Life,Chaos, and Transdisciplinarity: A Personal Journey

This article uses an autobiography as an object of research, to both illustrate some principles of chaos theory in analytic practice, and give those ideas a personal and social context, thereby producing a unique but explanation-rich history of chaos theory and recent intellectual history of transdisciplinarity and social research in the West. The ideas from Chaos Theory it uses and illustrates include: three-body analysis (Poincaré); fractals (Mandelbrot); fuzzy logic (Zadeh); and the butterfly effect (Lorenz).     

Coherence and correspondence criteria for rationality: experts' estimation of risks of sexually transmitted infections

The aim of this study is to examine both coherence and correspondence criteria for rationality in experts' judgments of risk. We investigated biases in risk estimation for sexually transmitted infections (STIs) predicted by fuzzy‐trace theory, i.e., that specific errors would occur despite experts' knowledge of correct responses. One hundred twenty professionals with specific knowledge of STI risks in adolescents were administered a survey questionnaire to test predictions concerning: knowledge deficits (producing underestimation of risks); gist‐based representation of risk categories (producing overestimation of condom effectiveness); retrieval failure for risk knowledge (producing lower risk estimates); and processing interference in combining risk estimates (producing biases in post‐test diagnosis of infection). Retrieval was manipulated by asking estimation questions that “unpacked” the STI category into infection types or did not specify infection types. Other questions differentiated processing biases from knowledge deficits or retrieval failure by directly providing requisite knowledge. Experts' knowledge of STI transmission and infection risks was verified empirically. Nevertheless, under predictable conditions, they misestimated risk, overestimated the effectiveness of condoms, and also suffered from processing biases. When questions provided better retrieval supports (unpacked format), risk estimates improved. Biases were linked to gist representations, retrieval failures, and processing errors, as opposed to knowledge about STIs. Results support fuzzy‐trace theory's dual‐process assumptions that different types of errors are dissociated from one another, and separate failures of coherence and correspondence among the same sample of experts. Copyright © 2005 John Wiley & Sons, Ltd.     

Log Data Analysis with ANFIS: A Fuzzy Neural Network Approach

This study explores the use of the Adaptive Neuro-Fuzzy Inference System (ANFIS), a neuro-fuzzy approach, to analyze the log data of technology-based assessments to extract relevant features of student problem-solving processes, and develop and refine a set of fuzzy logic rules that could be used to interpret student performance. The log data that record student response processes while solving a science simulation task were analyzed with ANFIS. Results indicate the ANFIS analysis could generate and refine a set of fuzzy rules that shed lights on the process of how students solve the simulation task. We conclude the article by discussing the advantages of combining human judgments with the learning capacity of ANFIS for log data analysis and outlining the limitations of the current study and areas of future research.