In Parts 1-5, I discussed the Team-Based Learning methodology (TBL), course design for an undergraduate INF 202 Introduction to Data and Databases class, and drew parallels between TBL and roleplaying games (RPGs). Part 6 reviewed the first week of classes. Part 7 detailed the design steps for the second section of the course.
Where We Are Going
This entry continues to explore the second and first full-TBL section of the course: data and spreadsheets.
Summing up in the last entry, I noted that not only does an instructor need to determine the learning objectives of the section, to find materials that adequately lay the foundations for those objectives, and to create a 10-question, multiple-choice Readiness Assessment Test (RAT) reviewing those materials, he or she should vary the learning taxonomy approach of the questions. In time, I realized one other design consideration was also important.
The RAT is part of a larger Readiness Assessment Process (RAP), which includes taking the RAT both individually and as a group, a mini-lecture review by the instructor on the most troublesome questions, group formulation of appeals to questions (including, most importantly, the rational for arguing that a group’s favorite answer is not the same as the instructor’s), and an instructor acceptance or denial of those appeals (thus allowing two “correct” answers to receive grading credit). In “wiring” the RAT for the entire process, I initially concentrated on the group test. I focused on creating at least one “wrong” answer that was plausible enough to attract a thoughtful student’s attention, but had some incorrect aspect sufficient for me to discount it in favor of my chosen answer. The intent was to spark group discussion over the question—one or more students would pick the misleading answer and the ensuing group debate would elucidate the concepts inherent in the question. As you can imagine, creating answers that are “plausibly” correct and “subtlety” incorrect is no mean feat. Indeed, in hindsight, it’s a design objective begging for lousy implementation.
Worse, such a design goal is both dishonest and counterproductive. As an instructor with (at least theoretically) a much better grasp of the material, it’s unfair and mean-spirited to design toward “fooling” the students into supporting a “gotcha” answer. This is a learning environment, not a political arena. The last thing I want to do is make the students hesitant to take a stand for risk of embarrassment. Moreover, I am asking the students to parse through the subtleties of an answer that mixes correct and incorrect features regarding newly introduced material. At best, that risks confusion. At worst, it emphasizes the incorrect notions to the detriment of the real lesson.
Unfortunately, it took me half a semester (three tests) before I realized my design error. Three tests of frustrating design time and puzzled/tentative students. Ah well, better late than never.
My epiphany came when I allowed myself to create questions with more than one correct answer. By removing the incorrect aspect of the answer and instead using it to highlight a different feature of the material, I both eliminated the deceit and expanded the learning. The group discussion could then proceed with full force because both sides would be arguing a “winning” case. Solid rationales could be marshaled in support of both answers. When the “correct” answer was revealed in the group’s scratch-off answer key, one side could claim victory (at least for a short time). The appeals process became more robust as the “losing” side could present their argument. In fact, once it became common knowledge that several RAT questions might have more than one correct answer, appeals would be encouraged. During my in-class review of the appeals, I could emphasize the worthy concepts inherent in the “losing” side’s approach. I would then accept the appeal and give credit to the second answer, increasing the appellants’ (and others) grades. Best of all, it seemed like the positive feelings among those who triumphed on an appeal was greater than the righteousness among those who backed the keyed answer initially. All good.
Now, I don’t want to overplay this approach. At most, two or three questions per 10-question test should be multi-pronged. Still, it’s a neat technique.
Taking my lead from the section-as-adventure idea (see Part 5), I wanted to create some overarching thread and coherence in the spreadsheet coursework. I had decided early on that the variegated grading instruments (in both types and weights) would necessitate a spreadsheet for tracking. Although I had definite ideas about what I wanted to see in this spreadsheet, I had no delusions that I had a monopoly on knowledge (particularly in the teaching arena). Also, grades are perhaps the most important feature of the course to the students. Finally, working through the grading spreadsheet would reinforce just what the students would be graded on and by how much. All in all, using a grading spreadsheet to explicate the various features of a spreadsheet as a data instrument made a great deal of sense.
So the “grading spreadsheet” adventure would open with the “database/system” design process (in limited form—goal, objective, requirements), proceed to laying out the grading instruments that needed to be tracked, make some calls on presentation, and end with the final calculations and, if they were really on the ball, VLOOKUPs. I didn’t have a conception of how I was going to incorporate Pivot Tables, but I hoped to see what the class could come up with.
To hammer the lessons home and provide some practical experience, the final assignment would be a functional grading spreadsheet with dummy data to make sure the formulas and VLOOKUPs worked. The climax of the story would be better student understanding of both spreadsheet capabilities and the grading schema of the course. No heroic sword fights, dramatic villainous reveals, or surprise plot twists, but it was an adventure that a data nerd could appreciate. Good enough for me.
A Crucial Sideroad
Happy with the idea of a “spreadsheet” adventure, I ran into an immediate design problem. I didn’t want to lose sight of the “introduction to data” portion of the course. Understanding what database solution best fit a data problem required some conception of “data.”
One of my fellow INF 202 instructors pointed out a wonderful article by Gene Bellinger called Knowledge Management—Emerging Perspectives (2004). In it, Bellinger provides an insightful, working definition of data, information, knowledge, and wisdom. Among information scientists, this debate can get fairly abstract and convoluted. Bellinger, on the other hand, made it fairly clear and highly practical. He focused on why it matters what the definitions are and what the ultimate goal was—using data in context (information) to identify patterns (knowledge) that can guide future behavior (wisdom). I definitely wanted to include that reading in this early section, and to devote an activity class to working through the concepts.
Unfortunately, the Bellinger article did not fit all that smoothly into the spreadsheet “adventure”. A clean way to mix the two didn’t occur to me, so I decided to make it a separate topic. Logically, talking about data before reviewing the ins and outs of a data store made sense, bu
t I wanted to avoid philosophical exercises as a basis for the students’ first group efforts. I decided to tackle the spreadsheet creation problem first and save the “what is data” discussion for the last class of the section.
Plotting the overall approach of the spreadsheet section and setting up its many activities consumed a significant amount of time. To my frustration, I also realized it was also not my only concern. The third TBL section—covering system development lifecycle and ER diagrams—was at most two weeks away. Once the students were done studying for RAT 2, the best among them would be looking forward to the next RAT and readings it covered. That meant I needed to settle on the third section’s readings early on in the second section.
As I’ve mentioned, the readings set the stage for the entire section. Thus, I needed to spend some time on design work for the third section before I could definitely settle on the readings. Consumed by section two RAT and activity construction, I was wholly uncertain where the time for that section three planning work would occur. Ugh.
More to come,
M Alexander Jurkat