14 - Can Confusion-Data Inform SFT-Like Inference? A Comparison of SFT and Accuracy-Based Measures in Comparable Experiments

Abstract

Systems Factorial Technology (SFT; Townsend & Nozawa, 1995) is a powerful framework that has allowed researchers to investigate the basic properties of cognitive processing systems for the past 20 years. Based on comprehensive mathematical underpinnings, SFT has been used to successfully investigate processing in simple detection, memory search, Stroop tasks, word processing and other domains through unique Response Time (RT) based analyses. SFT is thus undeniably one of the most important contributions to the field of mathematical psychology. Although SFT is very powerful, it is, by definition, intrinsically linked to the research design that underpins its mathematics. Therefore, SFT cannot currently be applied in other frameworks, such as unique identification of signals from a range of response options. In this chapter we explore another powerful approach, General Recognition Theory (Ashby & Townsend, 1986), that can be applied in such tasks. Specifically, we investigate the utility of applying GRT to ask similar questions that researchers may pose under SFT. We present data from three new experiments, based on similar SFT experiments, but focusing on accuracy measures and GRT analyses. We also provide a comprehensive set of simulation results to investigate the potential link between basic cognitive processing properties in SFT and GRT. We conclude by highlighting the need for future research to further combine both accuracy and RT approaches to allow a more comprehensive account of cognitive processing.