Student groups by academic ability

[Mercudenton]

I teach large 60+ survey classes. Student ability is very wide. I wondered this year about creating online and in-class discussion groups based on academic ability. I can see strong students stuck with weak ones when we do group discussions. As a strong student, I'd have hated that. I'd love the strong, engaged students to be able to talk together. I know there's a counter-argument that the strong students might pull the weak students up, but I wonder if this is idealistic.

So first question: anyone ever tried this?

Second question: how to achieve it? I asked registrar if I could have access to GPAs. Response was no (a) this would be FERPA violation and (b) it would not be appropriate to group students by GPA.

As for (a) I can see the issue, though I wonder if it could be classed under the FERPA "legitimate educational interest". I kind of feel professors should be able to have access to their students academic background; denying it is like denying a doctor access to patients' medical records -- it might be detrimental to the treatment. But I know FERPA is FERPA and must be obeyed.

As for (b) I tend to disagree. While I can see there is a pedagogical argument for and against, surely there might be merit in grouping students this way? We do actually have an honors and a remedial version of this class, so the institution has already kind of admitted that students might benefit from cohorts united by academic ability (and GPAs would be one of the key determining features here). Why not groups within a cohort based on similar principles (indeed, I don't really need to see the precise GPA, just a rough grouping of students).

Any alternative ideas? A pre-test? Voluntary disclosure of GPA? Or scrap the whole idea?!

[Hegemony]

I don't think GPAs are the way to go.  Just give them a quiz that helps them for some other reason — prepares them for an upcoming test or whatever.  Then divide half the class by putting together students with similar quiz scores, and half the class randomly.  Don't tell them you're doing this.  Keeps tabs on which is which, and evaluate whether it makes a difference.  You might even give them a little "group work evaluation questionnaire" at the end to get their impression of their experiences.  (Again, I wouldn't tell them what your variables were.)  Then you'll have a sense of whether it makes a difference.

[drbrt]

If you choose to do mixed ability groups, the research suggests you will get more out of pairings that are top-middle and middle-low rather than top-low. The gap in the t-l group is too big and it tends to cause the lower performing students to shut down and quit. For my classes, I use homogeneous groups, but I'm in math and I do group grades. It's a little easier to do performance groups in math, and the group grading means I get better buy-in from the top performers when they are grouped together. It also helps during class, because the learning assistants and I can really focus on the lower performing groups.

[spork]

I recommend against it. Odds are that you will end up with groups that are non-diverse in characteristics besides academic ability -- gender, ethnicity, SES -- and in the real world people have to work with others who are not like themselves.

Any group task should include components for which students are individually graded. Otherwise there will be free riders and students will resent you for it.

The experiment hegemony mentions: this would be an easily-publishable article in a pedagogical journal.

Your classroom environment sounds like it might be ideal for this system:

• Intersperse lecture with questions. Students respond to a question using clickers or some other technology.
• Display the results of students' answers to a question.
• Have students form pairs or triads and explain to each other how/why they arrived at the correct or incorrect answer.
• Quickly review the topic targeted by the question.
• Perhaps ask the question a second time (or ask a different question that targets the same concept) and display the results; students who still are not getting the correct answer should seek help from a teaching assistant/tutor.

[ciao_yall]

I don't think GPAs are the way to go.  Just give them a quiz that helps them for some other reason — prepares them for an upcoming test or whatever.  Then divide half the class by putting together students with similar quiz scores, and half the class randomly.  Don't tell them you're doing this.  Keeps tabs on which is which, and evaluate whether it makes a difference.  You might even give them a little "group work evaluation questionnaire" at the end to get their impression of their experiences.  (Again, I wouldn't tell them what your variables were.)  Then you'll have a sense of whether it makes a difference.

Except they figure it out really fast. To make conversation they ask each other questions like "how did you do on the last quiz?" Then they overhear a similar conversation... and realize why you purposely assigned them to work with each other. The strong students feel like they are being made to do the prof's work and the weaker students feel embarrassed.

For project groups, they worked on things they all chose to work on by interest. They were responsible for selling themselves to a group because they wanted to work on the leader's idea of marketing skateboards or Korean skin care products or whatever. And they were responsible for not getting themselves "fired" if they were really interested in that topic.

For informal groups I mixed it up. Geographic/proximity, counting off, color shirt you were wearing, month you were born, oatmeal raisin or chocolate chip, anything to get them talking to someone new.

[marshwiggle]

I recommend against it. Odds are that you will end up with groups that are non-diverse in characteristics besides academic ability -- gender, ethnicity, SES -- and in the real world people have to work with others who are not like themselves.

This raises a few interesting questions:

• Is there evidence that working in more homogeneous groups leads to better pedagogical outcomes?
• Is there evidence that working in more heterogeneous groups leads to better organizational outcomes?
• If the answer is "yes" to both of the above, how can you determine how to make the tradeoff, and should it be the same in every course?

[AvidReader]

I teach large 60+ survey classes. Student ability is very wide. I wondered this year about creating online and in-class discussion groups based on academic ability. I can see strong students stuck with weak ones when we do group discussions. As a strong student, I'd have hated that. I'd love the strong, engaged students to be able to talk together. I know there's a counter-argument that the strong students might pull the weak students up, but I wonder if this is idealistic.

There are already lots of good responses about grouping by ability. I don't do that often, but I do a lot of group work and I like to change things up a lot.

--When I teach a survey course, especially if I'm covering differing topics and/or give them some discussion autonomy, I often have them group by major/field/career. I've done this in some composition courses, where I've assigned articles in a few differing subjects and then had (for instance) all the nursing students discuss a medical article.

--I often try to put my most vocal participants together so that they won't dominate a group with quieter students. (In a big class, you can often figure this out by looking to see which students the others turn to when it's time to report in).

--Sometimes I give an informal quiz (e.g. "who is to blame for the death of Antigone and why?") and quickly count the responses (by hand-raising). Then I ask them to form X groups whose students have N differing responses and share their reasoning with one another. This has the added advantage of letting students who aren't verbally confident enter the discussion by reading their written responses.

--I've done many of the strategies ciao_yall suggests, and also, on a colleague's recommendation, started "counting off" this year. I go in a differing direction or pattern each time (front to back, left to right, spiral, snake, etc). 

AR.

[spork]

[. . .]

• Is there evidence that working in more heterogeneous groups leads to better organizational outcomes?

[. . .]
[/list]

Yes. Even the idiots at Harvard have figured this one out: https://hbr.org/2016/11/why-diverse-teams-are-smarter.

[Mercudenton]

Thanks for responses.

I think marshwiggle raises the crucial philosophical questions. What evidence is there either for homogeneous or heterogeneous groups? Also, we'd need to work out what our criteria for measuring "success" of such groups is.

It should be pointed out I am mainly thinking about discussions, not group projects to produce a defined assignment. Thus the heart of my concern is that students are able to enjoy meaningful, challenging and robust debate. In other words I am more concerned with the student experience of engaging questions and big ideas  than I am the student ability to produce a tangible product through their collaboration. In thinking about this I recognize I am clearly biased toward helping the strong students enjoy a better experience in a large survey class in which a large number of students couldn't care less. This is in part because our demographic is relatively low academic ability, and, anecdotally, the islands of stronger students sometimes get frustrated at the lack of interest in the topic shown by the majority of their peers ("It makes me so angry when no-one answers your questions and blows this class off" said one such student to me last year "why isn't everyone fascinated with this class?"). Also,  I went to a very mixed ability high school and was very frustrated at having to work with people who didn't care less or couldn't work at the same level.

In this sense I suppose I'm trying to create a classroom-in-a-classroom. I expect some would see this as elitist and indifferent toward lower-achieving students. I accept this is a fair critique, but again would suggest we do do this in colleges by having GPA entry requirements to begin with. I'd thus push back against Spok a bit when Spok says " in the real world people have to work with others who are not like themselves." This is true, but teaching people to work together is not the only point of academic discussions. My classroom is also the real world, and it in academia we do group people by academic merit, not to be elitist but because the act of academic discourse relies on certain shared understandings and skills in debate.  Isn't this why colleges have entry requirements? And why classes have pre-reqs? And honors programs have GPA requirements? And why professional conferences don't invite a freshman  to give the keynote?

In this regard the Harvard report is a bit of a red herring. It speaks about gender, ethnic, cultural diversity as bringing strengths. Of course. But it doesn't say that a business show hire people diverse in their ability to do the job for which they are hired. The Harvard report cannot be used to justify the superiority of a company of plumbers, some of whom are terrible and some of whom are good, to that of a company who employs only good ones. If the point of class discussion is simply to have good, in-depth, nuanced discussion, then surely the fundamental requirement for this task is some kind of ability to have a good, in-depth, nuanced discussion. Having established that, yes of course, it would be even better to have ethnic, gender, cultural diversity among those who are otherwise qualified. Of course, I get that that in reality, prior educational achievement and experience is pre-determined by the de facto socio-economic, ethnic structures of the United States, so it is not as simply as all this, and it may well be that, given the context, academic groupings end up also being ethnic and socio-demographic ones too.


Having said all that, I think it is hard to identify the exact way this would work, especially, as AvidReader points out, there are other characteristics that are important -- e.g. introvert/extrovert - - and students might quickly get suspicious of the whole process if they feel they've been zoned.

I'm therefore attracted to AvidReader's idea of grouping by major/career interest. I think this would achieve some of the effect I was looking for by grouping by "ability" in that it would cluster students most likely to be interested in certain styles of debate and interaction, without it being as individualized as a particular prior academic achievement or simplistically separating students based on an abstract notion of "ability". So the Lit/Philosophy students are most likely to engage within similar paradigms (; likewise the business students. There would also be no need for hiding the rationale behind the groups. I could be upfront about why students are in each group. This would also allow me as the instructor to set slightly different discussion topics on some issues in order to show the relevance of a gen ed. subject (that many think is irrelevant) to their major.

[Caracal]

GPA doesn't make much sense. If I'm teaching a class on philosophy why should I care that an engineering major has As. He might hate my subject or suck at it. Even in the class you're teaching, I don't think grades track that well with class participation.. It is true that the most of my students who do well on exams do tend to be more active participants and have more valuable contributions, and students doing quite badly in the class obviously aren't usually going to be very motivated, but it falls apart by the time you get to the B range. For me, this is the majority of my students. I have a lot of B students who are pretty checked out. They do enough to get by, but nothing else, and they don't show much interest in the class. But, I also always have students who get Bs who are interested in the class, want to participate and have interesting ideas. They aren't getting As either because they just don't put enough time into the class, or because they have trouble thinking in the way the discipline requires on assessments.

Dividing students by ability is a bad idea anyway, but since you're likely to not be differentiating between the largest group of students in the middle it seems pretty pointless.

[spork]

How, physically, are you going to have students engage in robust small-group discussions in a 60+ seat lecture hall?

And as caracal alludes to, how do you know that high grades 1) correlate with superior academic ability and 2) indicate willingness and ability to engage in discussion? A grade of A in a quantum physics course indicates far more academic ability than a grade of A in an introduction to marketing course. There are plenty of high GPA students who don't like to utter a single word in the classroom, just as there are usually a smattering of vocal idiots.

[Hibush]

hus the heart of my concern is that students are able to enjoy meaningful, challenging and robust debate. In other words I am more concerned with the student experience of engaging questions and big ideas  than I am the student ability to produce a tangible product through their collaboration. In thinking about this I recognize I am clearly biased toward helping the strong students enjoy a better experience in a large survey class in which a large number of students couldn't care less. This is in part because our demographic is relatively low academic ability, and, anecdotally, the islands of stronger students sometimes get frustrated at the lack of interest in the topic shown by the majority of their peers.

It sounds as if grouping by degree of caring about the class is more important than general ability or aptitude. There are various proxies you might use that would select the degree of caring.

[ciao_yall]

How, physically, are you going to have students engage in robust small-group discussions in a 60+ seat lecture hall?

And as caracal alludes to, how do you know that high grades 1) correlate with superior academic ability and 2) indicate willingness and ability to engage in discussion? A grade of A in a quantum physics course indicates far more academic ability than a grade of A in an introduction to marketing course. There are plenty of high GPA students who don't like to utter a single word in the classroom, just as there are usually a smattering of vocal idiots.

Maybe. Still, which helped you build your relationship with your toothpaste by creating a compelling brand narrative?

[Hibush]

How, physically, are you going to have students engage in robust small-group discussions in a 60+ seat lecture hall?

And as caracal alludes to, how do you know that high grades 1) correlate with superior academic ability and 2) indicate willingness and ability to engage in discussion? A grade of A in a quantum physics course indicates far more academic ability than a grade of A in an introduction to marketing course. There are plenty of high GPA students who don't like to utter a single word in the classroom, just as there are usually a smattering of vocal idiots.

Maybe. Still, which helped you build your relationship with your toothpaste by creating a compelling brand narrative?

I don''t think that is intro marketing, its the 300-level quantum marketing course.

Which brings to mind the idea that OP could build a compelling brand narrative around the discussion-group stratification method if it ends up working.

[Mercudenton]

How, physically, are you going to have students engage in robust small-group discussions in a 60+ seat lecture hall?
.

Well, as I noted part of this would be discussion forums online with groups. And the room I use has moveable tables and chairs. Not ideal, but better than fixed auditorium style seating.

how do you know that high grades 1) correlate with superior academic ability

Well, that's kind of the normal rough and ready expectation right? The D students have worse academic ability than the A ones? I get it's very approximate, and there's all kinds of variables, but surely we must expect some connection?!

A grade of A in a quantum physics course indicates far more academic ability than a grade of A in an introduction to marketing course. There are plenty of high GPA students who don't like to utter a single word in the classroom, just as there are usually a smattering of vocal idiots.

Well sure, but I'm talking about freshman students. No Quantum Physicists need apply. And to an earlier point, I do think a student getting As in Engineering would count as on average more likely to be ready to engage in a humanities subject than a student getting Ds. Your last point kind of speaks to what I'm trying to achieve, though. Why don't A grade students speak a word? I think there are a variety of answers to this but surely one might be that at least some A Grade students may be introverts who do not want to speak in a big setting or even (as this would have been me) want not want to stand out from others who they know are not as engaged or knowledgeable.