Recovering stimuli from memory: a statistical method for linking discrimination and reproduction responses.

Investigations of sensory memory have often relied on discrimination processes: participants judge whether a reference stimulus is identical to a remembered stimulus. Recently, investigators have used response modes in which participants directly reproduce stimuli. Findings using reproduction tasks have suggested new ways of thinking about biases that pervade psychophysics. To allow an assessment of whether results obtained using reproduction responses persevere for discrimination tasks, this paper presents a statistical model that links the two approaches. The model uses the pattern of discrimination judgements to estimate parameters that represent the means and standard deviations that would have occurred under a reproduction task. An analysis of stimuli that have been previously studied using reproduction responses illustrates the method. Simulations investigate bias and efficiency, large-sample confidence intervals and the performance of likelihood-ratio tests of the adequacy of special submodels.