… Because they are so much more Dudalicious. In honour of the David Gilmour (not the guitarist!) school of teaching, from now on I will only use statistical techniques designed by men. Sure, I could use Generalized Linear Latent and Mixed Models (GLLAMM), but just listen to the name of the damned thing. It’s like the Jane Austen of stats, and unsurprisingly it was developed by a woman (Sophia Rabe-Hasketh). Hardin and Hilbe just had a much more indefinably cool … manliness … about them, so I think for clustered binomial or count data I’ll just wing it with Generalized Estimating Equations. Luckily I don’t do much in the way of RCTs, because the classic text on experimental design by Cochran and Cox is half-authored by a woman – I can’t tell which bit she wrote so I’ll just have to dump the lot to be sure. This could be a bit tricky, because that stuff is pretty fundamental to how we think about efficiency in experimental design. No problem really, though, I’ll just make sure I apply for bigger grants and recruit more subjects. Typical of a woman to write a book about how to be thrifty with sample sizes really, isn’t it? Real men just recruit more subjects.
David Gilmour also doesn’t like Chinese authors, so if I’m going to follow his approach I’ll probably have to drop any adjustment for probability sampling, since a lot of the development work for those methods was conducted by Indians after independence. That shouldn’t be too bad because there are still some low-grade journals that let you publish without adjusting for your sampling process. Of course, to be sure I think I should develop a few stock phrases to deploy in explanation of why I’m avoiding certain methods:
Although region-level variables were available, they were not incorporated in this analysis because the methods required were developed by a woman
To avoid feminization of statistics, the clustering effects of school and classroom were not adjusted for in this analysis
Probability weights were not incorporated into the analysis, because that method was developed by Indians
I’m sure the peer reviewers will appreciate that, but just to be sure I’ll be sure to specify in all submissions that I not be reviewed by women. That should cover it.
Now, some of you might suggest that I should just relax and use all the techniques available to me, or at least not go through the canon with a fine-toothed comb checking the gender of every contributor – I mean, couldn’t I just drop the techniques only if I find out that they were written by a woman, without active screening? A kind of passive case-finding approach, if you will (but can I employ case-finding – it may have been invented by a woman. I should check that!) But this is not how the David Gilmour school works. You have to assess your authors first and foremost on their cool manliness:
Chekhov was the coolest guy in literature. I really think so. There’s a few volumes of his there, what a great looking guy. He is the coolest guy in literature; everyone who ever met Chekhov somehow felt that they should jack their behaviour up to a higher degree.
And really, when you look at the kinds of canon that are taught in English at high schools and first year uni courses, it is quite often the case that they are all (or almost all) male. Every statistician knows that those kinds of imbalances in a sample don’t happen by accident – that’s a deliberate selection bias. If it’s good enough for dudely English teachers it’s good enough for me, so I think from now on I should screen out any beastly feminized stats. Sure, you can’t get into any half-decent journals if you can’t use GLLAMM and good experimental design, but I say hell to that. It’s time to fight back! Men-only stats for the win!
In case anyone thinks I’m being serious, there’s been something of a storm of controversy about this David Gilmour chap, and I think you can see how stupid his approach is if you imagine trying it in a technical field. Stats being part of maths, it has its fair share of chick lit, but it is still male dominated; nonetheless, if you screen out the main work done by women, you suddenly lose a huge range of tools and techniques that are essential to the modern statistician. Surely the same applies in English literature, but moreso given the huge role women played in the development of the novel. Check this Crooked Timber thread for more entertaining take-downs of this position (with some prime grade Troll Meat thrown in the mix). It really is outstanding on so many levels that a literature teacher would judge who to teach in such a juvenile Boys Own Manual way; that they would take their responsibilities so lightly as to think that their sole task was to teach students their own opinion rather than … something useful … and that they would not try to hide it behind some more mealy-mouthed apologia. I mean really, there are a lot of very good female writers in the last two centuries and yet people like this David Gilmour chap manage to construct a syllabus without a single woman in it. Usually their argument would be along the lines of “I judge on merit” but you do have to wonder, don’t you? And then along comes a naif like Gilmour and makes it completely clear how these canons are really constructed – the women are screened out from the get go.
fn1: I really hope not, but this is the internet.