Domain and Indicator Selection
Eight domains were preselected based on expertise, team discussions, and relevance to human progress.
EXPERT SELECTION
Participants from various professional organizations were asked to nominate experts with optimistic, pessimistic, and realistic perspectives on the future of human welfare and societal change.
Gathered from different corners of the globe, 28 distinguished scholars nominated 68 experts on human welfare. Thirty-four experts were selected from the 68 based on expertise and ensuring discipline diversity in the sample.
Domain and Indicator Selection
Eight domains were preselected based on expertise, team discussions, and relevance to human progress.
Preselected indicators for each domain were chosen based on specific criteria including indicator specificity, global accessibility of data, data quality, and temporal resolution. Indicators and data sources needed to meet all criteria to be considered reliable.
Is the indicator concretely definable and does it reveal enough specific information to allow for measurement ? Successful indicators were ranked higher if they directly provided method of measurement (e.g., “number of”, “rate of”, etc.)
Is there any data for this indicator publicly available to indicate changes to human welfare?
Does the source provide data for at least one-third of all countries (i.e. minimum of sixty-five countries) and does each country report data at least from the last ten years? Indicators that represent data representative of just the world (e.g., size of smallest chips in mass production) were treated as passing this criteria.
Data transparency: Is there accessible metadata explaining methods of data aggregation and estimation for missing data?
Composition transparency: Is there accessible metadata explaining the relationships of sub-indicators for syntactic indicators (i.e., indicators that do not exist as real-world phenomena but are derived from multiple sub-indicators)? For example, socioeconomic status is a syntactic variable that can be derived from income, educational attainment, pre-existing wealth, etc. If the indicator did not represent a syntactic variable, it automatically passed this criteria.
Is the data for this indicator systematically updated (e.g., on an annual basis) or are there major gaps in temporal resolution with no data for recent years?
24 experts individually ranked the eight preselected domains by perceived importance. They selected 4 domains as the most important.
Furthermore, they voted on the most reliable markers for each of these domains.
Experts, superforecasters, and everyday people made predictions for each of these four domains.
Learn moreCopyright © 2025 Forecasting Human Welfare.
Powered by Orotron