Prediction vs. Explanation:
With the advent of Big Data and data mining, statistical methods like regression and CART have been repurposed to use as tools in predictive modeling. When statistical models are used as a tool of research, the goal is to explain relationships in a dataset, and make inference beyond the specific data to shed light on a phenomenon or problem. In this context, how well the model fits the data is important, as is any indication (through diagnostics) that some important relationship or variable may have been omitted. When statistical methods are used in data mining, the goal is to predict values in new data. The important metric in this context is how well the model does in making its predictions; traditional statistical metrics like goodness-of-fit are not as relevant.