The regression theoretical foundation and the tool of significance testing employed on big data are without statistical binding force. Thus, fitting big data to a prespecified small-framed model produces a skewed model with doubtful interpretability and questionable results.
Statistical and Machine-Learning Data Mining
Techniques for Better Predictive Modeling and Analysis of Big Data
by Bruce Ratner
10 Variable Selection Methods in Regression: Ignorable Problem, Notable Solution
10.2 Background
(P. 178)