I am grateful to Bi and Meyners for their comments on my letter and to the editors of FQ&P for allowing this discussion on equivalency testing. I am glad that the replies to my letter have opened up some of the questions concerning equivalency testing methods which were not adequately discussed by Bi in his two papers (2005 and 2007), one of which strongly advocated the TOST. Notwithstanding his acknowledgment of my criticism, Bi remains committed: ‘‘In my opinion, only the TOST, so far, deserves to be a standard procedure for similarity testing’’. There have been a number of times in our field when techniques were introduced and adopted without an adequate theoretical discussion of their properties. Good examples of these are the triangular method and the proportion of detectors or discriminators, neither of which were developed with an appreciation for the psychological aspects of decision making at the time that they were conceived. Forty years after the triangular method was introduced, models for the method that incorporated a decision rule emerged and had an important impact on how we relate methods to each other and determine relative power. The TOST is likely to follow other techniques that sounded like a good idea at the time, if we do not discuss our alternatives thoroughly and know what we are dealing with.
This article appears as:
Ennis, D. (2008). Rejoinder to Bi and Meyners. Food Quality and Preference, 19, 347-348.
Colleagues can request this journal article here:
Rejoinder to Bi and Meyners
Not a Colleague? Click here to join for free!