xF: Methodologies of problem solving research
Sessions: QF (10:30 - 11:45) and RF (2 - 3:15)
Organizers: Thomas Foster, Southern Illinois University Edwardsville and Kathleen Harper, The Ohio State University
Presenters:
Charles Henderson, Western Michigan University and/or Vincent Kuo, University of Minnesota
Thomas Foster, Southern Illinois University - Edwardsville
Leonardo Hsu, University of Minnesota
Kathleen Harper, The Ohio State University
David Van Domelen, Kansas State University
David Maloney, Indiana-Purdue Fort Wayne
We propose a targeted poster session on methodologies of problem solving research. The focus will be on the different types of data one might collect and the subsequent analysis, and how these decisions relate to the specific area probed by the research. Each presenter will be asked to describe a useful technique, explain valid applications of that technique, and discuss how the generated data is analyzed.
The following topics will be presented:
"An Overview of Research on Problem Solving in Physics"
David P. Maloney, Indiana University- Purdue University, Fort Wayne
Research on how students' attempts to solve physics problems has been going on for over a quarter of a century and we have learned much from this research. However, there is still much that is unclear, and the work to date has been of very limited value in helping students learn to solve problems. Why is that? This poster will raise a number of questions, e.g., What is a problem? Can a task be defined as a problem independent of the solver attempting the task? What are problem solving skills?, that I would argue have not been adequately addressed in most problem solving research. Some suggestions for directions in problem solving research will be given.
“Using Semi-Structured Interviews and Multi-Layered Concept Maps to Generate an Initial Model of Physics Faculty Beliefs about Problem Solving”
Charles Henderson, Western Michigan University Vince H. Kuo, Ken Heller, Pat Heller, University of Minnesota Edit Yerushalmi, Wiezmann Institute for Science
Explanatory models are a crucial part of any scientific field. A useful model not only allows scientists to explain empirical observations, but can also lead to the creation of new lines of research. The goal of a generative study is to develop new models or new elements of existing models. This type of study is appropriate when working in an area where little prior research has been conducted. In the first phase of a three-phase research program, we used interviews with 6 research university instructors to generate an initial model of instructor thinking about teaching and learning. As part of this model, we identified 3 qualitatively different ways that these instructors think about the problem-solving process: a linear decision-making process, a process of exploration and trial and error, and a creative process that is different for each problem. This poster will describe how this initial model was generated and how it can be used.
“Using Grounded Model Construction and Explicit Analysis Methods to Converge towards an Explanatory Model of Physics Faculty Beliefs about Problem Solving”
Vince H. Kuo, Ken Heller, Pat Heller, University of Minnesota Charles Henderson, Western Michigan University Edit Yerushalmi, Wiezmann Institute for Science
The goal of a convergent study is to refine and expand a scientific model. In the first phase of a three-phase research program, we used interviews with 6 research university instructors to generate an initial model that contained 3 qualitatively different ways that instructors think about the problem-solving process in the context of introductory physics. This poster will describe the second, more convergent phase of this research program. During this phase, components of the initial explanatory model were tested and refined based on an ongoing, more targeted analysis of an expanded data set, consisting of interviews conducted with physics instructors from community colleges, state universities and private colleges. This convergent analysis resulted in verification of the existence of three conceptions of the problem-solving process, refinement of the defining features of each process, as well as an increase in the level of detail within each process.
“Using Computers to Teach and Evaluate Problem Solving Skills”
Leonardo Hsu, University of Minnesota
Computers have long been used to help students learn to solve physics problems. However, the ability of computers to help evaluate students' problem solving abilities or to identify their difficulties in solving problems has not been fully exploited. Computers can generate log files giving detailed information about a students' use of computer tutorials. In this poster, I will discuss methods for analyzing log files and some of the questions which such analyses can address.
“Pencil and Paper Pitfalls - An Example”
Dave Van Domelen, Kansas State University
The Problem Decomposition Diagnostic was developed in 1998-2000 as an attempt at an objective test of certain aspects of problem-solving skills. However, while reasonably good results were obtained, the project was abandoned, largely due to some of the inherent difficulties in the development of such an instrument. The PDD's development and the reasons for its abandonment will be presented.
“Expert-Novice Comparisons to Illuminate Differences in Perceptions of Problem Solutions”
Kathleen A. Harper, The Ohio State University
Comparing the performances of experts and novices on particular problem solving tasks has been a popular technique throughout the history of problem solving research. This research has uncovered differences in the ways the two groups categorize, approach, and solve problems. Recently, a similar technique has been applied asking these two groups questions about pre-written solutions to problems, revealing differences in what the two groups think problem solving is. These results and their implications for instruction will be discussed.
“Real-time Coding versus Careful Reflection: A Tale of Two Rubrics”
Thomas M. Foster, Southern Illinois University - Edwardsville
Traditional quantitative evaluation of curriculum reforms requires a quasi-experimental design and testable statistics. But is this paradigm tenable when the curricula objective is problem-solving? There are a number of issues to overcome including population variances, changing contexts, and skill selection. Rubrics can provide a way to measure problem solving skills, but they come with their own issues including effort, validity, reliability, and the balance of each.
| Michael C. Wittmann tel: 207 - 581 - 1237 |
Rachel E. Scherr tel: 301 - 405 - 6179 |