ABSTRACT
Traditional statistical models and computational models offer fundamentally different, but complementary, approaches to understanding adolescent development. While statistical models summarize data to test theories, computational models generate data that would result if a given operationalization of a theory were true. In addition, they treat time fundamentally differently, with most statistical models using time as a variable anchored at baseline (Time 0), but computational models operationalizing time in stepwise measure with a new baseline derived at each iteration. Four models are used to illustrate the usefulness of computational models to developmental researchers.
