region." Regional Studies 52(3): 325-337. Murray, T., D. Maddison and K. Rehdanz (2011). "Do geographical variations in climate influence life satisfaction?" Kiel Working Paper (1694). Navarro, M., A. D'Agostino and L. Neri (2020). "The effect of urbanization on subjective well-being: Explaining crossregional differences." Socio-Economic Planning Sciences 71. Nelson, K. S., T. D. Nguyen, N. A. Brownstein, D. Garcia, H. C. Walker, J. T. Watson and A. Xin (2021). "Definitions, measures, and uses of rurality: A systematic review of the empirical and quantitative literature." Journal of Rural Studies 82: 351-365. Nübling, M., H. H. Andersen, A. Mühlbacher, J. Schupp and G. Wagner (2007). "Computation of Standard Values for Physical and Mental Health Scale Scores Using the SOEP Version of SF12v2." Schmollers Jahrbuch : Journal of Applied Social Science Studies / Zeitschrift für Wirtschaftsund Sozialwissenschaften 127(1): 171-182. O’Sullivan, R., A. Burns, G. Leavey, I. Leroi, V. Burholt, J. Lubben, J. Holt-Lunstad, C. Victor, B. Lawlor, M. Vilar-Compte, C. M. Perissinotto, M. A. Tully, M. P. Sullivan, M. Rosato, J. M. Power, E. Tiilikainen and T. R. Prohaska (2021). "Impact of the COVID-19 Pandemic on Loneliness and Social Isolation: A Multi-Country Study." International Journal of Environmental Research and Public Health 18(19): 9982. Office for National Statistics (2015). "Radial plots for the 2011 Area Classification for Output Areas." https://www.ons.gov.uk/methodology/geography/geographicalproducts/areaclassifications/2011areaclassifications/penpo rtraitsandradialplots. Office for National Statistics. (n.d.). "ONS Geoportal." Retrieved 12.09.2025, from https://geoportal.statistics.gov.uk/. Okulicz-Kozaryn, A. and J. M. Mazelis (2018). "Urbanism and happiness: A test of Wirth’s theory of urban life." Urban Studies 55(2): 349-364. Okulicz-Kozaryn, A. and R. R. Valente (2021). "Urban unhappiness is common." Cities 118. Ortega-Reig, M., C. Schürmann, A. Ferrandis Martínez and A. Copus (2023). "Measuring Access to Services of General Interest as a Diagnostic Tool to Identify Well-Being Disparities between Rural Areas in Europe." Land 12(5): 1049. Ory, B. and T. Huijts (2015). "Widowhood and Well-being in Europe: The Role of National and Regional Context." Journal of Marriage and Family 77(3): 730-746. Oswald, A. J. and S. Wu (2011). "Well-Being across America." The Review of Economics and Statistics 93(4): 1118-1134. Oswald, F., H.-W. Wahl, H. Mollenkopf and O. Schilling (2003). "Housing and life satisfaction of older adults in two rural regions in Germany." Research on Aging 25(2): 122-143. Pebesma, E. (2018). "Simple Features for R: Standardized Support for Spatial Vector Data." R Journal 10(1): 439-446. Pebesma, E. and R. Bivand (2023). Spatial Data Science: With Applications in R. Chapman and Hall/CRC. Pike, A., V. Béal, N. Cauchi-Duval, R. Franklin, N. Kinossian, T. Lang, T. Leibert, D. MacKinnon, M. Rousseau, J. Royer, L. Servillo, J. Tomaney and S. Velthuis (2023). "‘Left behind places’: a geographical etymology." Regional Studies 58(6): 1167-1179. Platt, L., G. Knies, R. Luthra, A. Nandi and M. Benzeval (2020). "Understanding Society at 10 Years." European Sociological Review 36(6): 976-988. Plaut, V. C., G. Adams and S. L. Anderson (2009). "Does attractiveness buy happiness? "it depends on where you're from"." Personal Relationships 16(4): 619-630. Podgorelec, S., M. Gregurović and S. K. Bogadi (2015). "Satisfaction with the quality of life on Croatian small islands: Zlarin, Kaprije and Žirje." Island Studies Journal 10(1): 91-110. Prati, G. (2024). "The Relationship Between Rural-Urban Place of Residence and Subjective Well-Being is Nonlinear and its Substantive Significance is Questionable." International Journal of Applied Positive Psychology 9(1): 27-43. Putnam, R. D. (1995). "Bowling alone: America's declining social capital." Journal of Democracy 6(1): 65–78. R Core Team (2024). _R: A Language and Environment for Statistical Computing_. . Vienna, Austria, R Foundation for Statistical Computing. Requena, F. (2016). "Rural-Urban Living and Level of Economic Development as Factors in Subjective Well-Being." Social Indicators Research 128(2): 693-708. Róbert, P., N. Geszler and B. Nagy (2022). "Parental determinants of subjective child well-being in Hungary." Intersections. East European Journal of Society and Politics 8(2): 156-174. Rodríguez-Pose, A. and V. von Berlepsch (2014). "Social Capital and Individual Happiness in Europe." Journal of Happiness Studies 15(2): 357-386. Rubin, D. B. (2001). "Using Propensity Scores to Help Design Observational Studies: Application to the Tobacco Litigation." Health Services and Outcomes Research Methodology 2(3): 169-188. | 96 Ruggeri, K., E. Garcia-Garzon, Á. Maguire, S. Matz and F. A. Huppert (2020). "Well-being is more than happiness and life satisfaction: a multidimensional analysis of 21 countries." Health and Quality of Life Outcomes 18(1): 192. Ryan, R. M. and E. L. Deci (2001). "On Happiness and Human Potentials: A Review of Research on Hedonic and Eudaimonic Well-Being." Annual Review of Psychology 52(Volume 52, 2001): 141-166. Ryan, R. M. and E. L. Deci (2017). Self-determination theory: Basic psychological needs in motivation, development, and wellness, The Guilford Press. Ryff, C. D., J. M. Boylan and J. A. Kirsch (2021). Eudaimonic and hedonic well-being: An integrative perspective with linkages to sociodemographic factors and health. Measuring well-being: Interdisciplinary perspectives from the social sciences and the humanities. New York, NY, US, Oxford University Press: 92-135. Saisana, M. and S. Tarantola (2002). State-of-the-art report on current methodologies and practices for composite indicator development European Commission Joint Research Centre. Ispra (VA) Institute for the Protection and the Security of the Citizen, Technological and Economic Risk Management Unit. EUR 20408 EN. Samavati, S. and R. Veenhoven (2024). "Happiness in urban environments: what we know and don’t know yet." Journal of Housing and the Built Environment 39(3): 1649-1707. Sánchez-Sellero, M.-C., B. García-Carro and P. Sánchez-Sellero (2021). "Synthetic Indicators of Quality of Subjective Life in the EU: Rural and Urban Areas." Prague Economic Papers 30(5): 529-551. Schenker, N. and J. F. Gentleman (2001). "On Judging the Significance of Differences by Examining the Overlap Between Confidence Intervals." The American Statistician 55(3): 182-186. Schiavina, M., S. Freire, A. Carioli and K. MacManus (2023). GHS-POP R2023A - GHS population grid multitemporal (1975-2030). J. R. C. J. European Commission. Schiavina, M., M. Melchiorri and P. M. (2023). GHS-SMOD R2023A - GHS settlement layers, application of the Degree of Urbanisation methodology (stage I) to GHS-POP R2023A and GHS-BUILT-S R2023A, multitemporal (1975-2030) J. R. C. J. European Commission. Schilling, O. and H.-W. Wahl (2002). "Family networks and life-satisfaction of older adults in rural and urban regions." Kolner Zeitschrift fur Soziologie und Sozialpsychologie 54(2): 304-317+415. Schmid, L., P. Christmann, A.-S. Oehrlein, A. Stein and C. Thönnissen (2023). "Life Satisfaction during the Second Lockdown of the COVID-19 Pandemic in Germany: The Effects of Local Restrictions and Respondents’ Perceptions about the Pandemic." Applied Research in Quality of Life. Schmitz, A. and M. Brandt (2022). "Health Limitations, Regional Care Infrastructure and Wellbeing in Later Life: A Multilevel Analysis of 96 European Regions." Social Indicators Research 164(2): 693-709. Schnaudt, C., M. Weinhardt, R. Fitzgerald and S. Liebig (2014). "The European Social Survey: Contents, Design, and Research Potential." Journal of Contextual Economics – Schmollers Jahrbuch 134(4): 487-506. Shapiro, A. and C. L. M. Keyes (2008). "Marital Status and Social Well-Being: Are the Married Always Better Off?" Social Indicators Research 88(2): 329-346. Shepherd, D., D. McBride, D. Welch, K. Dirks and E. Hill (2011). "Evaluating the impact of wind turbine noise on healthrelated quality of life." Noise and Health 13(54): 333-339. Shields, M. A., S. Wheatley Price and M. Wooden (2009). "Life satisfaction and the economic and social characteristics of neighbourhoods." Journal of Population Economics 22(2): 421-443. Shucksmith, M. (2012). Future Directions in Rural Development, Carnegie UK Trust. Shucksmith, M. (2018). "Re-imagining the rural: From rural idyll to Good Countryside." Journal of Rural Studies 59: 163172. Shucksmith, M. (2018). "Rural policy after Brexit." Contemporary Social Science 14(2): 312-326. Shucksmith, M., S. Cameron, T. Merridew and F. Pichler (2009). "Urban-rural differences in quality of life across the European Union." Regional Studies 43(10): 1275-1289. Silva, J. and Z. Brown (2013). More than the Sum of their Parts: Valuing Environmental Quality by Combining Life Satisfaction Surveys and GIS Data. OECD Statistics Working Papers OECD: 23. Sorensen, J. F. L. (2014). "Rural-Urban Differences in Life Satisfaction: Evidence from the European Union." Regional Studies 48(9): 1451-1466. Sorensen, J. F. L. (2021). "The rural happiness paradox in developed countries." Soc Sci Res 98: 102581. Sørensen, J. F. L. (2021). "The rural happiness paradox in developed countries." Social Science Research 98: 102581. Spellerberg, A., D. Huschka and R. Habich (2007). "Quality of life in rural areas: processes of divergence and convergence." Social Indicators Research 83(2): 283-307. StataCorp (2021). Statistical Software: Release 17.0. College Station, Texas, Stata Corporation. Steptoe, A., A. Deaton and A. A. Stone (2015). "Subjective wellbeing, health, and ageing." Lancet 385(9968): 640-648. Sutherland, A., I. Brunton-Smith and J. Jackson (2013). "Collective Efficacy, Deprivation and Violence in London." The British Journal of Criminology 53(6): 1050-1074. Tobiasz-Adamczyk, B. and K. Zawisza (2017). "Urban-rural differences in social capital in relation to self-rated health and subjective well-being in older residents of six regions in Poland." Annals of Agricultural and Environmental Medicine 24(2): 162-170. Tönnies, F. (1887). Community and association. London, Routledge & Kegan Paul. Townsend, P. (1979). Poverty in the United Kingdom: a survey of household resources and standards of living. Harmondsworth, Penguin Books. Townsend, P., P. Phillimore and A. Beattie (1988). Health and Deprivation: Inequality and the North, Croom Helm. Triadó, C., F. Villar, C. Solé, M. Celdrán and M. J. Osuna (2009). "Daily activity and life satisfaction in older people living in rural contexts." Spanish Journal of Psychology 12(1): 236-245. University of Essex and Institute for Social and Economic Research (2024). Understanding Society: Waves 1-14, 20092023: Special Licence Access, Acorn Type 2015. [data collection] Understanding Society: Study number 6614. Institute for Social and Economic Research and NatCen. Colchester, UK Data Service. | 97 University of Essex and Institute for Social and Economic Research (2024). Understanding Society: Waves 1-14, 20092023: Special Licence Access, Census 2011 Lower Layer Super Output Areas. [data collection] Understanding Society: Study number 6614. Institute for Social and Economic Research and NatCen Social Research. Colchester, UK Data Service. University of Essex and Institute for Social and Economic Research (2024). Understanding Society: Waves 1-14, 20092023: Special Licence Access, Census 2011 Output Area Classification. [data collection] Understanding Society: Study number 7629. Institute for Social and Economic Research and NatCen. Colchester, UK Data Service. University of Essex, Institute for Social and Economic Research, NatCen Social Research and Kantar Public (2024). Understanding Society: Waves 1-14, 2009-2023 and Harmonised BHPS: Waves 1-18, 1991-2009. [data collection] Understanding Society: Study number 6614. University of Essex, Institute for Social and Economic Research, NatCen Social Research and K. Public. Colchester, UK Data Service. van den Berg, A. E., J. Maas, R. A. Verheij and P. P. Groenewegen (2010). "Green space as a buffer between stressful life events and health." Social Science & Medicine 70: 1203-1210. Vaznonienė, G. and A. Wojewódzka-Wiewiórska (2021). "Territorial dimension of rural population wellbeing: cases of Lithuania and Poland " Economics and Sociology 14(4): 167-185. Veenhoven, R. (1994). How satisfying is rural life? Fact and value. Changing values and attitudes in family households, implications for institutional transition in East and West. Bonn, Germany, Society for agricultural policy research in rural society. 296: 41-51. Veenhoven, R. (2014). World Database of Happiness. Encyclopedia of Quality of Life and Wellbeing Research: 72577260. Veréb, V., C. Marques, L. Madureira, C. Marques, T. Keryan and R. Silva (2024). "What Is Rural Well-Being and How Is It Measured? An Attempt to Order Chaos." Rural Sociology 89(2): 239-280. von Möllendorff, C. and H. Welsch (2017). "Measuring Renewable Energy Externalities: Evidence from Subjective Wellbeing Data." Land Economics 93(1): 109-126. Voukelatou, V., L. Gabrielli, I. Miliou, S. Cresci, R. Sharma, M. Tesconi and L. Pappalardo (2020). "Measuring objective and subjective well-being: dimensions and data sources." International Journal of Data Science and Analytics 11(4): 279309. Wagner, G. G., J. R. Frick and J. Schupp (2007). "The German Socio-Economic Panel Study (SOEP) - Scope, Evolution and Enhancements." Schmollers Jahrbuch 127(1): 139-169. Wang, F. and D. Wang (2015). Place, Geographical Context and Subjective Well-being: State of Art and Future Directions. Mobility, Sociability and Well-Being of Urban Living. D. Wang: 189-230. Ware, J. E., M. Kosinski and S. D. Keller (1996). "A 12-Item Short-Form Health Survey: Construction of Scales and Preliminary Tests of Reliability and Validity." Medical Care 34(3): 220-233. Waterman, A. S. (2007). "On the importance of distinguishing hedonia and eudaimonia when contemplating the hedonic treadmill. ." American Psychologist 62(6): 612-613. Waterman, A. S. (2008). "Reconsidering happiness: A eudaimonist's perspective." The Journal of Positive Psychology 3(4): 234-252. Weckroth, M. and T. Kemppainen (2021). "(Un)Happiness, where are you? Evaluating the relationship between urbanity, life satisfaction and economic development in a regional context." Regional Studies, Regional Science 8(1): 207-227. White, M. P., I. Alcock, B. W. Wheeler and M. H. Depledge (2013). "Would you be happier living in a greener urban area? A fixed-effects analysis of panel data." Psychological science 24(6): 920-928. Wickham, H., R. François, L. Henry, K. Müller and V. Davis (2023). _dplyr: A Grammar of Data Manipulation_. https://CRAN.R-project.org/package=dplyr. Wiesli, T. X. and W. Przepiorka (2023). "Does Living in a Protected Area Reduce Resource Use and Promote Life Satisfaction? Survey Results from and Around Three Regional Nature Parks in Switzerland." Social Indicators Research 169(1-2): 341-364. Wirth, L. (1938). "Urbanism as a way of life." American Journal of Sociology 44(1): 1-24. Wolch, J. R., J. Byrne and J. P. Newell (2014). "Urban green space, public health, and environmental justice: The challenge of making cities ‘just green enough’." Landscape and Urban Planning 125: 234-244. Wurm, S., U. Ehrlich, F. Meyer-Wyk and S. M. Spuling (2023). "Prevalence of loneliness among older adults in Germany." Journal of Health Monitoring 8(3): 49. Yousaf, S. and A. Bonsall (2017). UK Townsend Deprivation Scores from 2011 census data. UK Data Service. http://statistics.digitalresources.jisc.ac.uk.s3.amazonaws.com/dkan/files/Townsend_Deprivation_Scores/UK%20Townsen d%20Deprivation%20Scores%20from%202011%20census%20data.pdf. Zajamšek, B., K. L. Hansen, C. J. Doolan and C. H. Hansen (2016). "Characterisation of wind farm infrasound and lowfrequency noise." Journal of Sound and Vibration 370: 176-190. | 98 9. Appendix Appendices list (Excel workbook): ▪ Appendix 1 (A2-1): List of publications included in the systematic review of rural area effects on subjective wellbeing and rural subjective wellbeing advantage in Europe in the empirical literature 2000-2023 ▪ Appendix 2 (A2-2): Rural area effects on subjective wellbeing and rural subjective wellbeing advantage in Europe in the empirical literature 2000-2023 ▪ Appendix 3 (A2-3): Rural area effects on subjective wellbeing and rural subjective wellbeing advantage in Europe in the empirical literature 2000-2023: Rural-urban classifications by geographical scale ▪ Appendix 4 (A3-1): Descriptive statistics of analytical sample - whole sample followed by welfare state groupand country-specific statistics ▪ Appendix 5 (A3-2): Population estimates - all countries followed by welfare state groupand country-specific statistics ▪ Appendix 6 (A3-3): Average subjective wellbeing scores (MPI) ▪ Appendix 7 (A3-4): Average subjective wellbeing scores (MPI) by settlement type ▪ Appendix 8 (A3-5_*): Multivariate regressions on life satisfaction (LS), Affective wellbeing (AW), Basic Psychological Needs Satisfaction (NS), Mental Resources (MR), Social wellbeing (SW), and Evaluative wellbeing (EW). ▪ Appendix 9 (A3-6_*): Country-specific multivariate regressions on subjective wellbeing (UK, NL, PL, ES, FR, D1, D2, PT) ▪ Appendix 10 (A4-1): Descriptive statistics of all variables used in the empirical analysis ▪ Appendix 11 (A4-2): Random Effects Generalised Least Squares Regressions of life satisfaction on rurality, area and individual characteristics, GB 2009-2023 ▪ Appendix 12 (A4-3): Random Effects Generalised Least Squares Regressions of life satisfaction on area and individual characteristics by rurality, GB 2009-2023 ▪ Appendix 13 (A5-1): Description of analytical samples used in wind turbine analysis ▪ Appendix 14 (A5-2): Propensity score matching statistics for all treatment groups ▪ Appendix 15 (A5-3): Fixed effects regressions of SF12 (MCS and PCS) on the presence of wind turbines, the number of wind turbines, and the inverse distance to wind turbines in a radius of 1 km to 6 km of residential homes. Germany 2002-2022 ▪ Appendix 16 (A5-4): Fixed effects difference-in-differences regressions of SF-12 (MCS and PCS) on the presence of wind turbines, the number of wind turbines, and the inverse distance to wind turbines within a radius of 1 km to 6 km of residential homes. Germany 2002-2022