tag:blogger.com,1999:blog-3722233.post6514655477374298191..comments2024-03-28T18:17:00.135-05:00Comments on Computational Complexity: Non controversial thoughts on rankingsLance Fortnowhttp://www.blogger.com/profile/06752030912874378610noreply@blogger.comBlogger7125tag:blogger.com,1999:blog-3722233.post-3942804719708402682014-11-21T09:29:43.423-06:002014-11-21T09:29:43.423-06:00At the end of the day what most prospective grad s...<i>At the end of the day what most prospective grad students want is to continue doing research after they finish their PhD. at a good school.</i><br /><br />Is that so? Over the past few years several of the top graduates from the big four went voluntarily to industry. We have seen the same phenomenon at my institution (top 20), where some of our top graduates spurned academic offers to go to industry. These star candidates got industry offers in the range of $130-200K.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-3722233.post-50125422325327382132014-11-16T03:35:09.653-06:002014-11-16T03:35:09.653-06:00You are suggesting information that we should put ...You are suggesting information that we should put for each department without asking what a grad student cares or should care about. Isn't it better to do a survey of grad students and recent graduates and ask them what is important for them? What would they like to know before choosing a grad school?<br /><br />Here are some that comes to my mind: the fields each department is active in, the star faculty (people like Valiant, Blum, ...), some measure of research activity (like number of papers in top conferences), the prospect of finding a tenure track academic position in one of top schools (with break down), etc.<br /><br />I personally like to know the likelihood of a graduate finding a good tenure-track academic position a few years after graduation for each school. This would penalize departments who take large number of students who do not find good tenure-track academic positions. It will also force departments to do more to help students find positions. At the end of the day what most prospective grad students want is to continue doing research after they finish their PhD. at a good school. I believe this is as important as the quality of research and the list of research areas of schools. An it is a quite objective and meaningful measure and is not easy to manipulate.<br /><br />There are other factors that one should take into account when selecting a school for graduate studies and an adviser many of which are not academic issues. But the prospective of ending up in a good academic job plus the quality of research are the top academic measures. The second one includes the subjective evaluation of departments by experts in an objective way: how likely is your department to hire someone from each school? Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-3722233.post-1486443078539693792014-11-13T19:53:18.439-06:002014-11-13T19:53:18.439-06:00How do you compute this GPV? Is it just the weight...How do you compute this GPV? Is it just the weighted (biased? unbiased?) variance, or something else?Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-3722233.post-66862664429049427582014-11-13T07:06:51.639-06:002014-11-13T07:06:51.639-06:00I propose the following ranking (or rather "o...I propose the following ranking (or rather "ordering"); it's one that all theory people should be familiar with.<br /><br />Alphabetically.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-3722233.post-8613678127771697292014-11-12T10:47:00.397-06:002014-11-12T10:47:00.397-06:00"Will this rank as one of Bill's top post..."Will this rank as one of Bill's top posts?" Fortnow.<br /><br />Sorry to see that is not happening. High rank needs high controversy.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-3722233.post-79485680967530991292014-11-11T20:08:19.393-06:002014-11-11T20:08:19.393-06:00From your concern about GPA, maybe schools should ...From your concern about GPA, maybe schools should also report GPV -- Grade Point Variance.<br /><br />It would be a plus because (1) it conveys more information to colleges / grad schools about the applicant, and (2) some students will obsess over it, meaning they'll probably understand variance better.Andy Parrishhttps://www.blogger.com/profile/12252029594014518238noreply@blogger.comtag:blogger.com,1999:blog-3722233.post-37149804529676740682014-11-11T11:32:49.085-06:002014-11-11T11:32:49.085-06:00I think it's interesting to flip this around a...I think it's interesting to flip this around and view it as a mechanism design problem. Assuming the entities being ranked (e.g., departments) receive utility from being ranked highly, what measure(s) would you want to use to encourage broadly beneficial behavior? (E.g., in the case of departments, behavior beneficial to the advancement of their discipline, behavior beneficial to the students who go through them, behavior beneficial to humanity, etc.) One thing for sure is you probably would want a measure that includes many different positive things that people can do. Anonymousnoreply@blogger.com