World Happiness Report 2023 4 • U sing the five characteristics (and income), it is possible to classify states into 3 clusters: common-interest states, special-interest states and weak states. In common-interest states, average life satisfaction is 2 points (out of 10) higher than in weak states and in special-interest states it is 1 point higher than in weak states. • I n those countries where average life satisfaction is highest, it is also more equally distributed – with fewer citizens having relatively low life satisfaction. Chapter 4. Doing Good and Feeling Good: Relationships between Altruism and Well-being for Altruists, Beneficiaries, and Observers • A person is being altruistic when they help another person without expecting anything in return. Altruistic behaviours like helping strangers, donating money, giving blood, and volunteering are common, while others (like donating a kidney) are less so. • There is a positive relationship between happiness and all of these altruistic behaviours. This is true when we compare across countries, and when we compare across individuals. But why? • Normally, people who receive altruistic help will experience improved well-being, which helps explain the correlation across countries. But in addition, there is much evidence (experimental and others) that helping behaviour increases the well-being of the individual helper. This is especially true when the helping behaviour is voluntary and mainly motivated by concern for the person being helped. • The causal arrow also runs in the opposite direction. Experimental and other evidence shows that when people’s well-being increases, they can become more altruistic. In particular, when people’s well-being rises through experiencing altruistic help, they become more likely to help others, creating a virtuous spiral. Chapter 5. Towards Reliably Forecasting the Well-being of Populations Using Social Media: Three Generations of Progress • Assessments using social media can provide timely and spatially detailed well-being measurement to track changes, evaluate policy, and provide accountability. • Since 2010, the methods using social media data for assessing well-being have increased in sophistication. The two main sources of development have been data collection/ aggregation strategies and better natural language processing (i.e., sentiment models). • Data collection/aggregation strategies have evolved from the analysis of random feeds (Generation 1) to the analyses of demographically- characterized samples of users (Generation 2) to an emerging new generation of digital cohort design studies in which users are followed over time (Generation 3). • Natural Language Processing models have improved mapping language use to well-being estimates – progressing from counting dictionaries of keywords (Level 1) to relying on robust machine-learning estimates (Level 2) to using large language models that consider words within contexts (Level 3). • The improvement in methods addresses various biases that affect social media data, including selection, sampling, and presentation biases, as well as the impact of bots. • The current generation of digital cohort designs gives social media-based well-being assessment the potential for unparalleled measurement in space and time (e.g., monthly subregional estimation). Such estimates can be used to test scientific hypotheses about well-being, policy, and population health using quasi-experimental designs (e.g., by comparing trajectories across matched counties).
RkJQdWJsaXNoZXIy NzQwMjQ=