I'm not sure whether to cheer or wince at the Gates-centered Next Generation Learning Challenge grant competition for various online learning tools (see the New York Times and Chronicle coverage). Wince potential: more millions being thrown in hopes that somehow, somewhere, "technology" will improve schooling. The good part: tools must be available publicly at the end of grants (at least from what I can tell). The very best part: at least from the initial coverage, it looks like this initiative focuses on community colleges, which serves the college students most in need of all sorts of assistance.
Let's just hope no one is funded to teach developmental math courses in Second Life.
No, that's not likely to happen, but of the four areas within the program structure (open courseware, blended learning, "deeper learning," and "learner analytics"), there are some where it's easier to make progress on than others. Open courseware for developmental classes and common gen-ed classes in community colleges offer the best promise, I think, because so many beginning courses have the same essential objectives, and if you can kill the $200 add-on fee of a commercial text, and make it more likely for someone to pass a class, you've suddenly saved several hundred dollars for a student, and that's scalable.
Blended course structures offer more flexibility than fully-scheduled classes, and among other things a community college with a substantial number of blended courses has less of a parking crunch. (That's a flip comment, but the space crunch is real for many places.) On the other hand, I am less sanguine that a giant effort here will be scalable. Blended courses have a great reputation, but the term is a catchall for a variety of structures, and I worry that the potential benefits of structural flexibility may be swamped by overenthusiasm for something we don't really understand well.
Unfortunately, the two jargon-monoxide categories are too likely to be wastes of money. In terms of "deeper learning" (which I read as "we'll throw money at people calling games classes"), I suspect small-scale, individual-faculty efforts will work better at investigating game-like course structures than big-ball-o'-wax efforts. NITLE members are more likely to crack this nut in an analytical sense than the likely applicants to Gates. To the Gates and Hewlett folks: pay Bryan Alexander's salaries, and maybe a few assistants, and let them go at it with no strings attached. Trust me, the ROI will be better than what you're likely to see from your plans.
Finally, to predictive systems to identify and intervene in the education of students at risk, aka "learner analytics" (which sounds like a Wall Street management consulting company title, to be honest: Lerner Analytics)… I wish this money had been spent in two different ways, not an attempt to scale up programs where the best research base the grant staff could identify apparently lies in nonrefereed sources, or where the research is clearly much more narrow in focus than the grant goals are. What two ways am I thinking of? Local support: a much smaller amount of money could be used for "quick and dirty" training and maybe even staff-salary support at large community colleges, to make sure that the largest CCs have enough people on hand who can slice and dice the messy data that institutions generally deal with (and do more than scramble to generate the whatever-someone-mandates reports that are required). State level: funding for national collection of multiple waves of longitudinal data on community-college students and the right outreach staff to make sure the data is used. Or help with cleaning the data for staff in states and other stakeholders with a need to look at data. SOMETHING other than blue-sky assumptions about institutional capacity.