This happens in reverse too. Teachers assume high achieving students will continue to achieve and give them more slack.<p>The trick to all my "writing" classes was to simply put the most effort in the first assignment. It was pretty clear that future marks were a function of prior marks, and I tested this hypothesis in different classes by modulating my effort on various assignments throughout the year.
I know I tended to rate my foreign" instructors poorly. Not that I want to be dislike them, but when I have trouble understanding their accent it is hard to learn from them. If I ever have to give a lecture in a foreign language I'm sure my students to have the same reaction (of course as a human you can be sure that if I actually spoke a language well enough for this to be reasonable I'd consider my accent perfect)
Would it be that hard to judge professors based on students success (in subsequent classes) using something like a pagerank algorithm? But that wouldn't necessarily correlate with institution popularity and success.<p>Academia seems to be becoming more of a customer service-oriented business where customer (student) perception is more important than the value/quality of the service provided.
I know how this is going to sound, but I'm genuinely curious: how do they measure whether the evaluations are gender biased, or whether the quality of the instruction is not equally distributed between the genders?<p>Before I get a bunch of hate and downvotes, I'm not suggesting that female teachers get lower evaluation scores because they're not as good on average as their male counterparts. But I assume that researchers can show <i>why</i> that's not the underlying source of the difference in ratings, since that would be an obvious objection to these findings. Just curious how they separate that out, that's all.<p>EDIT: nevermind, this was just laziness on my part. Here's the study, and it's actually not that hard to understand: <a href="https://www.scienceopen.com/document_file/25ff22be-8a1b-4c97-9d88-084c8d98187a/ScienceOpen/3507_XE6680747344554310733.pdf" rel="nofollow">https://www.scienceopen.com/document_file/25ff22be-8a1b-4c97...</a><p>TLDR: In one experiment, they randomly assigned students to a TA and compared how well the students did in the class to determine whether the instructors were of comparable quality. In another experiment they secretly switched the TA identities in an online course so the students who had the female TA thought they had the male TA, and vice versa. Very nice.
I spent one semester teaching at a major state university, many years ago. Here are some anecdata:<p>* I shared an office with a guy who was team-teaching a course with a professor. My office mate was an adjunct, the professor was coming up for tenure. They fought bitterly over the grading of the course. The prof wanted to give easy grades because he needed to boost his student evaluation scores.<p>* It was common knowledge that the best predictor of evaluation scores was the grades that the students expected to receive. It was as if the system was rigged to encourage grade inflation. That was 20 years ago.<p>* Female teachers received in their evaluations, among other things, violent threats.
I knew one CS professor who always had very good evaluations.<p>His trick? He brought free cookies on the day he was going to get evaluated.<p>He swore that it worked.
A really simple indicator of this the long term relevance of "hotness" in things like ratemyprofessor - mercifully they killed that, but it should never have been present in the first place - which showed that "hotness" was strongly correlated to a professor or TAs rating.<p>Couple with other studies that hotness ratings of women are more extreme (tending towards a binomial hot vs. not), and more negative (I would assume, but am not sure of actual real study to back it up, because "average" hotness gets demoted).
Is there an actual solution proposed here or is this just full of hot air?<p>He says adjusting can't work and that explaining bias doesn't do the job, so what is he suggesting? Don't consider a professors performance at all? Add some sort of standardized testing system?
Education is screwed up in many ways. Colleges would rather look good in the US News rankings than educate their students well. Students would rather have attractive professors than ones who teach them effectively. And yet the whole thing is massively popular and expensive.