Earlier work on expert problem solving has usually been disciplinary-specific and only carried out with a small number of informants. In this project, however, 52 experts across 10 separate fields were involved. The Experts were volunteers who were recruited through the research team's professional contacts. This means that the sample was of not random but biased towards people who had experienced relatively similar training at U.S. universities within the last 15–50 years.
That the project resulted in a set of only 29 separate decisions, suggests that there is indeed a generic decision making aspect to solving real-world problems. Note that the authors focused only on what decisions needed to be made, not how they were made. This is because how the experts made decisions was very different from discipline to discipline.
Below are the 29 decisions divided into six groups with their relative frequency given in percent.
A) Selection and goals (100%)
1. What is important in the field? (61%)
2. Opportunity fits solver’s expertise (77%)
3. Goals, criteria, constraints? (100%)
B) Frame problem (100%)
4. Important features and info? (100%)
5. What predictive framework? (100%)
6. How to narrow down problem? (97%)
7. Related problems? (97%)
8. Potential solutions? (100%)
9. Is problem solvable? (74%)
C) Plan process for solving (100%)
10. Approximations and simplifications? (100%)
11. How to decompose into sub-problems? (68%)
12. Most difficult or uncertain areas? (90%)
13. What info needed? (100%)
14. Priorities? (87%)
15. Plan for getting information (100%)
D) Interpret info and choose solutions (100%)
16. Which calculations and data analysis? (81%)
17. How to represent and organize information?(68%)
18. How believable is information? (77%)
19. How does info compare to predictions? (100%)
20. Any significant anomalies? (71%)
21. Appropriate conclusions? (97%)
22. What is best solution? (97%)
E) Reflect (100%)
23. Assumptions and simplifications appropriate? (77%)
24. Additional knowledge needed? (84%)
25. How well is solving approach working? (94%)
26. How good is solution? (100%)
F) Implications and communicate results (84%)
27. Broader implications? (65%)
28. Audience for communication? (55%)
29. Best way to present work? (68%)
A common reaction to the list is that “There is nothing new here” or “We already do all this”. What is new is that we did not know before that there is a relatively small set of decisions that are made by a wide range of disciplinary experts. This information should logically have consequences for teaching and learning. The authors suggest that in any given course only a smaller number of the 29 decisions would be practiced, but that students need to practice all 29 decision types during their education.
One further issue that the authors note is that typical textbook problem sets are usually well defined and have one single answer. Real life problems are messier and there are other decisions that students will have to make. These are reflected in the list of 29 decisions.
Comment: The area of expert/novice problem solving has been the focus of research for many years. One of the early proponents of teaching students expert-like problem solving techniques was Alan van Heuvelen. Having noticed that students of Newtonian mechanics often jumped straight to putting values into an equation, van Heuvelen (1991) suggested they should be required to first draw a sketch of the problem, to which they then added the forces involved before transferring these values to an equation for quantitative calculation. He called his paper “Learning to think like a physicist” although it soon transpired that expert physicists actually didn’t work in the way he had suggested! Rather it was found that physicists often “jump in in the middle” of a problem and then move backwards and forwards between different representations as they discover what is needed. Beginning with ideas like Playing physics jeopardy (1999)—where students were given the answer and had to create the physics problem—van Heuvelen’s work began to emphasise the movement between different representations. This culminated in the popular text book The Active Learning Guide for College Physics where students are encouraged to move backwards and forwards between different representations throughout the book.
The study described in this paper adds new information, suggesting that learning to make the 29 decisions should also be in focus. Going forward, this may well become a central theme, but for now it is up to the individual lecturer to make decisions about how this generic decision-making might be practiced in the discipline-specific course they are teaching.
Note:
Nobel Laureate Carl Wieman is the leader of the research group that produced this study. Watch his presentation at the NU2022 Conference at Stockholm University here>>
References
van Heuvelen, A. (1991). Learning to think like a physicist: A review of research-based instructional strategies. American Journal of Physics, 59(10), 891-897.
van Heuvelen, A., & Maloney, D. (1999). Playing physics jeopardy. American Journal of Physics, 67, 252-256.
Text: John Airey, Department of Teaching and Learning
Keywords: real world problem solving, expert/novice comparison, undergraduate science, generic decisions
