Next: About this document Up: Deming and Variability in Manufactured Goods Previous: Background information

Suggestions for teaching

In my first Chance course in the fall of 1991 Walton's book was the main text. We basically discussed the major ideas of that book in a group discussion, taking several chapters at a time. The class contained 12 students. We had begun the course with a unit on probability and coincidences. The students had some familiarity with sequences of chance events.

The core of the book is the first 20 chapters, roughly half the book. In fact, the first 3 chapters outline Deming's philosophy (the 14 points and the 7 deadly diseases), with the remainder of these 20 explicating it. I preceded our discussions by showing the ``Against All Odds'' video mentioned above. An easily digested introduction like this seemed to be a good idea and a short video or a short, non-technical article (like those mentioned in section 1 above) plays this role.

Deming's philosophy is radical, but the students tend to have spotty experience with the world of business and consequently may not find Deming easy to discuss. There are probably various ways to overcome this problem, but two come to mind. First you could have the students apply Deming's philosophy to a context they have all experienced, such as education. They have al l gone through 12 or more years of education and have pretty strong opinions about it. Ask them to break up in to smaller groups (say 4 or 5) and discuss a series of questions about education such as (i) who is the customer/producer in the education setting?, (ii) what are the (two or three) most important of Deming's points in the education setting and what has been your experience vis a vis them, and (iii) how could education improve its performance vis a vis these points?

Another way to draw on common experience is to have students visit some local industry. We did it as a class in the fall of 1991, visiting a local, successful manufacturer of sportswear. Having do ne this tour together enhanced our ability to discuss the book. A variation on this idea would be to have different students (working in small teams) visit different local industries.

We spent 6 2-hour periods on this unit. The first period was a general discussion of the first part of Walton (after viewing the video). The next period we did the red bead experiment and discussed it. (We even plotted the data, although the technical details of control charts were not the focus.) This was very enlightening. Even though they'd read about this experiment in Walton, actually doing it themselves was very revealing. The teacher got to ham it up, intoning in the manner of Dr. Deming himself. The third period was the plant tour, followed by the fourth period where we discussed the visit. (This discussion precluded our discussing the latter part of Walton, which is a series of case studies. We read them, but having a live example was much better.)

The fifth period we performed the funnel experiment (details below). This was less than successful. They saw through it; too many of them realized intuitively that the smallest variability would occur from the no-adjustment strategy. Our sixth and final 2-hour class period was a wrap-up discussion. I am sure one could make a successful unit using far less time than this. A discussion on the first few chapters of Walton as applied to education, perhaps augmented with the red bead experiment or a video might constitute a nice 2 to 3 hour introduction to Quality. Even though Deming is the focus of this unit it is probably a good idea to alert the students to the notion that total quality management (TQM) goes beyond Deming. The paper by Snee suggests this to some extent.

I gave one writing assignment in this unit. I asked the students to write a one page essay on how well the company we visited adhered to Deming's principles, realizing that these would necessarily be based upon first impressions, and that it would be impossible to speak to all points. Most students did a good job with this assignment and found it a useful exercise to tie the tour into the reading.

On the next interation of Chance, I would retain the red-bead experiment. I would focus the initial discussion by using small groups and trying out the discussion of education. I would arrange another plant tour (maybe more than one, or maybe have them do different ones in teams). I would forgo the funnel experiment until I come up with a better understanding of how to do it. I would also consider having the students begin the unit by actually going out and collecting some data to discover that variability affects quality. For example, something as simple as comparing the percentage of in-tact cookies in competing brands of animal crackers would be a good project (albeit somewhat expensive).

Red Bead Experiment

The red bead experiment (or parable of the red beads) is explained adequately in both Walton's and Deming's books. One can buy fancy kits to perform this experiment and one could spend some time at the workbench building a paddle to lift beads from a box 50 at a time, but I used lack of money, time and woodworking skill as motivation for the following quick and dirty solution. It worked fine.

I bought at the local discount store about $15worth of marbles that are used to line the bottoms of clay flour pots. These are approximately half an inch in diameter and conveniently (at my discount store, at least) came in two colors: black and clear. I bought 12 bags of black beads and 3 of clear. There are roughly 100 marbles per bag so my box contained 1200 black and 300 clear. (These numbers are not precise and I didn't bother to count out the beads exactly as this is not necessary for the experiment.) My experiment becames then the ``clear bead experiment''. Recall that for Deming the white beads are what we wish to produce, the red beads are defects, and there are roughly 20 per cent defects in the process.

The paddle conveniently was found on the games shelf of the Moore family home. Boggle is a commercially available word game in which 16 lettered dice are shaken into a 16 cell (4 by 4) paddle that fits perfectly my marbles. If one dips the Boggle paddle into the box of marbles and lets the excess marbles fall off one obtains a neat sample of 16 marbles. Sixteen is smaller than 50 and, yes, the normal approximation to the binomial may be inadequate at this level (for the control chart calculations given in Walton), but that is clearly not the point of this exercise. This set-up worked beautifully and provided us with a very lively class.

The Funnel Experiment

I include this description here because I adequately solved the losistics of how to set it up (maybe you can find improvements in this and I'd love to hear them) but was disappointed in how it went with the students and invite you to make it work and tell me about it. I think that my students in this particular class were a bit too sharp, or perhaps I should have performed the experiment before they knew as much as they did.

I enclose the relevant pages of Out of the Crisis to describe the goal and basic set-up. My main contribution is a few details on the set-up. I bought simple kitchen funnels at the local hardware store. I borrowed the holders from the physics department; they were clamp stands they use in a number of demonstrations. When I secured the funnels in the stands they still wiggled so I secured these with rubber bands. (The funnels were a bit cocked, which didn't matter. They were stable.)

The students dropped ball bearings (that barely fit through the funnel holes but didn't bind) through the funnel and onto a large piece of graph paper. (I photo-enlarged regular graph paper.) By putting a piece of carbon paper (yes, such stuff still exists) over the graph paper in the approximate area the bearing will land the dropped bearing makes a nice dark mark on the graph paper. Of course, you want to have someone snatch the bearing out of the air after just that first bounce so you don't get extraneous marks, but my students were talented enough to handle this and yours will be too. After each drop, write the number of the drop beside the last mark to keep track of the sequence of drops.

Next: About this document Up: Deming and Variability in Manufactured Goods Previous: Background information