Using two completely different high-dimensional but small sample datasets - one financial and one in health care - we document that the empirical methodology used, to determine key explanatory variables, when there are a large number of them, to explain outcome(s), can influence the findings.