The recent article by Jones (1984) discussing use of multiple-comparison procedures in journals of the Entomological Society of America seemed to place undue emphasis on significance testing at the expense of estimation. I demonstrate why multiple-comparison procedures are inappropriate for factorial experiments or treatments forming a progressive series. I suggest that these procedures are not helpful even for totally unstructured data. Some guidelines for statistical analysis are proposed and alternative approaches examined.
No takes yet. Share an insight, caveat, or question.
J. N. Perry (1986) studied this question.