Discussion Circles

Apologies to whoever I stole this idea from – I don’t remember who I should be crediting with it. It has, however, become a staple of my classes.

Previously, class discussions that I’ve worked into lessons have had problems. If the whole class tried to have a discussion together, a few very vocal students dominated the arena while others either tried in vain to compete or happily ceded the floor and retreated into themselves. If discussion groups were smaller, it was harder for non-participants to avoid notice, but discussions still depended on the willingness of a few people to keep a conversation going to prevent them from dissolving into a group of people sitting together, each checking his or her phone. Even groups that stayed on task would default to talkative students talking more and quieter students nodding along.

Discussion circles are a way of facilitating equally participatory conversation among students who naturally vary in their willingness to speak as themselves and voice opinions on either academic or familiar topics. They do this by:

  • Removing some of the burden on the students of representing themselves, because they are playing assigned roles rather than simply voicing their own thoughts,
  • Supplying pragmatically appropriate language, and
  • Encouraging participants, in various ways, to listen carefully to and respect each others’ contributions.

I use 3 versions of Discussion Circles sheets, each of which has 4 roles that participants need to play:

  • Discussion Leader
    • Chooses questions to ask and asks them
    • Begins and ends the meeting
  • Harmonizer
    • Thanks other members for participation
    • Asks for clarification
    • Rephrases others’ opinions
    • Encourages other members to participate
  • Reporter
    • Takes notes on the members’ contributions
    • Asks members to repeat or rephrase
  • Devil’s Advocate
    • Disagrees with other members’ contributions (constructively!)

These tasks are in addition to actually answering the questions that the Discussion Leader asks.

Each of these roles has a worksheet to fill out with sections for before, during, and after the discussion. These are turned in to the teacher afterward. The teacher, incidentally, is not involved in the discussions except to provide a list of questions and assign roles at the beginning.

The version of the worksheet that I use for at least the first 3 times that I do this activity is about 2 pages long per member. The “Before” and “After” sections are fairly involved and take about 10 minutes to do each. (The discussion itself can take anywhere between 20 minutes to an hour.)

You can get a copy of it here: Discussion circles online (called “online” because it is in a format that is easily distributable on Google Classroom. You can also print it.)

After they are used to the expectations of each role, I use a shortened version of the sheet. This one has a shorter “Before” section and no “After” section.

Find a copy here: Discussion circles lite online

Towards the end of the semester, I use a Turbo version of the sheet in which the participants switch roles with every question.

Get it here: Discussion circles turbo

In my Center for Excellence in Teaching and Learning group that meets on Fridays (basically a community of practice for new professors), I tried a revised Turbo version that had the job Quoter replacing Reporter.

Get it here: Discussion circles turbo 2

I give two grades for this assignment every time it is used: One grade for participation in the meeting and one for completing the worksheet. Now that we’re all online, the participation grade comes from a predetermined member recording their Zoom meeting and sharing the video with me.

Obviously, for the last few weeks of our spring 2020 semester, I’ve been distributing these online and having students share one sheet for the whole group rather than printing and handing out the sheets for an in-class discussion. I find that the distribution of responsibility in Discussion Circles, where everyone has to participate in order to complete their own sheet, suits the slightly impersonal nature of online synchronous discussions fine. Students often remark that they take more easily to some roles than others, but I try to make sure everyone plays every role at least once, so that even if they don’t “naturally” like to disagree with others, they will all be able to do so respectfully when it becomes important.

I find that Discussion Circles are a helpful scaffold for a lot of skills practice that we hope to see in class discussions, to the point that I rarely have a class discussion without them anymore. I hope you get some value from them too.

Thresholds in missed assignments predicting final grades

First, a note on a quick research project that yielded nothing interesting: I checked whether Canvas page views immediately before going remote, immediately after going remote, and the change in page views during that period were correlated with final grades, and they basically weren’t. I was wondering whether students who tended to check Canvas a lot (during the remote instruction period and before it) tended to do better in the class overall, and I didn’t find any evidence of that.

Also, and this will be mentioned again later, in checking the average scores for assignments over the past semester, I noticed that assignments that had to be done in a group were more likely to be completed than those that weren’t. This is interesting to me because setting up a Zoom meeting and talking to classmates, sometimes in other countries, would seem to be harder, not easier, than completing a worksheet by oneself. However, just taking two types of assignments from my Written Language class as examples, Reading Circles, which had to be done as a group via Zoom (or in person before we went remote in March), had a mean score of about 93% for the term, while Classwork, which included many assignments that were completed solo, had an average of 89%. In the Oral Language class, Discussion Circles (a sort of role-playing exercise with questions on an assigned topic) had an average scores of 99%, and Classwork 90%. It seems that Zoom meetings and the rare chance and synchronous interaction that they represent facilitate work, despite the pain of setting them up.

In other news, I have just completed my first academic year at the university IEP that I started at full-time last fall. As a celebration we got Thai takeout from one of the three good Thai restaurants in town (there are, mysteriously, no good Indian restaurants for 40 miles in any direction), and I immediately started blogging, vlogging, and tinkering with Google Sheets to fill the void left by work.

I’ve been slowly adding functionality to the Google Sheets that I use to do my end-of-course number crunching, mostly by figuring out new ways to use the FILTER function along with TTEST to see if there are statistically significant differences in my students’ final grades when they are separated into two populations according to some parameter. I put together a master Sheet for the year that included all of my classes between last August and now.

One possible factor that I had noticed anecdotally throughout the year was that students seemed more likely to fail or do poorly for assignments not turned in at all than for assignments done poorly. There was no shortage of work that was half-finished or ignored instructions, but the really low grades for the course were usually for students with work that was not even turned in.

So I set up a t-test on my Google Sheet to separate my students into two populations by the % of assignments that received a grade of 0 and look for a statistically significant difference in their final grades. Naturally, one expects students who have more 0s to do worse, but I still wondered where the dividing lines were – did getting 0s on more than 5% of assignments produce statistically significantly different populations? Did 10% do the trick? Is there a more graceful way of expressing this idea than “statistically significantly different”?

The relevant cells in my Google Sheet look like this:

As you can maybe figure out from the above, missing 10% of assignments (regardless of the points that those assignments were worth) produced a statistically significant difference in final grades: those who missed 10% or more of assignments had a final course grade of 66.9% (or D) on average while those who missed less than 10% had an average course grade of 90.8% (or A-).

On the other hand, getting full scores (which in my class means you followed all the directions and didn’t commit any obvious mistakes like failing to capitalize words at the starts of sentences) on more than 50% of assignments also produced a statistically significant difference in final grades: those who got full scores on 50% or more of assignments had a final course grade of 93.2% (or A) on average while those who got full scores on less than 50% had an average course grade of 78.4% (or C+). This isn’t the difference between passing and failing, but the ratio of full scores does produce two populations, one of which fails on average and one of which passes – see below.

Other significant dividing lines were:

  • Missing 3% of assignments
    • If you missed more than 3%, your average grade was 83.3% (B)
    • If you missed 3% or less, your average grade was 92.1% (A-)
  • Missing 5% of assignments
    • If you missed more than 5%, your average grade was 78.6% (C+)
    • If you missed 5% or less, your average grade was 91.7% (A-)
  • Getting full scores on 35% of assignments
    • If you got a full score on more than 35%, your average grade was 90.0% (A-)
    • If you got a full score on 35% or less, your average grade was 68.0% (D+)
  • Getting full scores on 70% of assignments
    • If you got a full score on more than 70%, your average grade was 96.6% (A)
    • If you got a full score on 70% or less, your average grade was 85.0% (B)

As you can see, I am not a prescriptivist on the use of the word “less”.

As you can also see, there are some red lines that pertain to the number of assignments that students can miss before they fall into a statistical danger zone: 10% of assignments missed, or only 35% of assignments with full scores. A student who fails to meet these thresholds is statistically likely to fail.

Statistics like these don’t carry obvious prescriptions about what to do next, but I worry a bit that the number of missed assignments will go up as classes are moved permanently online and assignments lose the additional bit of salience that comes from being on a physical piece of paper that is handed to you by a physical person. I also, for mostly bureaucratic reasons, worry that my grades seem to reflect less “achieving learning outcomes” and more “remembering to check Canvas” – although I’m sure this discrepancy is nearly universal in college classes.

I am considering giving fewer assignments per week that are more involved – fewer “read this article and complete this worksheet” and more “read this article, make a zoom discussion, and share the video and a reflection afterward”. We will see if that produces grades that reflect the quality of work rather than the mere existence of it.

Academic ESL and interlanguage: Partially totally effective or totally partially effective (or effective for other purposes)?

Three hypotheses for the observed effectiveness of academic ESL for preparing students for academic work in English:

  1. Academic ESL is perfectly effective at developing interlanguage, but academic ESL classes finish before the end of interlanguage development because students cease being ESL students and matricutate into regular degree programs. Students would still benefit from academic ESL after this point, but rarely have time due to their undergraduate or graduate class schedules. Some stunting occurs in students’ interlanguage because of the premature end of their ESL courses.
  2. Academic ESL is partially effective at developing interlanguage, and academic ESL classes finish at the end of their period of effectiveness. Students would not benefit from more academic ESL after this point because interlanguage development cannot occur through further academic ESL classes. Students are more likely to have student interlanguage development because of excessive time spent in ESL than a premature start to their degree programs.
  3. Academic ESL is partially effective at developing interlanguage but mainly effective at introducing compensatory strategies for students to use to make up for their lower language skills. Some of these strategies are specific to language learners and others are of use to any college student, but former ESL students in degree programs succeed by using them more than other students. Interlanguage development is less predictive of academic success than application of compensatory strategies.

Earlier this semester, we requested some data from our campus researcher, and he just got back to us. I won’t say what exactly he told us, but it pertained to average GPAs among different populations of undegrads, and it was good news for the apparent effectiveness of our IEP.

That said, we don’t know why our IEP appears to be effective. It is possible that we are getting better at our jobs. It is also possible that we are just recruiting better students. It’s possible that our students are far better than average, but we’re doing a worse-than-average job preparing them for college, resulting in performance that converges on the mean. Assuming that the work we do in class is at least part of the reason, it might help us to better focus our efforts in order to improve even more if we knew what part of what we do in class helps our students the most.

(For most of my career, I was used to the idea that interlanguage development started when students joined my class and stopped when they quit. In EFL, you can’t count much on outside factors to keep the interlanguage development ball rolling – students aren’t part of formal or informal organizations that facilitate regular English use and their identities accommodate English as a hobby at most. I tried as the owner of an eikaiwa to get students to start pastimes that included English, only to realize that as an eikaiwa teacher, I was the pastime. In short, I was used to thinking of English class as a self-contained unit; anything I wanted my students to do with English we had to do together.

I realized partway through my first year teaching community college ESL in California that we were by design only giving our students a partial education. We wanted to send them off into English 100 with maybe a bit of a head start and without a lot of baggage, but we expected English 100 to continue the work of interlanguage development. I’m sure some of us thought that ESL would still benefit our students, but they had to get on with their credit-bearing classes eventually, and some of us probably thought that ESL was inherently limited in what it could accomplish. There are also those who think that the one and only way that a student will come to understand adjective clauses is if the teacher explains adjective clauses and have never heard of interlanguage.)

Anyway, this would make a good long-term study project: find a decent sample of former academic ESL students in their undergrad years, give them the TOEFL or IELTS (which they wouldn’t have taken in a few semesters at least), survey them on their “compensatory strategies” (defining those would be a lot of work), and measure those against their undergraduate GPAs.

By the way, I’ve started recording some old blog posts as vlogs, seeing how different people tend to read ELT blogs and watch ELT-related content on YouTube. Feel free to stop by and leave a comment about how I don’t look like you expected.

Dr. Brute Facts doesn’t care about your feelings

Of all of the world’s doctors, and particularly at St. Jude’s Hospital, Dr. Judd Shapiro is a physician with a rare gift. While most doctors consider the physical body a vessel for a living, thinking person first and foremost, and the apperception and treating of the physical body a worthy enterprise because of the comfort that it might bring such a person, Shapiro considers the person to be a mere collection of physical phenomena, of value only as the sum total of the stuff of which it is made. To be even more precise, “it”, the physical body, barely exists to him; it demands no more recognition of oneness or continuity than does the air in a balloon or the water that makes up a particular section of river. This is not just a matter of point of view, but more of the acuity of his senses: he sees the interactions of particles and waves in front of him much like you or I see a fly landing on a window, and thinks primarily on that scale, the scale that is most intuitive and meaningful to him. The scope of his thought can be widened to include constructs we might call “life”, including human life, and on some occasions he be persuaded to adopt our conventions of grouping matter into “cells”, “organs”, and “people”. In this way he has a grasp, in both the general and particular senses, of the true inner workings of the human body, and any body that happens to be in his examination room. He knows both the sequences of DNA (that is, can recite the base pairs) that commonly predispose one to abnormal hemoglobin production as well as the precise location and corpuscular hemoglobin concentration of any given red blood cell in the bloodshot, overworked eye of any of the nurses. The difficulty of working with him lies not in his intellect but where he is inclined to focus it, in convincing him to think of the physical phenomena that he is attuned to as part of a “body” or a “patient” at all. To him, arrangements of matter in the coordinated pattern we call a “human” are but a convenience, a mental shorthand, for those without direct access to the atomic motion that underlies it all. When his opinion is solicited, he has the habit of emerging from a trance-like state to declare, “facts, not feelings” – which, as far as we can tell, is an enjoinder to let go of our parochial attachment to bodies and minds of the patients under our care and focus on simply what is objectively there. To him, a “body” is a random selection of the possible arrangements of matter, only intrinsically appealing to humans because we appreciate things of roughly our own size, and in particular things whose cells contain DNA recombinable with our own. He seems to see no point in putting his answers to us, when we can get them, in body-scale terms, instead phrasing them in the the brute factual language of nature. They often come in descriptions of the states of electrons somewhere nearby, which would be useless to us even if it were clear that these electrons were in the patient’s body and relevant to his complaint. If we manage to elicit a comment on a particular condition, we find that just as often as not he has taken the side of the disease over the patient. At one point, a fellow doctor contributed weeks of study to discern the meaning of one off-hand comment Dr. Shapiro had made about a T being where a C should be (we were lucky that in this case he favored us with an explanation in terms of molecules rather than quanta), thinking that it might hold a clue as to the particular mutation in liver cells that led to a patient’s hepatoblastoma, only to find that Dr. Shapiro’s remark was about the liver’s healthy cells – Dr. Shapiro seemed to see no reason that the patient’s whole liver should not be cancer. Apparently, cancer displays all of the hallmarks of what we call “living”, with the added bonus of not disturbing his trances with irrelevant questions.

I wonder if it would not be a terrible burden on his gift of perception to favor the intuitions of others as to the relative value of different arrangements of matter, seeing that he accepts the incomparably more arbitrary construct that is “employment” as a “doctor” and the remuneration it brings.