Disentangling Selection from Causation in the Empirical Association between Crime and Adolescent Work