Measuring relative cue weighting : a reply to Morrison

Morrison (this issue) criticized the analytical and statistical methods that Escudero and Boersma (2004) used for assessing the participants' cue weightings in their listening experiments. He proposed that logistic regression constitutes a better method for measuring perceptual cue weighting than Escudero and Boersma's "edge difference ratio." The present paper starts by summarizing and illustrating Escudero and Boersma's experiment and analysis method and then addresses five of Morrison's objections"”namely the alleged ceiling effect, the alleged superiority of logistic regression, the problem of discarding data, the (dis)confirmation of two-category assimilation, and Escudero and Boersma's grouping of the data. We will argue that although logistic regression is a very good method for measuring cue weighting, there was nothing wrong with Escudero and Boersma's methodology in these five respects.