OR/MS Today - June 2003 Issues In Education Teach the Process By Francis W. Wolek I'm a real nut about apprenticeship! I always believed there had to be lessons in an ages-old approach to education. My studies showed that some techniques, such as demonstrating tools, are already part of modern education. My quest to determine if we were missing something encouraged a brief apprenticeship as a blacksmith. That experience really clicked with my research on apprenticeship: the pedagogical heart of apprenticeship learning is repetition of the process that an especially talented worker uses for work. The blacksmith taught us by demonstrating a process and then asking us to practice it over and over. One example is "Squaring the Round" or converting a two-foot round rod into a two-foot, four-inch square one. My message here is that OR/MS educators should: a) design well-structured work processes; b) demonstrate their effectiveness; c) design progressively challenging exercises for student practice; d) tutor students on early exercises when needed; and e) promote productive reflection about each student's successes and failures using the process. (See my 1999 article for theoretical underpinnings and development.) A Sample Process My colleagues and I teach the following process for forecasting: 1) plot the data; 2) visually identify possible discontinuities, trends and seasonality; 3) verify the business forces creating each such pattern; 4) normalize the data by removing discontinuities; 5) use exponential smoothing to project data that contains only randomness; 6) use seasonally adjusted regression to project data with a business validated trend and/or seasonality; 7) determine if the forecast error (i.e., MAPE) is acceptable; and 8) interpret the forecast model for management. (See my 2001 article for further detail on discontinuities.) We start the students out by asking them to make a forecast using a data series from a company that sells holiday (Halloween, Easter, Valentines, etc.) supplies to retailers. The series has 19 points and contains an upward trend, seasonality and a downward discontinuity in one period. We canvass student solutions and logic before demonstrating our forecasting process. Part of that demonstration is a staged dialog with management about the business forces that created the discontinuity (a fire in the warehouse that damaged goods), the trend (increasing popularity of Halloween) and the seasonality (Halloween is the company's biggest holiday in terms of sales). The process ends with a managerial interpretation of the coefficients in the forecasting model (the positive slope indicates an annual growth of seven percent, Halloween sales are 14 percent above the growth trend, and average accuracy is two percent of actual sales). We explain the logic behind the forecasting process such as that plots provide more insight than tables, and that mathematical projection assumes that underlying business forces continue into the forecasted period. Students are then referred to a set of exercises to practice exactly the same process on real — but disguised — data. Graduate students are also asked to use the exact same process on a forecast for a company for which they or a friend work. Key Points in Apprenticeship I've used a process-based approach for more than five years in teaching operations management. Forecasting was a natural topic, but process teaching has also been successful for the analytical hierarchy process and statistical quality control. Some topics such as project management and operations strategy have proven only partially amenable, but I keep working on them. Along the way, my studies of apprenticeship and my own experience have taught some interesting lessons. Each step in a process should be concrete and yield useful results on its own. The process should be robust so that students gain insights even if they make an error or fail to complete the full process. Don't shortchange the exercises. A buddy developed a great process anchored on innovative software. Unfortunately, he claimed that he and his trainees lacked time for exercises. He never understood why trainees stopped using his process after initial trials. Insist on exact replication and do not allow experimentation. Honestly, I believe in the Socratic method, but students will not learn a process if they fiddle with it. The time for Socratic coaching is following reflection on the individual surprises that occur to each student. Be prepared for abandonment. A key aspect of skill development is abandoning previous ideas. For example, many of us were taught that we should use a black box approach in forecasting and avoid all explanation of underlying business forces. Develop your own talent at using a process. The credibility of your process and expertise hinges on your personal skill. Developing that skill requires extensive practice, self-testing and reflection. Do that and you will be rewarded not only with credible talent, but also with insights from your own reflection. Further Reading Wolek, Frank, 1999, "The Skill Building Processes of Apprenticeship," International Journal of Life-Long Education, Vol. 18, No. 5, pp. 395-406. Wolek, Frank, 2001, "Forecasting in a Dynamic Environment: How to Deal with Jumps and Spikes in Data," Journal of Business Forecasting, Vol. 20, No. 2, pp. 19-22. Wolek, Frank, 2003, "Forecasters as Business Strategists: Crossing the Threshold," Journal of Business Forecasting, (in press). Frank Wolek is on the faculty of the Department of Decision and Information Science at Villanova University. Table of Contents OR/MS Today Home Page OR/MS Today copyright © 2003 by the Institute for Operations Research and the Management Sciences. All rights reserved. Lionheart Publishing, Inc. 506 Roswell Rd., Suite 220, Marietta, GA 30060 USA Phone: 770-431-0867 | Fax: 770-432-6969 E-mail: lpi@lionhrtpub.com URL: http://www.lionhrtpub.com Web Site © Copyright 2003 by Lionheart Publishing, Inc. All rights reserved.