James Heckman has written many papers (and see this) about the importance of developing non-cognitive "soft skills" in young children. He argues that such early age investments can build up these skills and that these skills pay off in terms of higher future wages. A standard approach for measuring such soft skills is the Big Five Personality test. Such a test measures you on; 5 major dimensions of personality: Openness, Conscientiousness, Agreeableness, Extraversion, and Neuroticism.
I would like to make a research suggestion in this Big Data age. Each Uber customer has a personal Uber rating. I propose that researchers use this as a "Crowd Sourced" Big 5 personality test. Those of us with high scores are charming individuals who are agreeable and conscientious!
As you know, the University of California's salaries are in the public domain. A hard working graduate student could take each UC professor and staff member's salary and link by name and city (i.e Matthew Kahn of Los Angeles) to our Uber data. Uber knows our name and our location and our Uber rating. By merging these data, one could then run a regression including zip code of residence fixed effects (Uber knows our home address and thus our zip code) and you could run regressions of the form;
Annual income of UC Employee = controls + b*Uber score + U
This would be a fun and easy paper to write. Is b greater than zero?