Educational Engineering

Engineering in other disciplines proceeds through some general phases. Each phase requires specific skills from the engineer and specific tools. The first phase is invention where something new is created and played with until its value is demonstrated or it is abandoned. The new invention is very unreliable and often will only work for the inventor, much like most educational innovations. The engineer must have the original idea and facility with the topic matter. In the next phase, general engineering, the engineer uses broad instruments to characterize his new invention and begin to optimize it. Perhaps the engineer looks at fuel efficiency and breakdown rate. The invention becomes better understood and more reliable. The general analysis will take the engineer only so far, eventually the engineer begins to notice that modifications that improve one system degrade the performance of another system. In the last phase, scientific engineering, the invention is completely modeled and understood. The engineer can then take the invention to the limits of its theoretical performance. Our test case for educational engineering, University Physics II, has finished the general engineering phase, and scientific characterization is underway.

To scientifically engineer a class, one must understand the operation of the class, and more importantly the information presented and the final state of student understanding in minute detail. The process modeling part of our research seeks to model the information content of a science class over a very finely characterized set of fundamental intellectual steps, called processes. The set of conceptual processes involved in introductory electricity and magnetism was developed by sampling traditional sources and used to develop an extensive conceptual inventory construction tool.

The information delivered by the course is only partially mastered by students. To completely characterize the course, this partial mastery must be measured. Detailed work on the uncertainty in student knowledge is underway.

The information presented in a class is only one facet of class operation. The actions students take in response to a class also effect learning. Jennifer McGee, now at Colorado, performed a detailed measurement of the effect of different student behavior on their test average and conceptual gain. Shawn Ballard looked in detail at the effect of student writing behavior on performance.

The processes used for class characterization were identified by sampling traditional textbooks. This was part of an ongoing research effort to understand the textbook as a part of a physics course.