Last updated 16-Nov-95 [http://www.stat.uchicago.edu/~thisted/courses/227/sel412.html]



There is an error in Selvin's formal statistical test for the change in due to the addition of q additional predictors to a regression model. The F statistic in (4.12) should read



The denominator is in error in the book. This ratio assesses the increase in relative to the proportion of unexplained variation remaining in the full model.