tag:blogger.com,1999:blog-1331441403058020963.post6385647139264294827..comments2024-03-28T06:49:24.930-04:00Comments on International Political Economy at the University of North Carolina: On IR Professionalization, Grad School, and What I Think I've Been Doing for the Past Four YearsThomas Oatleyhttp://www.blogger.com/profile/14092437150746625670noreply@blogger.comBlogger9125tag:blogger.com,1999:blog-1331441403058020963.post-9148512188707771532016-09-24T02:10:21.545-04:002016-09-24T02:10:21.545-04:00Did you know that you can create short urls with L...Did you know that you can create short urls with <b><a href="http://shortener.syntaxlinks.com/r/LinkShrink" rel="nofollow">LinkShrink</a></b> and <b>earn cash for every click on</b> your shortened urls.Bloggerhttps://www.blogger.com/profile/07287821785570247118noreply@blogger.comtag:blogger.com,1999:blog-1331441403058020963.post-41911395599446527472012-04-15T14:36:33.327-04:002012-04-15T14:36:33.327-04:00Anon -
I took Dan's point to mean that profes...Anon -<br /><br />I took Dan's point to mean that professionalization was leading to a lack of good *theory*, not to poor execution of methods used to evaluate theory. What you're suggesting is something else, which I'd characterize as "We need *more* professionalization so that we stop seeing so many crap methods". I might agree with that but it's a different kind of argument.<br /><br />I do agree that the discipline incentivizes article count over quality. (The quality is *supposed* to be ensured by peer review.) It also incentivizes innovative new approaches and new ways of thinking, but there's a higher risk/reward there. I think Dan's career so far demonstrates that. If you can do it well then you'll do well in the discipline, as he has. But not everybody can do that sort of work well (particularly early in their careers), and for them it's less risky to tweak POLITY on MIDS.<br /><br />I think there's some value in the latter too. I'm very much okay with purely empirical studies of the sort that Voeten is describing in his Monkey Cage post.Kindred Winecoffhttps://www.blogger.com/profile/14330671232391851377noreply@blogger.comtag:blogger.com,1999:blog-1331441403058020963.post-81785872985667869592012-04-14T10:03:37.842-04:002012-04-14T10:03:37.842-04:00Kindred--
In my example, what was interesting was...Kindred--<br /><br />In my example, what was interesting was that this overwhelmingly non-quantitative audience focussed on the problem of the easy-to-quantify dependent variable my colleague was looking at being a bad proxy for the underlying concept of interest. I agree. But that hasn't stopped a number of published articles in JCR, ISQ and such from using the same (bad) measure, and indeed my colleague remains confident that reviewers for the journals he'd submit will understand that "this is what people use".<br /><br />This gets to Dan's larger point in that GOOD methods training would indeed fight against that. But I worry that the pressures of narrow professionalization lead to poor methods training and poor methods evaluation where we collectively suspend disbelief and applaud work that applies a new estimation tweak to the usual X from POLITY correlated with Y from MIDS, because that's the kind of work that lets you crank out enough articles/year. Studies that use "better" methods might yield better results from a scientific standpoint, but because those studies would be more difficult and especially, more time consuming (e.g., collecting original data), they produce a lower publication count and so that way is an inferior career strategy.AnonProfnoreply@blogger.comtag:blogger.com,1999:blog-1331441403058020963.post-44623028756351094832012-04-13T14:41:57.967-04:002012-04-13T14:41:57.967-04:00Dan -
The problem is that the incentive structure...Dan -<br /><br />The problem is that the incentive structure in the discipline does not allow both of "more methods training" and "less pressure to publish". If you learn methods that will help you to publish, using them to publish is a dominant strategy. Moreover, there's no good way to signal that the time you spent not publishing was spent actually thinking in a way that will lead to better future theorizing. <br /><br />Again, it's about signaling. And I don't think there's any way to change the structure of this game. <br /><br />AnonProf -<br /><br />Thanks for commenting. I'm not sure I understand what you're saying. For example: "the problems with inference in an article far more often lie in the basics of research design, defining and measuring variables, or whether a quantitative approach makes sense" to me sounds like a lack of good methods training. I'm not *just* referring to "the details of model estimation" when I refer to methods training, which is why I specifically mentioned things like assumptions regarding the data-generating process, knowledge about non-Gaussian distributions, taking potential bias seriously, etc. That's all part of methods, to me, as are the things you identified as being problematic.<br /><br />So your colleague that had "fundamental problems in whether his empirical data actually measured the concepts of interest" was making a <i>methodological</i> mistake, not a theoretical mistake. It sounds like the audience picked up on that criticized his methodological approach rather than his theoretical argument.<br /><br />See what I mean?Kindred Winecoffhttps://www.blogger.com/profile/14330671232391851377noreply@blogger.comtag:blogger.com,1999:blog-1331441403058020963.post-59689127435269722452012-04-13T09:48:44.317-04:002012-04-13T09:48:44.317-04:00I have to agree with Dan Nexon, that comment about...I have to agree with Dan Nexon, that comment about not being able to judge substantive work or even do things like identify the assumptions in that work without having had a lot of stats training seemed odd. I've had quite a bit of stats myself but the problems with inference in an article far more often lie in the basics of research design, defining and measuring variables, or whether a quantitative approach makes any sense than they do in the details of model estimation. <br /><br />Unless of course we all decide that we're more concerned with playing games with models and datasets than real-world inference. <br /><br />I'm reminded of a junior colleague's recent project, someone just out a top program with a heavy methods emphasis. He was doing some interesting, even innovative stats work, but with some really fundamental problems in whether his empirical data actually measured the concepts of interest. Sure enough when he presented it the audience nodded at the stats and then beat him bloody on conceptual mismatches and measurement problems in his variables. His reply was basically to shrug and say they was probably right, but nonetheless this dataset was "what everyone uses". And he's right on that, within the JCR/JPR/CMPS/etc community. What struck me as illustrating Nexon's point was a real sense from this new guy of, "Look, people, how will I get my manuscript submitted this semester if I can't just 'use what everyone uses'"?AnonProfnoreply@blogger.comtag:blogger.com,1999:blog-1331441403058020963.post-14960962178665368332012-04-12T17:20:42.324-04:002012-04-12T17:20:42.324-04:00Nice critique. Sorry I didn't see it sooner. I...Nice critique. Sorry I didn't see it sooner. I'd go further: we're in real danger of a mixed-methods equilibrium trap in which people combine statistical and narrative inference for no good reason other than getting a job.<br /><br />The difficulty that I think you are having is the same that Eirk had: understanding that my argument is about configurational effects. So there's no tension in my saying that there should be more methods training and saying that graduate students should be given more space to think and less pressure to produce publishable work. Right! Now! It is when you combine lots of emphasis on relatively narrow methods training, hubris about the nature of "social science," and pressure to professionalize as quickly as possible that theorization suffers. <br /><br />I'm troubled by your claim that you had no ability to evaluate substantive courses before you had lots of "methods training." Plenty of work stipulates the terms by which its authors intend it to be judged. Not all work can be best judged by the kind of training you discuss. Indeed, if you think it can, that's probably indicative of what I'm concerned about.<br /><br />And yes, get as much training as possible in grad school. It gets much harder to find time later in your career.Dan Nexonhttps://www.blogger.com/profile/14802717151098392988noreply@blogger.comtag:blogger.com,1999:blog-1331441403058020963.post-77643741771445249242012-04-12T17:19:37.840-04:002012-04-12T17:19:37.840-04:00This comment has been removed by the author.Dan Nexonhttps://www.blogger.com/profile/14802717151098392988noreply@blogger.comtag:blogger.com,1999:blog-1331441403058020963.post-50051442012675150222012-04-03T01:03:08.897-04:002012-04-03T01:03:08.897-04:00I'm going into a phd programme in political sc...I'm going into a phd programme in political science (U of T) this upcoming fall, and I'm quite interested in this ongoing debate. I have almost no ability to use and abuse statistical methods, though I'm quite conversant in my philosophy of science. I've got a couple publications in decent but field-specific journals (terrorism-related) but these are all based off of fairly simple historiographical and comparative work.<br /><br />I worry that I'll end up some awful sort of theorist whose obtuse work bears no relevance to anyone other than those within my own echo-chamber.<br /><br />I will take methods courses like they are going out of style. Why is that even a figure of speech?Simon Frankel Pratthttp://saidsimon.wordpress.comnoreply@blogger.comtag:blogger.com,1999:blog-1331441403058020963.post-24764346702195599282012-03-28T23:19:38.890-04:002012-03-28T23:19:38.890-04:00Definitely agree about the philosophy of science p...Definitely agree about the philosophy of science part. Really liked the essay. I definitely (still?) didn't know what I was doing in my first year.Zachary Joneshttp://www.zmjones.comnoreply@blogger.com