1. Measuring the predictability of life outcomes with a scientific mass collaboration
- Author
-
Kristin E. Porter, Malte Möser, Flora Wang, Bingyu Zhao, Wei Lee Woon, Yoshihiko Suhara, Adaner Usmani, Erik H. Wang, Kun Jin, Samantha Weissman, William Eggert, Hamidreza Omidvar, Andrew Or, Lisa M Hummel, Gregory Faletto, Ben Sender, Qiankun Niu, Viola Mocz, Antje Kirchner, Catherine Wu, Karen Ouyang, Ian Lundberg, Allison C. Morgan, Abdulla Alhajri, Arvind Narayanan, Khaled AlGhoneim, Louis Raes, Ilana M. Horwitz, Barbara E. Engelhardt, Ben Leizman, Crystal Qian, Drew Altschul, Guanhua He, Jeanne Brooks-Gunn, Ridhi Kashyap, Eaman Jahani, Ryan James Compton, Anna Filippova, Sara McLanahan, Tejomay Gadgil, Claudia V. Roberts, Muna Adem, Julia Wang, Jeremy Freese, Alexander T. Kindel, Daniel E Rigobon, Naijia Liu, Lisa P. Argyle, Mayank Mahajan, Jonathan D Tang, Moritz Hardt, Ethan Porter, Diana Mercado-Garcia, Andrew Halpern-Manners, Anahit Sargsyan, Duncan J. Watts, Alex Pentland, Sonia P Hashim, Dean Knox, Onur Varol, Ryan Amos, James M. Wu, Thomas Davidson, Emma Tsurkov, Bernie Hogan, Areg Karapetyan, William Nowak, Jingwen Yin, Livia Baer-Bositis, Landon Schnabel, Chenyun Zhu, Noah Mandell, Ahmed Musse, Yue Gao, Josh Gagné, Stephen McKay, Jennie E. Brand, Abdullah Almaatouq, Katy M. Pinto, Andrew E Mack, Austin van Loon, Bedoor K. AlShebli, Helge Marahrens, Xiafei Wang, Bryan Schonfeld, Sonia Hausen, Kengran Yang, Maria Wolters, Brandon M. Stewart, Naman Jain, Moritz Büchi, Nicole Bohme Carnegie, Redwane Amin, Caitlin Ahearn, Kirstie Whitaker, Bo-Ryehn Chung, Diana Stanescu, Thomas Schaffner, Patrick Kaminski, David Jurgens, Kivan Polimis, Kimberly Higuera, Zhilin Fan, Matthew J. Salganik, Debanjan Datta, Connor Gilroy, E H Kim, Katariina Mueller-Gastell, Karen Levy, Brian J. Goode, Zhi Wang, Tamkinat Rauf, Lundberg, Ian [0000-0002-1909-2270], Almaatouq, Abdullah [0000-0002-8467-9123], Altschul, Drew M [0000-0001-7053-4209], Carnegie, Nicole Bohme [0000-0001-7664-6682], Kashyap, Ridhi [0000-0003-0615-2868], McKay, Stephen [0000-0002-5080-8417], Morgan, Allison C [0000-0003-2926-2162], Raes, Louis [0000-0003-2640-7493], Argyle, Lisa P [0000-0003-3109-2537], Büchi, Moritz [0000-0002-9202-889X], Jin, Kun [0000-0002-0118-1021], Varol, Onur [0000-0002-3994-6106], Watts, Duncan J [0000-0001-5005-4961], Apollo - University of Cambridge Repository, Department of Economics, and Research Group: Economics
- Subjects
Male ,Adolescent ,Computer science ,Social Sciences ,02 engineering and technology ,Cohort Studies ,Machine Learning ,03 medical and health sciences ,0302 clinical medicine ,Life ,Multidisciplinary approach ,Predictive Value of Tests ,Benchmark (surveying) ,0202 electrical engineering, electronic engineering, information engineering ,Econometrics ,Humans ,Family ,Predictability ,mass collaboration ,Child ,life course ,Multidisciplinary ,Fragile Families ,Correction ,Infant ,Fragile Families and Child Wellbeing Study ,prediction ,Outcome (probability) ,Mass collaboration ,machine learning ,L310 Applied Sociology ,Child, Preschool ,Computational Social Science ,Life course approach ,020201 artificial intelligence & image processing ,Computational sociology ,Female ,030217 neurology & neurosurgery - Abstract
© This open access article is distributed under Creative Commons Attribution-NonCommercialNoDerivatives License 4.0 (CC BY-NC-ND). 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.
- Published
- 2020
- Full Text
- View/download PDF