Information-processing architectures in multidimensional classification: A validation test of the systems factorial technology.

Abstract

A growing methodology, known as the systems factorial technology (SFT), is being developed to diagnose the types of information-processing architectures (serial, parallel, or coactive) and stopping rules (exhaustive or self-terminating) that operate in tasks of multidimensional perception. Whereas most previous applications of SFT have been in domains of simple detection and visual-memory search, this research extends the applications to foundational issues in multidimensional classification. Experiments are conducted in which subjects are required to classify objects into a conjunctive-rule category structure. In one case the stimuli vary along highly separable dimensions, whereas in another case they vary along integral dimensions. For the separable-dimension stimuli, the SFT methodology revealed a serial or parallel architecture with an exhaustive stopping rule. By contrast, for the integral-dimension stimuli, the SFT methodology provided clear evidence of coactivation. The research provides a validation of the SFT in the domain of classification and adds to the list of converging operations for distinguishing between separable-dimension and integraldimension interactions.Keywords information processing; response times; classification; mental architecture; integral and separable dimensions A fundamental issue in the psychology of perception concerns how information from multiple dimensions is processed in tasks such as detection, recognition, and classification (e.g., Ashby, 1992;Garner, 1974; Kantowitz, 1974;Lockhead, 1972;Schweickert, 1992;Sternberg, 1969;Townsend, 1984). Consider, for example, a simple detection paradigm in which there is a potential target in the left visual field and one in the right visual field. On each trial, the observer's task is to simply detect whether one of the targets is present. One basic question is whether the processing of the information operates in serial fashion or in parallel fashion. In serial processing, information from each visual field is gathered sequentially, one field at a time. By contrast, in parallel processing, information from both visual fields is gathered simultaneously. A second question is whether the processing is exhaustive or self-terminating. In the self-terminating case, processing would stop as soon as a single target is detected. By contrast, in the exhaustive case, processing would continue until the information has been gathered from both visual fields, regardless of whether a target had already been detected in one of them. A third question of interest in the present research is whether the processing may be coactive. In the present example, the intuition is that information from the separate visual Correspondence concerning this article should be addressed to Mario Fific or Robert M. Nosofsky, Department of Psychological and Brain Sciences, Indiana University, Bloomington, IN 47405. E-mail: mfific@indiana.edu or nosofsky@indiana.edu. NIH Public Access NIH-PA Author ManuscriptNIH-PA Author Manuscript NIH-PA Author Manuscript fields ma...

Parallel interactive retrieval of item and associative information from event memory

“…In addition, error rates tend to be higher in our study than in those to which SFT is typically applied. We therefore obtained predictions for the SFT measures using a combination of simulations (Diederich & Busemeyer, 2003;Fific et al, 2008;Eidels et al, 2011) and analytic methods (Townsend & Thomas, 1994), as described in Appendix A. The resulting predictions allow for both a large proportion of errors (from near 0% to 96%) and violations of selective influence; while we lose the statistical power available with low error rates and selective influence, we can be confident that we have provided a fair qualitative characterization of each possible retrieval architecture.…”

“…Although these explorations were quite thorough, they cannot be as definitive as analytic proofs, which remain an active area of exploration. For example, while the analytic form of the SIC is known for some classes of coactive system (Townsend & Nozawa, 1995;Fific et al, 2008;, there as yet is no general proof that covers all coactive systems, let alone those with more complex interactions of the kind we considered. While the converging evidence from our various analyses and their relation to other results in the literature (as discussed both above and below) puts our conclusions regarding memory on firm footing, our efforts also contribute to the ongoing project to understand the qualitative signatures of different cognitive systems and to develop experimental means to detect them (Townsend & Wenger, 2004;Eidels et al, 2011;Little et al, 2011;Yang et al, 2014).…”

Numerical predictions for serial, parallel, and coactive logical rule-based models of categorization response time

“…Even research in visual search, where the ambiguity of the existing data has been enough to prompt the declaration of the serial-versus-parallel debate as a dead-end, has shown renewed interest in differentiating serial and parallel models by using better experimental techniques (Thornton & Gilden, 2007) and by focusing on full RT distributions (Donkin & Shiffrin, 2011;Wolfe, Palmer, & Horowitz, 2010). In categorization, questions about the serial versus parallel processing of stimuli comprising multiple dimensions have only recently begun to be asked (Bradmetz & Mathy, 2008;Fifić et al, 2010;Fifić et al, 2008;Lafond, Lacouture, & Mineau, 2007;Little et al, 2011). Consequently, having models that can produce predictions at the level of full RT distributions is central to uncovering the underlying architecture used to process multidimensional stimuli.…”

“…Theoretical and methodological advances have made it possible to design experiments that can differentiate serial and parallel models (i.e., the systems factorial technology; Fifić, Nosofsky, & Townsend, 2008;Townsend & Nozawa, 1995;Townsend & Fifić, 2004;Townsend & Wenger, 2004a). Even research in visual search, where the ambiguity of the existing data has been enough to prompt the declaration of the serial-versus-parallel debate as a dead-end, has shown renewed interest in differentiating serial and parallel models by using better experimental techniques (Thornton & Gilden, 2007) and by focusing on full RT distributions (Donkin & Shiffrin, 2011;Wolfe, Palmer, & Horowitz, 2010).…”

Double Jeopardy in Inferring Cognitive Processes

“…The proposed SFT approach was designed to explore conditions under which the fundamental properties of mental processes, such as the order of processing (serial, parallel, coactive), stopping rule (terminating, exhaustive), process independence and capacity, could be inferred from data (e.g., Townsend and Ashby, 1983;Schweickert, 1985;Egeth and Dagenbach, 1991;Townsend and Nozawa, 1995;Schweickert et al, 2000). The SFT has been used in the context of various cognitive tasks: For perceptual processes (e.g., Townsend and Nozawa, 1995;Eidels et al, 2008;Fific et al, 2008a;Johnson et al, 2010;Yang, 2011;Yang et al, 2013), for visual and memory search tasks (e.g., Egeth and Dagenbach, 1991;Townsend, 2001, 2006;Townsend and Fific, 2004;Fific et al, 2008b;Sung, 2008), for face perception tasks (Ingvalson and Wenger, 2005;Fific and Townsend, 2010), and for classification and categorization (e.g., Little et al, 2011Little et al, , 2013). 3 The current study doesn't evaluate model complexity as a quantitative criterion for model selection and falsification.…”