Working Paper: NBER ID: w24592
Authors: John F. Helliwell; Hugh Shiplett; Christopher P. Barrington-Leigh
Abstract: This paper presents a new public-use dataset for community-level life satisfaction in Canada, based on more than 400,000 observations from the Canadian Community Health Surveys and the General Social Surveys. The country is divided into 1215 similarly sampled geographic regions, using natural, built, and administrative boundaries. A cross-validation exercise suggests that our choice of minimum sampling thresholds approximately maximizes the predictive power of our estimates. Our procedure reveals robust differences in life satisfaction between and across urban and rural communities. We then match the life satisfaction data with a range of key census variables to explore ways in which lives differ in the most and least happy communities. The data presented here are useful on their own to study community-level variation, and can also be used to provide contextual variables for multi-level modelling with individual life satisfaction data set in a community context.
Keywords: life satisfaction; community health surveys; subjective wellbeing; Canada; demographic factors
JEL Codes: C81; I31; R12
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| community characteristics (R23) | life satisfaction (I31) |
| happier communities (R23) | community belonging and social trust (Z13) |
| demographic and economic factors (J11) | life satisfaction outcomes (I31) |
| urban areas (R11) | life satisfaction (I31) |
| community characteristics (R23) | average life satisfaction (I31) |