Monday, June 17, 2013

Confessions of a (slightly below) average writer

I re-took the GREs last winter (my third time), partially for a lark, and partially in preparation for a harebrained scheme to get a second doctorate if I can't find a faculty post.

Anyway, despite the fact that the format has been completely revised since I last took it, and the scoring totally different, I got exactly the same results, percentile-wise, on the Verbal and Quantitative sections that I did before. But, of course, I had to wait several weeks before I got the score from the Analytical Writing test. I didn't worry about it, though, because my writing is one of my strongest skills.

And then I got the score.

4.0 out of 6.

49th percentile.

Bear in mind the last time I took the test, before my written rhetoric had been honed by five years of grad school, I scored a 5.5. My prose has never been described as anything but excellent, clear, and entertaining. As an undergraduate, I seriously considered a career writing fiction. And, not to brag, but I got a perfect score on the Verbal section the last two times I took it. So it came a something of a shock to discover that, according to ETS, my writing is (slightly below) average.

I have long known that the writing assessments of standardized tests are flawed. I had students who had tested out of their freshman comp requirement because of their writing SAT scores, but who couldn't construct an argument if they bought it from Ikea. And based on the ETS website, the written sections are only scored by a single human reader and a computer, unless the computer's score differs from the human's. Forgive me, but I am a little skeptical that language processing algorithms have reached the point that they can evaluate the quality of an argument.

My point here is not to gripe. It's to illustrate how problematic it is for us to rely on standardized tests to predict how students will perform in graduate programs. If a published scholar holding an advanced degree in the humanities only scores a 4 out of 6, the metric is useless. I assume (or at least hope) that this score is irrelevant for applicants with strong essays and writing samples, but considering the cost of the test (not to mention test prep classes), the test's failure to accurately predict one's ability to produce graduate-level writing is troubling.

For what it's worth, I shelled out another $60, in order to have my essays regraded by hand. And I got the same score. I intend to re-re-re-take them again in the fall, after I've had time to study their mechanical grading rubric, so the next time the essay asks me analyze the statement, "One should always consider the consequences before taking risks," instead of providing a well-crafted argument about the pitfalls of teleological ethics, I'll just say risk is bad because it's risky...

No comments:

Post a Comment

All comments are moderated by imperial fiat. You have the right to say what you want, but I don't have to give you a forum. Don't be a schmuck: that is the whole of the Torah. The rest is just commentary.