Abstract
Considerable attention is typically given to Type I and Type II errors when conducting empirical research. This article presents an error, often ignored in marketing and consumer behavior research, termed Type IV error. This error results from the improper investigation of interactions in an analysis of variance. A review of research published in Journal of Marketing Research and Journal of Consumer Research found widespread occurrence of Type IV errors. Illustrative improper interpretations of interactions are discussed and approaches for properly investigating interactions are presented. Situations where interactions need to be tested and interpreted are noted. These situations are contrasted with those where it is more appropriate to examine cell mean differences. The correct use of graphs of cell means is also discussed. Guidelines are presented for avoiding Type IV errors.
Original language | English (US) |
---|---|
Pages (from-to) | 17-26 |
Number of pages | 10 |
Journal | Journal of the Academy of Marketing Science |
Volume | 24 |
Issue number | 1 |
DOIs | |
State | Published - Jan 1 1996 |