| Domain | Category | Change | Effect on 0-10 Life Satisfaction | Dynamics over time | Confidence in effect and causality | Data type | Data source | Country | Reference | Location in paper | Comments |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Education | Duration | Extra year of compulsory education | -0.03 (± 0.098) converted from 1-7 to 0-10 LS | Persistent effects | High for UK; since effect found from 1972 UK compulsory school changes. Marginal result also found in other Western countries | Panel | BHPS 1996-2008 | 🇬🇧 | Clark & Jung, 2017 | Page 11, paragraph 1 (based on Table 3) | |
| Environment | Air pollution | Increase of 1-day SO₂ level by 10 μg m⁻³ (equivalent to 3.9 ppb) | -0.02 (± 0.02) on 5-point LS | Temporary effect | Effect robust in cross-sectional data; includes high-resolution geographic fixed effects. | Cross-sectional | CCHS 2005-11 | 🇨🇦 | Barrington-Leigh & Behzadnejad, 2017 | In text, bottom of page 16 of paper | |
| Increase of average SO₂ level by 10 μg m⁻³ (equivalent to 3.9 ppb) | -0.08 | Unknown | High; effects driven by unanticipated changes in power plant emissions due to policy | Panel | GSOEP 1983-2011 | 🇩🇪 | Luechinger, 2009 | Table 4, column II (IV estimate) | |||
| Increase of average PM10 level by 10 μg m⁻³ (equivalent to 3.9 ppb) | 0.014 on a 3-point happiness scale | Unknown | Medium to high; effects of air pollution significantly exogenous for single individual | Cross-sectional | GSS (USA) 1984-96 | 🇺🇸 | Levinson, 2012 | Results section paragraph 1 | |||
| Land use | Increase of 1 hectare of greenspace within 1km of household | +0.0066 (± 0.0049) | Seems permanent | Medium to high; panel data-based set but no clearcut exogenous variation; similar results by studies in the UK | Panel | GSOEP 2000-2012 | 🇩🇪 | Krekel et al, 2016 | Table B.2 | Effects strongest for older residents | |
| +0.0031 converted from 1-7 to 0-10 LS | Seems permanent | Medium to high; panel data-based set but no clearcut exogenous variation | Panel | BHPS 1991-2008 | 🇬🇧 | White et al, 2013 | 0.0020 in Table 2, Column 5 | Cited by / taken from DOHC in Frijters and Krekel...? | |||
| Increase of 1 hectare of vacant land (abandoned areas) within 1km of household | -0.0395 (± 0.0002) | Unknown | Medium; panel data-based but no clearcut exogenous variation | Panel | GSOEP 2000-2012 | 🇩🇪 | Krekel et al, 2016 | Table B.2 | Effects strongest for older residents | ||
| Construction of wind turbine within 4km around household | -0.1405 (±0.0782) | Seems temporary; effect disappears after 5 years | High; wind turbine construction exogenous for household in surroundings, difference-in-differences with treatment at multiple points in time | Panel | GSOEP 2000-2012 | 🇩🇪 | Krekel & Zerrahn, 2017 | Table 2, column 1 | |||
| Weather | Daily rainfall of 6mm above average | -0.008 (± 0.0012) on 5-point LS | Temporary effect | Effect is statistically significant and robust in cross-sectional dataset, but not in panel dataset | Cross-sectional and panel | CCHS 2005-11, NPHS 2004-10 | 🇨🇦 | Barrington-Leigh & Behzadnejad, 2017 | Table 2, Columns 7 and 8 | Women and individuals with poor health condition are more affected | |
| Work | Employment status | From employment to unemployment | −.054 (±0.022) on 5-point happiness-in-life | Short and long term effects | High. Panel data, fixed instrumental effects | Panel | NPHS 1994-2007, CCHS 2009-11 | 🇨🇦 | Latif, 2010 | Table 3, Column 2 | Not statistically significant for individuals aged 54 and older |
| -0.46 (±0.078) | Immediate effect higher, then reducing, but no adaptation | High. Large effects found in longitudinal cross-sections, recession-related and employment-shock related (plant closures) | Panel | GSEOP | 🇩🇪 | Flèche et al, 2019 | Table 4.2 | ||||
| -0.71 (±0.059) | Immediate effect higher then reducing, but no adaptation | Immediate effect higher then reducing, but no adaptation | Panel | BCS70 | 🇬🇧 | Flèche et al, 2019 | Table 4.2 | ||||
| From full-time employed to part-time employed wanting more hours | -0.108 (±0.016) | Largely permanent | Effect very robust in cross section and panels, but causality unclear | Cross-sectional and panel | GWP 2006-08 | 🇦🇺 🇨🇦 🇳🇿 🇺🇸 | De Neve & Ward, 2017 | Table 6.3, Column 8 "NA+ANZ" | Particularly strong effect for men | ||
| From full-time employed to part-time employed not wanting more hours | +0.080 (±0.043) | Largely permanent | Effect very robust in cross section and panels, but causality unclear | Cross-sectional and panel | GWP 2006-08 | 🇦🇺 🇨🇦 🇳🇿 🇺🇸 | De Neve & Ward, 2017 | Table 6.3, Column 8 "NA+ANZ" | Particularly strong effect for men | ||
| From unemployment to out-of-labour force | -0.23 (± 0.13) | Unknown | Cross-sectional data precludes causal claims | Cross-sectional | CCHS 2009-10 | 🇨🇦 | Shi et al, 2019 | Table 4.2 | |||
| From working to retired (at age 55 or older) | +0.056 (± 0.047) on 5-point happiness-in-life | Unknown | High. Panel data, fixed instrumental effects | Panel | NPHS 1994-2007 | 🇨🇦 | Latif, 2011 | Table 2, Column 4 | No significant effect for ages 45-54 | ||
| Job satisfaction | One unit change on 0-10 scale of non-financial job satisfaction | +0.15 (± 0.04) | Unknown | Cross sectional data but findings consistent between ESC and GSS data. Causality unclear. | Cross-sectional | GSS17, ESC2 | 🇨🇦 | Helliwell & Huang, 2010 | Table 1, Column 2 | Income effect instrumented for ESC data, adjusted in GSS data | |
| Type of job | Employment in an occupation that is below an individual’s skills or work experience (non-immigrants) | -0.280 (± 0.049) | Unknown | Cross-sectional data precludes causal claims | Cross-sectional | CCHS 2009-14 | 🇨🇦 | Hou & Frank, 2017 | Table 3, Column 2 | Lower income just one of the important factors for non-immigrants. | |
| Employment in an occupation that is below an individual’s skills or work experience (immigrants) | -0.055 (± 0.096) | Negative effect tends to diminish with increased length of stay in Canada | Cross-sectional data precludes causal claims | Cross-sectional | CCHS 2009-14 | 🇨🇦 | Hou & Frank, 2017 | Table 3, Column 4 | Lower income the main intermediate factor linking over-education to life satisfaction for immigrant | ||
| Being in a white collar job versus a blue collar job | Approx. +0.80 | Unknown | Effect very robust in cross-section and panels but causality unclear | Cross-sectional and panel | GWP 2006-08 | 🌐 | De Neve & Ward, 2017 | Approximated from job categories in Table 6.5 (?) | White collar includes: managers, officials, clerical and office workers; blue collar includes construction, transportation, farming | ||
| Commute | From no commute to 1 hour car commute | -0.012 (± 0.041) | Unknown | Low. Findings disputed and causality unclear. | Panel | BHPS 1996-2008 | 🇬🇧 | Dickerson et al, 2014 | Table 2, Column 2 | ||
| -0.20 (± 0.098) | Unknown | Low. Findings disputed and causality unclear. | Panel | GSOEP 1985-2003 | 🇩🇪 | Stutzer & Frey, 2008 | Table 1, Column 2 | ||||
| Increase in commute (by ???) | -0.18 (± 0.1176) on 10-point LS | Unknown | Low. Unclear units on time allocation commuting variable . | Cross-sectional | GSS 24 | 🇨🇦 | Hilbrecht et al, 2014 | Table 12, Column 2 | Particularly strong effect for women; Significant indirect effects for time spent in physically active leisure and seriousness of traffic congestion | ||
| Work conditions | Flexible work hours | +0.19 (± 0.1176) | Unknown | Cross-sectional data precludes causal claims | Cross-sectional | GSS 24 | 🇨🇦 | Hilbrecht et al, 2014 | Table 12, Column 3 | ||
| Finances | Income | Doubling of household income | +0.16 (± 0.196) | Persistent effects with elation peak | High. Effect found in panels, cross-sections, and shock-related (lotteries). | Panel | BCS70 | 🇬🇧 | Flèche et al, 2019 | Table 2.1 | Height disputed and income measurement problematic. |
| +0.5 | Persistent effects with elation peak | High. Effect found in panels, cross-sections, and shock-related (lotteries). | Panel | GSOEP 1991-2001 | 🇩🇪 | Frijters et al, 2004 | Table 2 | ||||
| Increase in difference between own log income and log income of a provincial reference group | +0.194 (± 0.135) | Unknown | Medium. Panel data, significant negative effect as found in other Canadian literature. | Panel | NPHS 1994-2009 | 🇨🇦 | Latif, 2016 | Table 5, Column 2 | Reference group contains all individuals with a similar education level that are inside the same age bracket and residing in the same province | ||
| Financial satisfaction | High financial stress (self-rated) | −0.864 (±0.086) | Unknown | Cross-sectional data, considering the possibility of an indirect effect of income through financial stress uncovers a strong effect of financial stress on life satisfaction, but an effect not clearly linked to income | Cross-sectional | GSS 19-24 | 🇨🇦 | Brzozowski & Spotton, 2020 | Table 2, Column 2 | Measurement includes those who report 3 or higher on a 5-point stress scale and also choose "finances" as their primary source of stress | |
| Prosocial spending | Donated to charity in the past month | +0.28 (±0.047) on 11-point Cantril ladder | Unknown | Cross-sectional data, relies on correlational analysis, supported by limited experimental data | Cross-sectional and panel | GWP 2006-08 | 🇦🇺 🇨🇦 🇳🇿 🇺🇸 | Aknin et al, 2013 | Region-specific coefficient using survey results from US, Canada, Australia, NZ | ||
| +0.27(±0.039) on 11-point Cantril ladder | Unknown | Cross-sectional data, relies on correlational analysis, supported by limited experimental data | Cross-sectional and panel | GWP 2006-08 | 🌐 | Aknin et al, 2013 | |||||
| Health | Physical health | From excellent to poor physical health (self-rated) | -2.19 (± 0.17) | Unknown | Cross-sectional data precludes causal claims | Cross-sectional | CCHS 2009-10 | 🇨🇦 | Shi et al, 2019 | Table 2, Column 1 | Obtained from control variables |
| From healthy to poor physical health (self-rated) | -1.080 (± 0.122) | Permanent effect, with initial peak | High as found everywhere, including to health shocks. | Panel | NCDS 1958-2009 | 🇬🇧 | Frijters et al, 2014 | Table 4, column 2 | |||
| -0.96 | Permanent effect, with initial peak | High as found everywhere, including to health shocks. | Panel | GSOEP 1983-2011 | 🇩🇪 | Ferrer-i-Carbonell & Frijters, 2004 | Unclear but likely taken from Table 3. See additional comments column | Based on a 3-point change in a 1-5 self-report measure of physical health | |||
| Satisfied with health status, at age 60 or older | +0.292 (±0.059) on 10-point LS | Unknown | Medium. Cross-sectional data precludes causal claims, yet findings are consistent with many studies suggesting health is the strongest single predictor of late-life SWB | Cross-sectional | WVS 2005-07 | 🇨🇦 🇬🇧 🇳🇿 🇺🇸 | Zelikova, 2013 | Table 2, Column 7 | |||
| Smoking | From smoking daily to not at all | +0.12 (± 0.04) | Unknown | Cross-sectional data precludes causal claims | Cross-sectional | CCHS 2009-10 | 🇨🇦 | Shi et al, 2019 | Table 2, column 1 | Obtained from control variables | |
| Nutrition | From 0 to 8 portions of fruit and vegetables a day | +0.16 (±0.08) | Unknown | Cross-sectional data precludes causal claims | Cross-sectional | CCHS 2009-10 | 🇨🇦 | Shi et al, 2019 | Table 2, column 2 | ||
| +0.24 (±-0.03) | Effect lasts while treatment lasts | Medium. Fixed-effect estimates consistent with small RCTs and public health campaign results, but magnitude very unclear | Panel | HILDA 2007, 2009 | 🇦🇺 | Mujcic & J.Oswald, 2016 | Table 2, column 1 and 2; in text near beginning of page 3 | ||||
| Mental health | From depression to full mental health | +0.71 | Permanent, little evidence of a peak | High as found everywhere, including large clinical trials | Panel | BHPS | 🇬🇧 | Flèche et al, 2019 | Table 16.2 | Based on 4-point change on a 0-12 scale | |
| From excellent to poor mental health (self-rated) | -3.13 (±0.30) | Unknown | Cross-sectional data precludes causal claims | Cross-sectional | CCHS 2009-10 | 🇨🇦 | Shi et al, 2019 | Obtained from control variables | |||
| Social capital | Friendships | From 0 close friends to 3-5 close friends | +0.241 (±0.017) on 10-point LS | Unknown | Cross sectional data; consistent with broader literature | Cross-sectional | GSS17 | 🇨🇦 | Helliwell & Wang, 2011 | Table 3, Column 1 | Impact is much smaller for those who are married or living with a partner, suggesting friends and spouses provide some similar happiness benefits |
| Seeing close friends more frequently | +0.096 (±0.051) on 10-point LS | Unknown | Cross-sectional data precludes causal claims, but consistent with | Cross-sectional | GSS17 | 🇨🇦 | Helliwell & Wang, 2011 | Table 3, Column 4 | Frequency of visits with family and especially with friends add significantly to LS above and beyond the effects of having such networks in place | ||
| From 0 close relatives to 3-5 close relatives | +0.526 (±0.149) on 10-point LS | Unknown | Cross sectional data; consistent with broader literature | Cross-sectional | GSS17 | 🇨🇦 | Helliwell & Wang, 2011 | Table 3, Column 1 | Paper includes several categories of numbers of close relatives (1 or 2, 3-5, 6-10, 11-20, over 20), an increase from one category to the next is about 0.15 | ||
| Seeing close relatives more frequently | +0.096 (±0.051) on 10-point LS | Unknown | Cross sectional data; consistent with broader literature | Cross-sectional | GSS17 | 🇨🇦 | Helliwell & Wang, 2011 | Table 3, Column 1 | Frequency of visits with family add significantly to LS above and beyond the effects of having the network in place | ||
| Can count on friends | +0.414 (±0.090) on 11-point Cantril ladder | Unknown | Low. Cross sectional data with regional effects; causality unclear | Cross-sectional | GWP 2006 | 🌐 | Helliwell & Wang, 2011 | Comes from Y/N response to question: "If you were in trouble, do you have relatives or friends you can count on to help you whenever you need them, or not?" | |||
| Romantic relationships | From single to married/partnered | +0.28 (±-0.10) | Permanent effect with initial peak | High. Ubiquitous finding around the world | Panel | BHPS | 🇬🇧 | Flèche et al, 2019 | Table 5.2 | ||
| +0.1 | Permanent effect with initial peak | High. Ubiquitous finding around the world | Panel | GSOEP 1983-2011 | 🇩🇪 | Ferrer-i-Carbonell & Frijters, 2004 | Taken from Frijters and Krekel's table-- not exactly sure where this coefficient came from. Maybe Column 1: fixed effect ordered logit 0.08 in Table 3 ? | ||||
| +0.60 (±0.022) | Unknown | High. Panel data, fixed instrumental effects | Panel | NPHS 1994-2007, CCHS 2009-11 | 🇨🇦 | Latif, 2010 | Table 3, Column 2 | ||||
| From never married to married at 50 or older | +0.20 (±-0.078) | Permanent effect with high initial peak | Medium: cohort study findings so causality unclear | Panel | BHPS | 🇬🇧 | Flèche et al, 2019 | Table 9.1 | |||
| Never married, age 60 or older | -0.122 (±-0.078) | Unknown | Medium. Cross-sectional data precludes causal claims, yet consistent with broader literature as found widely | Cross-sectional data | WVS 2005-07 | 🇨🇦 🇬🇧 🇳🇿 🇺🇸 | Zelikova, 2013 | Table 2, Column 7 | |||
| From partnered to separated | -0.40 (±-0.14) | High intial effect, then some adaptation | High as found everywhere. | Panel | BHPS | 🇬🇧 | Flèche et al, 2019 | Table 5.2 | Note that most find new partners and don't stay separated. Lone men suffer more. | ||
| Immigration | Being an immigrant parent (female) | -0.210 (±0.106) on 5-point LS | No apparent improvement over time, "years since arrival" variable is statistically insignificant | Medium. Cross sectional data, effect persists with controls for personal characteristics such as ethnicity, income, etc; consistent with broader literature | Cross-sectional | CCHS 2002-10 | 🇨🇦 | Burton & Phipps, 2010 | Table 5, Column 3 | No statistically significant effect for female immigrant children once mediating variables (language, ethnicity) are added | |
| Being an immigrant parent (male) | -0.218 (±0.133) on 5-point LS | No apparent improvement over time, "years since arrival" variable is statistically insignificant | Medium. Cross sectional data, effect persists with controls for personal characteristics such as ethnicity, income, etc; consistent with broader literature | Cross-sectional | CCHS 2002-10 | 🇨🇦 | Burton & Phipps, 2010 | Table 5, Column 4 | No statistically significant effect for female immigrant children once mediating variables (language, ethnicity) are added | ||
| Discrimination | Experience religious discrimination | -0.39 | Unknown | Cross-sectional data precludes causal claims | Cross-sectional | GSS27 | 🇨🇦 | Vang et al, 2019 | Table 4, Column 2 | Significant positive interaction term suggests higher religiosity mitigates the negative effect of religious discrimination | |
| Trust | Social trust (self-reported trust in "most people") | +0.131 on 10-point LS | Unknown | Cross-sectional data precludes causal claims; statistically significant positive effect on life satisfaction and domain satisfaction in all domains | Cross-sectional | GSS17 | 🇨🇦 | van der Horst & Coffé, 2012 | Table 3, Column 1 | Social trust measured by a binary variable where 0 is ‘‘one cannot be too careful in dealing with people’’ and 1 is ‘‘most people can be trusted’’. | |
| Believe a lost wallet is likely to be returned if found by neighbours | +0.172 (±0.088) on 10-point LS | Unknown | Cross sectional data; consistent with GWP findings and broader literature | Cross-sectional | GSS17 | 🇨🇦 | Helliwell & Wang, 2011 | Table 3, Column 3 | Respondents who live in high-density census tracts and are highly mobile are less likely to believe a neighbour would return their wallet | ||
| 0.117 (±0.088) on 11-point Cantril ladder | Unknown | Medium. Cross sectional data includes regional fixed effects; generally consistent with broader literature | Cross-sectional | GWP 2006 | 🌐 | Helliwell & Wang, 2011 | Table 2-a, Column 6 | ||||
| Believe a lost wallet is likely to be returned if found by police | 0.138 (±0.094) on 11-point Cantril ladder | Unknown | Medium. Cross sectional data includes regional fixed effects; generally consistent with broader literature | Cross-sectional | GWP 2006 | 🌐 | Helliwell & Wang, 2011 | Table 2-b, Column 6 | |||
| Believe a lost wallet is likely to be returned if found by a stranger | +0.074 (±0.098) on 11-point Cantril ladder | Unknown | Low. Cross sectional data includes regional fixed effects; but effect is statistically insignificant. | Cross-sectional | GWP 2006 | 🌐 | Helliwell & Wang, 2011 | Table 2-c, Column 6 | |||
| +0.237 (±0.098) on 10-point LS | Unknown | Cross sectional data precludes causal claims but is consistent with GWP findings and broader literature | Cross-sectional | GSS17 | 🇨🇦 | Helliwell & Wang, 2011 | Table 3, Column 3 | ||||
| Trust in neighbours | +0.336 (±0.140) on 10-point LS | Unknown | Cross sectional data precludes causal claims but is consistent with broader literature on community-level trust | Cross-sectional | GSS17 | 🇨🇦 | Helliwell & Wang, 2011 | Table 3, Column 5 | Respondents who live in high-density census tracts and are highly mobile are less likely to trust their neighbours | ||
| Trust in co-workers | +0.638 (±0.149) on 10-point LS | Unknown | Cross sectional data precludes causal claims; | Cross-sectional | GSS17 | 🇨🇦 | Helliwell & Wang, 2011 | Table 3, Column 5 | |||
| Confidence in police | +0.361 (±0.114) on 10-point LS | Unknown | Cross sectional data precludes causal claims | Cross-sectional | GSS17 | 🇨🇦 | Helliwell & Wang, 2011 | Table 3, Column 5 | |||
| Belonging | Sense of belonging to the community | +0.781 (±0.110) on 10-point LS | Unknown | Cross sectional data precludes causal claims but is consistent with broader literature suggesting community-level belonging is most important | Cross-sectional | GSS17 | 🇨🇦 | Helliwell & Wang, 2011 | Table 3, Column 5 | A sense of belonging to one’s community is strongly associated with neighbourhood trust | |
| Sense of belonging to the province | +0.274 (±0.114) on 10-point LS | Unknown | Cross sectional data precludes causal claims | Cross-sectional | GSS17 | 🇨🇦 | Helliwell & Wang, 2011 | Table 3, Column 5 | |||
| Sense of belonging to Canada | +0.336 (±0.137) on 10-point LS | Unknown | Cross sectional data precludes causal claims | Cross-sectional | GSS17 | 🇨🇦 | Helliwell & Wang, 2011 | Table 3, Column 5 | A sense of belonging to Canada is strongly associated with general social trust | ||
| Crime | Violent crime | Victim of violent crime | -0.396 | Effect largely in first year (only statistically significant in first year) | High but specific: effects are for unanticipated events that were recorded | Panel | HILDA 2002-12 | 🇦🇺 | Johnston et al, 2018 | Table 3 (?) Effect of -0.398 for females and -.300 for males | |
| Fear | A doubling fear of crime | Approx -0.30 | Unknown | Medium. Panel data-based, often replicated, but drivers of fear not exogenous | Panel | Nationwide representative study on victimization and crime-related issues, 2010 | 🇩🇪 | Hanslmaier, 2013 | "derived from the relative effect of fera of crime versus effect from unemployment in a log-odds setting" (note on this reference in Frijters handbook) | Derived from relative effect of fear of crime versus effect from unemployment in a log-odds setting |
This project was made possible by the Social Sciences and Humanities Research Council of Canada (supporting Barrington-Leigh) and the Arts Research Internship Award Programme at McGill University (supporting Katja Lemermeyer).
For how to cite this database, and for the method and principles behind it, see Barrington-Leigh and Lemermeyer, "A public, open, and independently-curated database of happiness coefficients", Journal of Happiness Studies, doi 10.1007/s10902-023-00652-4, April 2023.
This is a demonstration effort and should be considered a draft. Questions and comments should be sent to Chris Barrington-Leigh and Katja Lemermeyer, McGill University, September 2020--.