In this era of big data and complex statistical modeling, scientists use sophisticated computational tools to search through a large number of models, looking for meaningful patterns. The challenge is then to judge the strength of a large number of apparent associations that have been found. This statistical problem has become known as \Post-selection inference," the assessment of significance and effect sizes from a data-set after mining the same data to find these associations. In this talk I will discuss new methods for computing p-values and confidence intervals in regression, that correctly account for the adaptive selection of the model.

This is joint work with Jonathan Taylor, Ryan Tibshirani, Will Fithian and Richard Lockhart.