“…The maximum likelihood estimators (p, 9 , O ) of (n, y, v ) satisfy In paired comparison experiments arising in the area of marketing research, order effect parameters are usually considered to be nuisance parameters which complicate the analysis, but which must be accounted for within the analysis or by randomization of the order of presentation in order to avoid biases in the estimates of the worth parameters. On the other hand, in paired comparison experiments arising in psychology, the 'order' of the items in a pair may represent a spatial or a temporal positional effect, and the order effect may be more important to the experimenter than the relative worth parameters themselves, since the stimuli are often controlled by the experimenter in these situations (see Harris, 1957). Within this second context, we examine the relative efficiency of two design configurations in testing for 'significant order effects and for differences among the item-worth parameters.…”