Exploring the Interaction of Implicit and Explicit Processes to Facilitate Individual Skill Learning

Facilitative interactions of model- and experiencebased processes: Implications for type and flexibility of representation

“…This is a particularly important oversight, since nearly every real-world task of any complexity (e.g., radiologist, fighter pilot, truck driver) relies upon both experience-and model-based knowledge for its performance. Our results join those of prior work (e.g., Sallas et al, in press;Stanley et al, 1989) in suggesting that the two types of processing can sometimes be fruitfully combined to facilitate learning (but not always; Domangue et al, 2004;Sun & Mathews, 2005).…”

“…As learning progresses, participants are able to rely on their memories both types of processes operate in nearly all tasks (e.g., Mathews, 1997;Reber, 1989;Seger, 1994;Willingham, Nissen, & Bullemer, 1989). Because most tasks involve both processes, one important but insufficiently studied question is how these processes interact (Mathews, 1997;Mathews et al, 1989;Sun & Mathews, 2005). The purpose of the following experiments is to explore this interaction and its implications, particularly for the type and flexibility of knowledge representations which result from these processes.…”

“…Thus, we argue that participants can learn to control a dynamic system using primarily experience-based processing. Most previous attempts to study the impact of model-based processing on learning this task have focused on the effect of instructing participants to try to discover the rules governing the system (e.g., Berry & Broadbent, 1988;Sun & Mathews, 2005), and the result has often been performance decrements rather than facilitation.…”

“…In considering the nature of the interaction of model-and experience-based processes, we first argue that-although both processes are typically involved in most tasks-it is possible to manipulate conditions in such a way as to emphasize one type over another (Sun & Mathews, 2005). In the following experiments, we examine one specific type of interaction-namely, the effect of providing partial or full model-based information before training on participants' ability to learn to control a dynamic system (Berry & Broadbent, 1984).…”

“…1 In this conception, experience-based knowledge is acquired relatively directly from the environment, and features of family resemblance are abstracted over multiple encounters with category members (e.g., Brooks, 1987;Estes, 1986;Hintzman, 1986;Medin & Schaffer, 1978). This type of processing is fairly error-tolerant (e.g., Sun & Mathews, 2005), and the resulting knowledge is often difficult to articulate (e.g., Lewicki, Czyzewska, & Hoffman, 1987;Reber, 1989). More broadly, important features of stimuli may be learned without the intention of doing so.…”