Sunday, April 15, 2012

Better Late Than Never

I watched Moneyball last night. I'm late to the party, I realize, but my interest in baseball is about at the same level as my interest in watching paint dry. Just ain't my thing. What I didn't realize, however, is the film really isn't a baseball movie: It's a data movie. What if you selected a team by the numbers---setting aside most everything you knew about players as people: work ethic, level of fitness, age, how weird their pitch looked, etc? According to the plot ("based on a true story..."), you can do pretty well. I'm oversimplifying things, of course, but I'm not interested in nitpicking the film. What I'm curious about is what this strategy would look like in an educational setting.

What are the metrics associated with teaching or being an administrator? The "Value Added Models" I've seen don't even begin to capture the complexity of classroom work. But suppose that you could collect data about being a teacher. What would you include? If baseball players have stats to describe the various aspects of their work...why don't teachers?

One thing about the way baseball was portrayed in Moneyball was the impersonal nature of it all. You could arrive at a stadium to play for one team to discover you'd been traded and would be playing for the opposing team that very night. A stronger player was picked up...so someone else had to be sent down. Baseball was portrayed a game of skills. Teaching is not a game and is built on relationships. What data could we collect to reflect this difference? I know the new teacher evaluation tools might be a start, but even then, sample size would be limited to a few times a year. A baseball season is over 150 games. A school year is (roughly) 180 days. I'd like to think that some sort of daily data collection would give you a better idea of what is happening in a classroom. What would you put on your scorecard?

There is, of course, the question of how the data would be used. And I would guess this is the number one reason why no one is asking what we could collect. Instead of the Dance of the Lemons at the end of each school year, what if school/grade/subject assignment was based on what the needs of a school were and how your strengths supported that? What if you could focus on what you do best instead of trying to do everything?

I'm still pondering my own answers to these questions. It's more thought experiment than realistic pursuit. In an age when more and more people collect personal data about themselves, I think it's only a matter of time before we start seeing teachers take control of their own data about their work.

Monday, April 2, 2012

How to Excel without Excel

I just submitted my first article for publication. It looks at how to use data visualization in the classroom as a form of feedback for teachers and students. I won't know for a few weeks whether or not it has been selected for publication, but even if I don't see ink on wood pulp, the research and writing process has been a wonderful one. And I am much indebted to those of you who provided content or copy edits.

Of course, I am not the first one to write about using data in the classroom. In fact, in an odd bit of serendipity today, I stumbled onto three jewels from the 1921 - 1922 issues of the American School Board Journal. All were written by E.L. Bowman, Director of Vocational Eduction [sic]; Erie, PA. The first one I'd like to share is actually the second in the series, Graphic Aids in School Administration: Watchtowers in Charts, published in February and March of 1922. (I pulled the document together from its various hiding places in an 800+ page anthology on Google Books. You're welcome.)

Lest the word "watchtower," give you visions of visitors knocking at your door, rest assured that the author uses the term as it relates to the forestry service.
The school executive should erect his watchtower. He should build for himself a structure which will keep him in touch with his situation, and enable him to control with precision the forces for education in his community. Unless the executive keeps always in touch with the situation, he will probably be finding another place in which to exercise his powers. Since the less effort he expends in gathering the data the more force he will have for other essential duties, he should plan and put into execution a routine for his subordinates that will reduce his task visualizing the situation to a glance at a chart or graph.

Bowman describes how he builds his graphs---even detailing some best practices and showcasing different visualizations.


He goes into great detail about how he constructed the pie chart shown below, but warns "The circle chart has many disadvantages. Although the quantities are really plotted as arcs of the circumference, the eye sees them as areas of segments subtended by these arcs. It is a well established fact that it is practically impossible to compare areas accurately, whereas it is entirely possible to compare parallel lineal measures easily and accurately. Hence the circle charts, though often used, are likely to be misleading." Atta boy.


In the second part of the article, Bowman describes uses of maps to plot data, giving tips on how to thrust your map tacks. He provides ideas for building a master schedule of classes, as well as a variety of ways to track progress.


And all of this with his handy ruler, protractor, and slide rule. Dude was definitely ahead of his time.

I don't know the story behind E.L. Bowman, but after finding today's treasures, I'm definitely a fan. Here's hoping I get a chance to launch a renewed in data visualization, his watchtower.