Naked Statistics: Stripping the Dread from the Data

by Charles Wheelan

If you put enough junk variables in a regression equation, one of them is bound to meet the threshold for statistical significance just by chance.