A RETEST OF THE RESPONSE‐BIAS EXPLANATION OF THE WORD‐FREQUENCY EFFECT
John R. Morton · British Journal of Mathematical and Statistical Psychology · 1968
This paper investigates whether the relation between the frequency of occurrence of a word and its sensory threshold can properly be ascribed to response bias. Seven models for the recognition process are tested against data of Brown & Rubenstein (1961). The models are of two kinds: probabilistic single‐threshold models and information‐processing models. It is concluded that the data can only be accounted for within the former class of models by assuming a differential effect of the stimulus as well as response bias. For the class of information‐processing models, the data require the assumptions that there is equal sensitivity for words of all frequencies and a lower criterion for more common words. The relation between the two acceptable models makes it apparent that the notions of ‘stimulus effect’ and ‘response effect’ are by no means as clear as has previously been thought.