ICD stuff

Welcome to the ICD Stuff blog.

July 03, 2009

What do ICDers do when they grow up?

Here is a nice article in Wired about Googlenomics, featuring my co-author and friend Hal Varian. This describes a number of ways that Google has combined vast data mining resources with economics to do some incentive-centered design.

Hal is a micreconomist par excellence, who has made important contributions to both theory and empirical work. He was one of the first economists who took the study of the Internet and related phenomena seriously.

He was my colleague at Michigan, and involved in some of the early meetings in which a group of us developed the plan to create the School of Information. The year we launched, however, he departed for Berkeley, where a year later he was dean of their new School of Information (called SIMS at the time, but since renamed). For the past few years he has been on leave from Berkeley to be the Chief Economist at Google.

Posted by jmm at 03:06 PM | Comments (0) | TrackBack (0) | Permalink »

<< Another take on manipulating Wikipedia | Main

May 08, 2009

Another take on manipulating Wikipedia

(Scott Adams, http://dilbert.com/strips/, 8 May 2009)

Posted by jmm at 10:54 AM | Comments (0) | TrackBack (0) | Permalink »

<< Incentives for filling out teaching evaluations? | Main | What do ICDers do when they grow up? >>

April 27, 2009

Incentives for filling out teaching evaluations?

The University of Michigan switched to an online course evaluation system this year (Fall 2008 was the first full run).

One of the primary concerns during the study that went into the decision, and during the implementation, is what would happen to response rates. Selection bias for respondents is a concern for any survey, and there are obvious (and documented) reasons to expect that response is correlated with student satisfaction with a course and a teacher, which definitely would create a bias, the more so the lower is the response rate. With the traditional fill-in-a-form system, if administered during a class late in the semester when attendance is high (perhaps because students are getting concerned about what will be on the exam!), through convenience and peer pressure the response rate can be fairly high (as I recall, the norm at UM was above 70%). With an online system, evals are filled out at the student's convenience. This might catch students who would have missed the evaluation class day, but might lose many others.

The evidence UM collected from other campuses (and from two schools at UM that had already implemented their own online system), response rates do tend to be a bit lower, though not much.

My ICD colleague, Yan Chen, and I have chatted a bit about providing incentives to students to complete the online evaluations. One idea that she'd heard seemed effective in another department is to award a nominal amount of "extra credit" to students who submit an evaluation (in some systems, even though the evaluation content is anonymized, it is possible to track whether a student submitted something).

Here is a blog entry on the use of course evaluation incentives from the Center for Teaching, Learning and Technology at Washington State University. Their impression is that the incentives (including extra credit and randomized cash awards) have had little impact on participation (at least at the level at which the incentives were offered). Their conclusions are mostly based on anecdotes, but they provide a link to a discussion of a comparative study that was done in their College of Engineering which, it appears, was being submitted for publication.

Posted by jmm at 01:36 PM | Comments (0) | TrackBack (0) | Permalink »

<< Wits & Wagers | Main | Another take on manipulating Wikipedia >>

April 25, 2009

Wits & Wagers

The other day I created a prediction market for School of Information students, faculty and staff to see what kind of forecast we could generate for the coming year's enrollment.

One of our staff told me that she had played a board game that was quite similar, and indeed it is: Wits & Wagers is more or less a prediction market board game. It doesn't aggregate the side bets into a market prediction, but it does allow players to take a position in various claims (actually, they are trivia questions, not predictions).

(Thanks to Kelly Kowatch.)

Posted by jmm at 02:39 PM | Comments (0) | TrackBack (0) | Permalink »

<< Beautiful, but Hollywoodized mind | Main | Incentives for filling out teaching evaluations? >>

April 24, 2009

Beautiful, but Hollywoodized mind

All economists have followed the story of John Nash as he recovered from 35+ years of paranoid schizophrenia in the mid 1990s, and was awarded the Nobel Prize for his Ph.D. dissertation in which he invented the core concepts of non-cooperative game theory. After all, every last one of us learned Nash equilibrium in our first year of graduate school (or earlier), and an awful lot of us use the concept daily in our research and teaching.

And so, like many others, I read Sylvia Nasar's biography A Beautiful Mind as soon as it came out. Then rushed to see the Russell Crowe movie by the same name.

Oh no! In the key scene of the movie (well, to us economists), when Nash excitedly explains his equilibrium concept to his buddies while sitting in a bar...the screenwriters got it wrong. The "Nash equilibrium" Crowe describes in fact is not an equilibrium. Not even close: every one of the three players in the game described has an incentive to deviate and go for the blonde. Grave disappointment. How could Hollywood spend many millions on a biography of a game theorist, and get his most famous, and let's face it, rather straightforward concept wrong?

Over the years, when I would teach introductory game theory to students, I usually mentioned the popular movie to motivate them a bit, and quickly sketched the highlights of Nash's strange life. Then I would say "but the Nash equilibrium in the bar scene is wrong" and let them figure it out as an exercise. (Easy in recent years, since they can go find the clip on YouTube.)

Of course, most anyone who knows a bit of game theory and saw the movie surely spotted the mistake: it's a bonehead error. But just today I discovered that his mistake has really made it in pop culture: it's the punch line of an xkcd comic (http://xkcd.com/182/)!