Paper Authors

Abstract

NOTE: The first page of text has been automatically extracted and included below in lieu of an abstract

Session 1330

Identifying the Components of Modeling Through Protocol Analysis

Paul S. Steif, Marina Pantazidou Carnegie Mellon University/National Technical University of Athens

Introduction

The art of engineering involves a variety of skills, and one of them is modeling. While the terms “model” and “modeling” are not simple to define, within engineering, Piel and Truxal1 offer a helpful account: “a model is the simplest possible system description that includes all important aspects.” One might add to this “at the appropriate level of detail and accuracy”, which helps to capture the significant amount of judgment involved in modeling.

While the importance of modeling process is obvious, and while on the surface it appears to be a mainstay of engineering education, we would argue that engineering instruction focuses much more heavily on model analysis, rather than on model formulation or development. Indeed, there would appear to be little more on model formulation than “watch me do it”; this suggests that educators operate on the hope that students will somehow draw together their various exposures to modeling experiences to become competent. Departing from this norm, we advocate explicit modeling instruction, based on these premises: while modeling is a complex mental task, it can be articulated, and that only with such articulation can we help students learn this skill. To this end, this paper describes an approach to uncovering the basic elements of modeling.

Within the educational literature there does not appear to be yet an articulation of the constituent components of the modeling process. Fortunately, though, the literature of cognition and instruction has moved from studies of well-defined subjects, such as physics, to open-ended tasks, such as design. Goel and Pirolli2 studied the structure of design by constructing design problems from several fields. They recorded interviews with designers engaged in these problems and asked them to “think aloud” as they proceeded. The transcribed interviews (“protocols”) were divided into statements, which were labeled (“coded”) according to a protocol-coding scheme similar to existing schemes of previous design studies. From this analysis, they identified twelve invariants, or, twelve components, of the design task.

With such an approach as guidance, we also sought to determine whether similar task decomposition is meaningful from the viewpoints of cognitive theory and instruction. Lovett and Greenhouse3 applied cognitive theories to instruction in statistics and performed a task decomposition. They supported their approach by appealing to Anderson and Lebiere’s4 theory of cognition, according to which declarative knowledge (“know that”) can be broken down in chunks, upon which procedural knowledge (“know how”) operates via production rules. As support for the instructional benefits of a task decomposition, they cited the work by Catrambone5, who demonstrated that both labeling and visually isolating subtasks in examples improved the performance of students in solving novel problems. In earlier work, Lovett6 had categorized approaches to task decomposition into a matrix of theoretical versus

Proceedings of the 2004 American Society for Engineering Education Annual Conference & Exposition Copyright 2004, American Society for Engineering Education

• Citation

• Format
  • APA
  • APA - LaTeX bibitem
  • MLA
  • MLA - LaTeX bibitem
  • Bibtex
  • EndNote - RIS

Pantazidou, M., & Steif, P. (2004, June), Identifying The Components Of Modeling Through Protocol Analysis Paper presented at 2004 Annual Conference, Salt Lake City, Utah. 10.18260/1-2--13847