Monday, 13 October 2008

The paper that caught my attention at QME Conference in New York

The annual QME conference is getting more popular. They even closed the registration this year. Despite the undercurrent of gloom in New York that everyone talked about at least briefly, the sessions were full of the curiosity and open debate that has come to mark this conference. The hosts, New York University Stern School of Business made it all go smooth under the leadership of Tulin Erdem and Russ Winer. I found all the sessions interesting but was struck most by one. This was on Saturday just before lunch: 11:30 - 12:30 pm Session 3 
A Simple Nonparametric Estimator for the Distribution of Random Coefficients in Discrete Choice Models 
Patrick Bajari (Minnesota), Jeremy T. Fox (Chicago), Kyoo il Kim (Minnesota), and Stephen Ryan (MIT) discussant: Andres Musalem (Duke).

This paper is all about empirical methodology, and in my research I don’t normally spend a lot of time on that kind of thing. But then, once in a way comes along an idea that makes us all sit up because it is so clever, and so obvious in hindsight, and in a word so beautiful. This paper has such an idea. I will let you find out for yourself what it is. I had perused the paper before I got to New York, and was curious to find out if my initial reaction to the paper was justified. The speaker, Jeremy Fox, soon confirmed my assessment with a presentation that was as informative and disarming as it was definitive. The participants’ comments, especially on the validity of the idea’s simplicity claim were it to be compared to Bayesian methods, and of the exact applications where its full force would be present were just as interesting. It occurs to me that this idea will soon find a place in the estimation toolbox of most marketing scholars. What may need to be sorted out is for which models it would be most beneficial to exploit the idea. I am not an expert in this area, but I will watch this with interest.

Every year I look forward to the SICS conference in Berkeley in the summer, and now after attending the last two years, I will add QME to my calendar. SICS with its emphasis on analytical modeling happens to be more central to my research but QME is also a well organized and stimulating conference and there is much to gain from its emphasis on empirical methodologies and to a lesser extent on substantive findings. Perhaps QME will return to the more freewheeling discussion of past meetings that is also an essential feature of SICS. And finally, I should put in a word for UTD-FORMS conference which will take place in February 2009, and that will be the third year in a row. With its format similar to SICS and QME, and emphasis on substantive issues in marketing this conference could well be seen as the third leg of a tripod bringing current research in marketing to everyone. Not coincidentally, I was a discussant in all three conferences in 2008. And I really enjoyed them all.


Ron said...

Jeremy nailed the presentation, despite the senseless interruptions from the moose in the front row ;)

Mosong said...

Yes, the idea is real simple and clever. Have people started to apply it to their work?