ABSTRACT
How predictable are life trajectories? We investigated this question with a scientific mass collaboration using the common task method; 160 teams built predictive models for six life outcomes using data from the Fragile Families and Child Wellbeing Study, a high-quality birth cohort study. Despite using a rich dataset and applying machine-learning methods optimized for prediction, the best predictions were not very accurate and were only slightly better than those from a simple benchmark model. Within each outcome, prediction error was strongly associated with the family being predicted and weakly associated with the technique used to generate the prediction. Overall, these results suggest practical limits to the predictability of life outcomes in some settings and illustrate the value of mass collaborations in the social sciences.
Subject(s)
Social Sciences/standards , Adolescent , Child , Child, Preschool , Cohort Studies , Family , Female , Humans , Infant , Life , Machine Learning , Male , Predictive Value of Tests , Social Sciences/methods , Social Sciences/statistics & numerical dataABSTRACT
An emerging idea in psychopathology conceives of disorders as networks of mutually-reinforcing symptoms that constitute the disorder rather than simply reflect it. This is similar to how social scientists already view socioeconomic status, and has affinities to how physical health problems compound in later life. Social, psychological, and physical conditions might therefore be thought of as networks of problems with 'causal bridges' that span different levels and bring low SES, mental health challenges, and physical health problems into pervasive relationships with one another. The network view suggests a more heterogeneous and less reductive view on genetic causes which accords with the highly diffuse causal architecture now known to be the hallmark of complex behaviors and traits.
