Attractive female students' grades dropped during remote learning: Study

[ergative]

The one thing I would never care about it what my surgeon *looks like*.
If he plays the scalpel like Coltraine on the sax, then I'm in (well, maybe a bit less *rough* than Coltraine, but you know what I mean). Don't give a darn what they look like.

\

'Ya know, if we take all this stuff seriously, we would ask for bad looking surgeons! They were ignored or discriminated against in med school, but made it anyway, so they must be better!

Excellent point.

I, too, am not surprised by this study. My institution forces us to grade anonymously--everything is submitted through Moodle under site-assigned participant numbers, and only de-anonymized after grades are assigned. Sure, for small classes in which students have discussed their work with me, I usually know who wrote what, but for larger classes, or assignments where everyone responds to the same prompts, I have often been quite surprised by who performed well and who performed badly. I've been tempted on many occasions to go back in and soften criticism of weaker assignments (although I don't raise the grades) when I learn that the student is someone who I 'know' is conscientious/bright/etc. In a microcosm demonstration of why internet comments are so toxic, my feedback could be much harsher when I didn't know the face receiving it. (Now I'm much better at giving gentle feedback to anonymous participant numbers, even if they turn in real stinkers, and when I've been doing a huge batch of grading on that mode, I find that my peer reviews tend to take on that gentler tone as well. Maybe Reviewer 2 grades anonymously too much.)

The point is, I've seen in my own head the difference between how I grade an anonymous assignment, and how I grade an assignment for a student I know. And, because all people are obviously the same as me, I simply don't believe that anything that can't be graded automatically can possibly be immune from this kind of influence. And if enough instructors pay attention to physical attractiveness (not you, of course, you would never be so shallow), that can easily shift the average enough to be picked up by this kind of study.

[Wahoo Redux]

The one thing I would never care about it what my surgeon *looks like*.
If he plays the scalpel like Coltraine on the sax, then I'm in (well, maybe a bit less *rough* than Coltraine, but you know what I mean). Don't give a darn what they look like.

\

'Ya know, if we take all this stuff seriously, we would ask for bad looking surgeons! They were ignored or discriminated against in med school, but made it anyway, so they must be better!

He was a very good surgeon, thankfully.

But yeah, his fanbase was a little surprising.

We are animals.  We just can't help it.

[clean]

Am I the only one that gives grades to students based on well-defined and objective performance metrics during the semester without taking into account their physical appearance?

I didnt read the article.
However, when I first read the heading and the first post (and reviewed the following ones) my thoughts turned to
'Good looking engineering woman' are likely to have plenty of opportunities to get help on homework - tutoring - whatever from Other Students (particularly the proverbial nerdy guy engineers in the class). 
Like the author of the post I copied, I make sure that I dont associate anyone's name with the papers I grade.  The student's name is on the first page, and I grade one page at a time.  (in that I grade page one for everyone, and then page 2 for everyone.... so that I am grading the same problems at the same time without knowing who the student's are, much less what the students look like). 

[apl68]

The one thing I would never care about it what my surgeon *looks like*.
If he plays the scalpel like Coltraine on the sax, then I'm in (well, maybe a bit less *rough* than Coltraine, but you know what I mean). Don't give a darn what they look like.

\

'Ya know, if we take all this stuff seriously, we would ask for bad looking surgeons! They were ignored or discriminated against in med school, but made it anyway, so they must be better!

Excellent point.

I, too, am not surprised by this study. My institution forces us to grade anonymously--everything is submitted through Moodle under site-assigned participant numbers, and only de-anonymized after grades are assigned. Sure, for small classes in which students have discussed their work with me, I usually know who wrote what, but for larger classes, or assignments where everyone responds to the same prompts, I have often been quite surprised by who performed well and who performed badly. I've been tempted on many occasions to go back in and soften criticism of weaker assignments (although I don't raise the grades) when I learn that the student is someone who I 'know' is conscientious/bright/etc. In a microcosm demonstration of why internet comments are so toxic, my feedback could be much harsher when I didn't know the face receiving it. (Now I'm much better at giving gentle feedback to anonymous participant numbers, even if they turn in real stinkers, and when I've been doing a huge batch of grading on that mode, I find that my peer reviews tend to take on that gentler tone as well. Maybe Reviewer 2 grades anonymously too much.)

The point is, I've seen in my own head the difference between how I grade an anonymous assignment, and how I grade an assignment for a student I know. And, because all people are obviously the same as me, I simply don't believe that anything that can't be graded automatically can possibly be immune from this kind of influence. And if enough instructors pay attention to physical attractiveness (not you, of course, you would never be so shallow), that can easily shift the average enough to be picked up by this kind of study.

In the past I've been somewhat skeptical of the push to make grading anonymous over concerns of bias.  But the evidence that it can make a difference does seem to be accumulating. 

[Caracal]

It's not really that essay assessment is "subjective". It's that people often don't grade these things anonymously. We're all inclined to give the benefit of a doubt to people we like, or who we know work hard, etc. And we're all more likely to be kindly disposed towards people we think are pretty. We're also more likely to recognize them, etc., which impacts "participation" grades.

But yeah, not surprising.

I think it is quite subjective.  When I am grading essay style questions, I am fairly certain my expectations/standards change as I am grading.  Your grade would most likely be impacted by the sequence in which you were graded.  Obviously I try to avoid this, but this is only one obvious example.  The nature of grading an essay is just not as clean as a multiple choice question where there is a single correct choice, there is too much discretion for it to be completely objective.

It isn't like the differences in grades were that big. If I'm reading the study correctly, the  perceived attractiveness premium is less than a tenth of a point on a gpa scale. Do a study of 600 students and you can pick it up, but the actual effect on any particular student is generally going to be pretty minimal. All kinds of things come into my grading that shouldn't, because that's just how it goes. Like Kron, I suspect the biggest ones are the order I grade in, but I'm sure there are plenty of other things. The goal isn't to eliminate this stuff, but to minimize it.

Unconscious bias is generally going to have small effects when people are assessing a discrete exam or assignment. The danger is when the thing being assessed is more amorphous. The classic example is the way the behavior of black boys in schools is perceived in ways that result in more and harsher discipline than for everyone else. In a college classroom, I'd be more worried about unfairness in handling of requests for extensions, absences, etc.

[ciao_yall]

Unconscious bias can also creep in for so-called objective factors. Do we grade grammar "errors" more harshly when they are from a student using African-American dialect? Do we set up True/False questions that might have implicit cultural bias?

[marshwiggle]

Unconscious bias can also creep in for so-called objective factors. Do we grade grammar "errors" more harshly when they are from a student using African-American dialect? Do we set up True/False questions that might have implicit cultural bias?

If our True/False questions are so open to interpretation that they can exhibit cultural bias, then we need to start making questions based on objective factors, so that  "True" means factually correct and "False" means factually incorrect. If that's not possible, then it doesn't belong in a "True/False" question.

[Kron3007]

Unconscious bias can also creep in for so-called objective factors. Do we grade grammar "errors" more harshly when they are from a student using African-American dialect? Do we set up True/False questions that might have implicit cultural bias?

Grammatical errors are objective in terms of rules etc but become subjective when grading an essay given that you need to assess it as a whole.  This is part of why grading essays is subjective and I try to avoid it.

Multiple choice and true false questions are more objective.  They can have issues with bias, but can be designed to minimize this.  Creating good MC questions is actually quite hard.  In contrast, essay questions will always have a degree of subjectivity in grading.  Perhaps the are necessary in many fields, but they do present problems and steps should be taken to recognize and minimize them.

[Caracal]

Unconscious bias can also creep in for so-called objective factors. Do we grade grammar "errors" more harshly when they are from a student using African-American dialect? Do we set up True/False questions that might have implicit cultural bias?

Grammatical errors are objective in terms of rules etc but become subjective when grading an essay given that you need to assess it as a whole.  This is part of why grading essays is subjective and I try to avoid it.

Multiple choice and true false questions are more objective.  They can have issues with bias, but can be designed to minimize this.  Creating good MC questions is actually quite hard.  In contrast, essay questions will always have a degree of subjectivity in grading.  Perhaps the are necessary in many fields, but they do present problems and steps should be taken to recognize and minimize them.

Yes, multiple choice or true/false questions don't do a good job measuring the skills that I want students to acquire in my classes. I don't want students to memorize a bunch of facts but not be able to understand how they fit together or why they are important. I want them to be able to make arguments and provide evidence for them. Subjectivity by itself isn't really a problem, I'd argue that it's built into my discipline. It wouldn't bother me if I give a paper an A- and a colleague would give the same paper a B+. If those differences were because of some inappropriate factor, that's different of course. But, really if you look at the study, it doesn't show that appearance has some huge impact on grades, so I'm not sure we really need to spend that much time worrying about instructor bias in grading.

[Kron3007]

Unconscious bias can also creep in for so-called objective factors. Do we grade grammar "errors" more harshly when they are from a student using African-American dialect? Do we set up True/False questions that might have implicit cultural bias?

Grammatical errors are objective in terms of rules etc but become subjective when grading an essay given that you need to assess it as a whole.  This is part of why grading essays is subjective and I try to avoid it.

Multiple choice and true false questions are more objective.  They can have issues with bias, but can be designed to minimize this.  Creating good MC questions is actually quite hard.  In contrast, essay questions will always have a degree of subjectivity in grading.  Perhaps the are necessary in many fields, but they do present problems and steps should be taken to recognize and minimize them.

Yes, multiple choice or true/false questions don't do a good job measuring the skills that I want students to acquire in my classes. I don't want students to memorize a bunch of facts but not be able to understand how they fit together or why they are important. I want them to be able to make arguments and provide evidence for them. Subjectivity by itself isn't really a problem, I'd argue that it's built into my discipline. It wouldn't bother me if I give a paper an A- and a colleague would give the same paper a B+. If those differences were because of some inappropriate factor, that's different of course. But, really if you look at the study, it doesn't show that appearance has some huge impact on grades, so I'm not sure we really need to spend that much time worrying about instructor bias in grading.

A well designed MC question can require much more than regurgitating facts, but I do recognize they are not suitable for all fields.  For me, they are good, but I also have other assignments in class that are much more subjective to grade.  I wouldn't argue that 100% of grades should be based on MC questions. 

The main thing IMO is to recognize the potential issue and do our best to minimize it.  Where I see a bigger issue with this is when you may have multiple TAs grading the same assignment and such. 

[Caracal]

Unconscious bias can also creep in for so-called objective factors. Do we grade grammar "errors" more harshly when they are from a student using African-American dialect? Do we set up True/False questions that might have implicit cultural bias?

Grammatical errors are objective in terms of rules etc but become subjective when grading an essay given that you need to assess it as a whole.  This is part of why grading essays is subjective and I try to avoid it.

Multiple choice and true false questions are more objective.  They can have issues with bias, but can be designed to minimize this.  Creating good MC questions is actually quite hard.  In contrast, essay questions will always have a degree of subjectivity in grading.  Perhaps the are necessary in many fields, but they do present problems and steps should be taken to recognize and minimize them.

Yes, multiple choice or true/false questions don't do a good job measuring the skills that I want students to acquire in my classes. I don't want students to memorize a bunch of facts but not be able to understand how they fit together or why they are important. I want them to be able to make arguments and provide evidence for them. Subjectivity by itself isn't really a problem, I'd argue that it's built into my discipline. It wouldn't bother me if I give a paper an A- and a colleague would give the same paper a B+. If those differences were because of some inappropriate factor, that's different of course. But, really if you look at the study, it doesn't show that appearance has some huge impact on grades, so I'm not sure we really need to spend that much time worrying about instructor bias in grading.

A well designed MC question can require much more than regurgitating facts, but I do recognize they are not suitable for all fields.  For me, they are good, but I also have other assignments in class that are much more subjective to grade.  I wouldn't argue that 100% of grades should be based on MC questions. 

The main thing IMO is to recognize the potential issue and do our best to minimize it.  Where I see a bigger issue with this is when you may have multiple TAs grading the same assignment and such.

Yeah, I never have TAs, but I recall that when I was a TA we sometimes would take a few exams at the beginning of the semester, and sort of calibrate our grading standards. To some extent, I do that when I'm grading by myself too. When I grade out of class essays, I pencil in a grade for the first four or five essays and then come back to them and look at them as a group. Partly, that's because I usually find that I'm being harsher about the first few essays I look at than I want to be-a student paper has to be pretty good to not be painful to read. It also helps me make sure I'm being consistent, however, and also that I'm not factoring things into grades that I shouldn't be. Usually this comes up when I'm grading something where I find the argument to be sexist/racist/generally messed up. It helps to have points of comparison so I can say "ok, this paper is no worse in its argumentation than this other paper that I gave a B to.

[apl68]

Yeah, I never have TAs, but I recall that when I was a TA we sometimes would take a few exams at the beginning of the semester, and sort of calibrate our grading standards. To some extent, I do that when I'm grading by myself too. When I grade out of class essays, I pencil in a grade for the first four or five essays and then come back to them and look at them as a group. Partly, that's because I usually find that I'm being harsher about the first few essays I look at than I want to be-a student paper has to be pretty good to not be painful to read. It also helps me make sure I'm being consistent, however, and also that I'm not factoring things into grades that I shouldn't be. Usually this comes up when I'm grading something where I find the argument to be sexist/racist/generally messed up. It helps to have points of comparison so I can say "ok, this paper is no worse in its argumentation than this other paper that I gave a B to.

I recall when I was a TA having done something similar with some of the classes that I TA'ed with.  To make sure all the assistants and the prof were on the same grading page.

[Myword]

How can they judge or know which females are attractive? Attractive to whom?

Assuming this is true, which I doubt, it suggests that higher grades are given to attractive females in class by male professors, but when the same students were not visible, they got lower grades--the grades they deserved, probably.
Might also be true with handsome men

This survey does not look legit. Probably a feminist attempt to show that women get discriminated. An agenda. Well, actually studies also show that attractive people are more likely to get better jobs, advantages and nice grades. ( PRETTY lookers can count as"class participation" points)  Obviously, no one admits that appearances are part of a grade but it must happen.  there is no inherent or innate correlation between looks and grades.

[marshwiggle]

How can they judge or know which females are attractive? Attractive to whom?

As it says in the article, they showed student ID photos to a bunch of students who rated them.

Assuming this is true, which I doubt, it suggests that higher grades are given to attractive females in class by male professors, but when the same students were not visible, they got lower grades--the grades they deserved, probably.
Might also be true with handsome men

As it notes in the article, some of the profs were female as well. They didn't point out any difference between male and female profs in the results, so either they didn't look for it specifically or it didn't appear. (In my recollection, when someone studied how hiring committees responded to male and female names on an application, male and female committee members has similar biases; it wasn't dependent on the sex of the person doing the evaluation.)

This survey does not look legit. Probably a feminist attempt to show that women get discriminated. An agenda. Well, actually studies also show that attractive people are more likely to get better jobs, advantages and nice grades. ( PRETTY lookers can count as"class participation" points)  Obviously, no one admits that appearances are part of a grade but it must happen.  there is no inherent or innate correlation between looks and grades.

The whole point of the idea of unconscious bias is that it's unintentional. No-one sets out to give higher grades to certain students for reasons that have nothing to do with their performance, but as long as there is some subjectivity in how we evaluate people, we can't help letting things affect our evaluations that, in principle, shouldn't.

[Ruralguy]

This is actually a well studied phenomenon in many arenas. Why would it be a surprise in academia?
I think we especially don't like it when someone suggests we may not be 100% pure of thought. Oh gosh, we
might have a bias and one we aren't even aware of! Not us! Can't be!