Finally! Media Computation book series done!

I have sitting next to me Problem Solving with Data Structures Using Java: A Multimedia Approach by Mark Guzdial and Barbara Ericson, the fourth and final Media Computation book! This joins the Python CS1 book, the Java CS1 book, and Barb’s (with Steve Cooper and Wanda Dann) Alice and Media Computation book.  I started the first Python book in 2002, so it’s been eight years to finally finish the series.  Woo hoo!  I look forward to not writing more textbooks in the forseeable future.

(BTW, yes, it’s a shockingly high price.  It was a surprise to us, too. We hope that market forces will cause that to drop.)

Boredom vs. Failure Part 2: The New Demographic

On the way to Rochester on Monday (across two planes, 3 hour delay, and 2 hour flight), I read Richard E. Mayer’s Multimedia Learning, 2nd Edition which I highly recommend.  It’s a synopsis of literally dozens of studies, categorized as supporting (and sometimes disputing) 12 principles of how to design multimedia for effective and transferable learning.  I saw a lot of great ideas for how to improve computing education. That observation also has a downside.  Some of Mayer’s well-supported principles aren’t appearing anywhere in computing education that I can see.

One figure really struck me, and I’m including it here (hoping that a single figure counts as ‘fair use’):

This figure summarizes four studies where student data were split based on the amount of prior knowledge that students had about the field of study.  In one experiment, the experimental treatments were integrating text and pictures (text next to pictures) and the other where they were separated.  In the other three experiments, students were shown text with supporting illustrations vs. only text.

What’s striking about these four results is the huge difference for students with low knowledge.  Doing it right matters a lot for these students.  What’s also striking is how it doesn’t make much difference for the high knowledge students.  In fact, in the first experiment, the low-knowledge students even did better than the high knowledge students when given integrated text plus illustrations.

Herein lies the challenge: High knowledge students, being put in a situation which isn’t much better than the alternate treatment, where they’re not being challenged, can be bored.  This figure highlights the trade-off that I mentioned in a previous post.  Do we risk failing the low-knowledge students by catering to the high-knowledge students, or do let the high-knowledge students be bored but help the low-knowledge students succeed?

Can we get success and challenge for all students?  That’s a great research goal!  However, we don’t know how right now, at least, that I can find in the research literature.  And since we’re not currently practicing the learning design principles that support low-knowledge students’ success, I suspect that we’re a long way from reaching that perfect path.

All of this is particularly salient for me this morning. The Atlanta Journal Constitution reported this morning that Georgia is now one of the first four states in the US to have majority low-income, underrepresented-minority students.  More than half of all students in Georgia public schools qualify for free lunches, and are Black, Latino, or Asian-Pacific Islanders.  These students tend to be lower-knowledge than the majority, higher SES students.  If we want to succeed at educating these students, we have to figure out how to use the principles that work best with them.  That may mean boring some middle and upper class White kids.  I think that’s better than failing out more than half the students, but that is a hard decision to make.