Identifying Worker and Workplace Predictors of Life Satisfaction in Age-Sex Sub- Populations: An “Emelometrics” Approach

The paper applies a blend of econometric and machine learning techniques to individual-level data from the 2008 National Study of the Changing Workforce to identify covariates that appear to predict life satisfaction across four age-sex sub-populations; specifically, females age < 49; females age 49+; males age < 49; and males age 49+. Covariates such as logged absolute earnings, stress, sleep troubles, job satisfaction, marital status, depression, health status, job physicality, and some types of workplace flexibilities and are among the strongest predictors of life satisfaction across the sub-populations. The findings largely prevail though a wide array of sensitivity analyses, including alternative specifications and alternative estimation strategies, among them exhaustive-search algorithms and variable importance results from random forest algorithms. Prediction results from confusion matrices are reported for all binary-response models alongside coefficient estimates (where possible); the logit models delivered the best prediction accuracy. JEL Classifications: J28, J30, J32, J,63, J81