Assessing Interactions through Concurrent-Judgment Experiments

Abstract

Many important psychophysical questions concern the interaction or combination of different components of a stimulus. Classical psychophysical methods for assessing whether two stimulus aspects are coded independently (eg, masking and summation) provide limited information about the nature of whatever interactions are discovered. In both older work in detection and recent work in complex pattern discrimination, we have used a double-judgment paradigm in which the observer rates two aspects of a stimulus simultaneously. The paradigm provides a rich source of information about the codes underlying each psychophysical decision. It is unique in permitting us to investigate effects resulting from correlations in noise. We review the theoretical, technological, and methodological results that led us to develop this approach. Procedural antecedents lie in theories of dimensional interaction, in signal detection theory, and in information theory. Analytically, we draw on methods from several branches of statistics, including categorical data analysis and structural equation modeling. Also key to our work are advances in computational power: both our experimental procedures and our data analysis would have been difficult or impossible two decades ago.