Things I'd like to revise: 1. passing of alphalevel around 2. repetitive computing of outlying observations in display.delstats, summary.gvlmaDel