home - login - register

PERWiki

Parent Article


Referencing Articles (2)


Article Functions


Search the Wiki


More


Major research themes in PER

Traditional physics education research can be thought of as building on several major themes or topics, often represented by early researchers. Major early themes of PER were:

Operational Definitions and Tutorials

The University of Washington influence on PER centers around instruction using operational definitions (Physics by Inquiry) and identifying and addressing specific student difficulties (Tutorials in Introductory Physics). Both these efforts have strongly influenced the process of teacher education, a major element of their work.

The leading voices of this work were Arnold Arons and Lillian McDermott at the University of Washington Physics Education Group. McDermott's students have since greatly expanded this research program at other institutions. Her students and post docs are listed in the family tree. The most concise versions of McDermott's research program can be found in her award papers: the Millikan lecture of 1991 and the Oersted lecture of 2001. In these two papers, she describes the need for research as a guide to curriculum development and how such research can address the needs of teachers.

In the 1970s and 80s, research at UW consisted mostly of interviews (audiotaped but not videotaped). In early papers, Piagetian tasks are used and explicitly named as such. Research results were applied to the development of Physics by Inquiry, a curriculum for future teachers. In the late 80s and early 90s, these materials were adapted to help students in more traditional engineering classes. Peter Shaffer played a major role in this shift in the group. The Tutorials in Introductory Physics were first developed then. The research opportunities of working with larger courses and more students led to the development of new research tools, such as pretests (which preceded tutorial instruction). Fewer interviews were carried out, and surveys and examination questions became more important. With the arrival of Stamatis Vokos and Paula Heron, research topics expanded past the introductory sequence and into advanced topics such as thermodynamics and relativity. Most work since the early 90s has been on Tutorials, and less on Physics by Inquiry.

Media and Applied Topics

Where the UW PEG studied understanding of basic concepts in at times highly abstracted situations, researchers such as Bob Karplus, Bob Fuller, and Dean Zollman spent the 70s to the 90s investigating student reasoning about applied topics and the implementation of media in instruction.

Zollman and Fuller worked closely to bring media into teaching, first with films, later with videodiscs, and then videotape and digitized video. Because of the ability to show things not typically seen (or show in slow motion things not otherwise easily observed), one had the ability to investigate how students thought about the physics they were seeing. Fuller and Zollman built on this program throughout the 70s and 80s.

Knowledge-in-pieces: agents, primitives, and resources

The work on mental agents, begun at MIT by Marvin Minsky and Seymour Papert in the 60s and 70s, grew with Andy diSessa's work on phenomenological primitives (80s and 90s) and became a dominant theme in PER with Hammer's descriptions of conceptual and epistemological resources (90s and 2000s).

diSessa played a major role in the further development of a knowledge-in-pieces approach, carrying the topic from Minsky's early work to Hammer's later developments. diSessa's subject "J" is one of the best analyzed students in PER, and data from his interviews with J can be found in several major papers. Perhaps the longest paper in PER, "Toward an epistemology of physics," presents a detailed description of phenomenological primitives (p-prims), one kind of knowledge piece. In the paper, a strong argument is made against "misconceptions" research, and in favor of a manifold model of mind.

The "pieces" approach had an effect not just on modeling student reasoning but on the development of computer modeling tools. LOGO, developed in the 1970s, slowly developed into languages such as StarLogo and NetLogo, which are still actively used and developed. Boxer, another outgrowth of LOGO, is no longer in active development. All these systems allow the user to program using simple tools such as defining nearest-neighbor interactions and letting emergent properties develop. This method stands in contrast to "black box" uses of computers (see MBL tools).

Around 2000, the University of Maryland became a central point for research built on a manifold model of mind. Wittmann, in his 1998 PhD, had already applied diSessa's p-prims to student reasoning about wave physics. Hammer's arrival at Maryland in 1998 greatly extended this work as Hammer and co-workers greatly expanded on Minsky and diSessa's original models with the resources framework.

MBL Tools

With the growth of new technologies and opportunities for funding from FIPSE, new tools were developed that allowed students to gather data and interpret its conceptual meaning more easily. Because of the time period, the term "micro-computer based labs" was used (today's computers were called micro-computers in the days of mainframes). Priscilla Laws (Dickinson), David Sokoloff (Oregon), and Ron Thornton (Tufts) are the best known curriculum developers and researchers. Sokoloff and Thornton's work in developing surveys is described in the section on surveys. The curricula they developed are now collected in the Physics Suite, and include Workshop Physics, RealTime Physics, Tools for Scientific Thinking, and the Activity-Based Tutorials (the last is described in hybrid groups).

The development of early MBL tools rested on the historical coincidence of FIPSE funding in the early 80s and the creation of computer tools (motion sensors and force probes, in the beginning) which could be used with new graphical user interfaces. Early work was done on Atari computers, but the Macintosh was the major platform in the first decade of MBL tool use. Only after Windows 95 came out was a native Windows interface made for the software which analyzed the data coming from motion and force sensors, for example.

Later innovations of MBL tools came with the use of video to gather data about moving objects. VideoPoint, first released in the mid 90s, was never as popular as the motion sensors, though.

RealTime physics and the idea of technology-valuing instruction have begun to play an international role due to the outreach work of David Sokoloff. He won a Millikan Award in part for this work.

Survey Development

It is hard to overstate the publicity created when students did badly on conceptual surveys of physics. Perhaps the best known survey is the Force Concept Inventory from 1992 (Wells, Swackhammer, and Hestenes). Others include the Mechanics Baseline Test (again Hestenes et al.), the Force and Motion Conceptual Evaluation (Sokoloff and Thornton), the Conceptual Survey of Electricity and Magnetism (Maloney, van Heuvelen, and others), and many more. The major period of survey development was from the late 80s to the mid 90s. Major publicity came about in the 90s because of Eric Mazur's presentations of Harvard students' performance on the FCI.

Surveys brought together many existing research questions into a single questionnaire which was easily administered to a class both before and after instruction. Research from the 70s and 80s on student understanding of mechanics was collected in the FCI, for example. One problem, of course, was that some questions could be answered correctly for the wrong reasons, and a multiple choice survey had no space for the necessary free response writing.

By comparing results of matched students, one could see if the class had improved. Hake, in 1996, introduced the idea of normalized gain, in which results are scaled based on the possible gain (so, a 10% gain from 80% and a 40% gain from 20% both have a normalized gain of 0.5). Hake showed that students with certain kinds of instruction typically had normalized gains in specific ranges. The normalized gain is applied to results from many surveys, but some (such as the FMCE) are not always accurately described with the measure. More detailed analysis of surveys exists in the form of Bao's Model Analysis, for example.

The effect of Hake's study and the existence of the FCI were enormous. They brought publicity to the idea that PER existed and helped define the emphasis of PER on conceptual understanding in the larger community.