August 1990

ANOVA, MANOVA, My Black HenComments on Repeated Measures

Author Affiliations

From Massachusetts General Hospital, Harvard Medical School, Boston.

Arch Gen Psychiatry. 1990;47(8):775-778. doi:10.1001/archpsyc.1990.01810200083012

The article by Ekstrom et al1 has two very different messages: statistical reporting needs to be improved, and a common method of analysis (repeated-measures analysis of variance [RM ANOVA]) appears to be improperly used, at the expense of a possibly more appropriate technique (multi

See also p 770. variate ANOVA [MANOVA]). The reader whose eyes do not immediately light up at the thought of a bare-knuckle statistical brawl may want to know what the fuss is about. Who needs all those details, and why should I care if my statistician takes his or her ANOVA straight, without an M?

WHY REPORT THE DETAILS?  In textbooks and classrooms, methodologists remind future experimentalists that the design of investigations and analysis of data work like the two oars of a rowboat: neglect one and you go in circles. To design is to plan, to have an intention and a goal. When it is

First Page Preview View Large
First page PDF preview
First page PDF preview