Well-Being Measures at Scale for Policy Research
(The blog had been on hiatus during 2019 as CREOS launched. This post is part of a series of catch-up blog post summarizing talks presented over the last 10 months.)
The Dubai Future Foundation invited me to lead a conversation on the significance of the concept of well-being for social-science and policy. These slides framed the conversation, and summarize of the subsequent discussion of opportunities and challenges of measuring well-being at scale.
Human well-being has four major dimensions: affective, health, economic, and ethical/political. Policy analysis typically focuses on only one of these dimensions, and measured coarsely — e.g. ranking countries by GDP, while ignoring the distribution of wealth (not to mention health, freedom, and life-satisfaction).
Focusing on GDP and other incomplete measures of well-being limits often leads to the wrong conclusion about, what makes good policy. For example in a paper we published recently The Happiness-Energy Paradox: Energy Use is Unrelated to Subjective Well-Being we find that although energy use does have a limited relationship with GDP it does correlate with subjective well-being. This suggests that conservation is compatible with direct national-interest in developed countries, and highlights the need for more sophisticated measures of policy outcomes.
One challenge to making policy analysis more sophisticated is measuring well-being at scale. The discussion highlighted a number of existing sources of big data that could be repurposed for broad use well-being measurements:
- Facial expression
- … short term subjective measures (e.g. happiness)
- … measures of durable emotional states (e.g. anxiety/depression)
- Quantified self systems, health mobile apps for
… extensive collection of self-evaluative measure - Public social profiles to …
… measure emotive statements and communication patterns indicative of well-being
… detect life-events - Administrative and financial data …
… detect life events - Always-on listening devices (e.g. alexa) to measure ….
… voice tone, voice emotion, conversational patterns - Location tracking
… movement variability proxies for well-being
For those interested in these questions and related areas of interest, writing on modern approaches to privacy principles and protectins are linked from my web site.