Seminar on Active Learning pedagogy

I’ve joined an Eberly Center seminar / reading group on the educational approach known as Active Learning (AL).

[Obviously our Statistics department should use Active Learning, the educational technique, to teach a course on Active Learning, the approach to experimental design. We can call it Active Active Learning Learning.]

The basic idea in AL is to replace traditional lectures and “passive” student behavior (listening quietly to an instructor speak, perhaps taking notes to transcribe the lecture) with almost anything more “active”: discussion with your neighbor or the whole class; short clicker quizzes; labs or larger projects; etc. The goal is to help your students stay attentive, motivated, and engaged, so they are constructing and synthesizing knowledge throughout class—instead of just being passive receptacles for the words you say.

I’ve tried variants of AL when teaching, and generally I liked the outcomes, but I hope the reading group will help me think through some challenges I had in implementation.

Notes-to-self below, posted in case any readers have thoughts or suggestions:

Week 1

Video example and engagement

We watched an example video of Deborah Ball teaching a class using Active Learning. It was a nice demo overall, but even during a good discussion among Ball and several students, you can see that many students still seem not to be engaged in class: slouching, nodding off, checking their laptops…

Does AL really increase the engagement rate in class? Or does it just help improve the learning of those students who’d be engaged anyway?

We discussed strategies to keep student engagement high. You can require everyone to write something short and turn it in, or vote with clickers or show-of-hands, or anything else that’s a low-bar-to-entry activity that’ll demand at least a little thought and attention—even if they choose not to take part in group discussion afterwards. (See also “act one” of Dan Meyer’s three-act approach to teaching math: Hook them with a clear question where anyone can hazard a short guess/opinion. Here’s also his list of ways to ask for an informal answer before the formal one.)

Or, you could cold-call on students. This should be done with care; you want to make it random, not pick on the dozing kids, so that it’s clearly not intended as punishment. But if you do it right (and the students do feel it’s fair), it can keep them awake and on their toes, knowing that they might be called on at any moment. If doing this, you have to start on day 1 and set it as an expectation for the whole course. To ensure randomness (and learn students’ names), you could print names and photos on index cards that you shuffle. (If grading participation, you can mark whether they responded on this index card.) Finally, it’s less anxiety-producing if you ask questions that have many good answers, not just one single correct answer.

More examples, note-taking, and challenges

Next, we read Chapter 8 from Bowen’s Teaching Naked. Bowen’s point is that we should use all our fancy edu-tech outside of class, so we can optimally spend the precious, limited, face-to-face class time on in-person interaction among the professor and students. And during that time, lectures can have their place, but let’s not default to using lectures in all situations. Think about whether a lecture would really work better than other activities for your goals for that class session.

However, I disagreed with Bowen’s advice that “Lectures do their best work when students do not take notes.” Maybe he means simple transcription of a fast-talking speaker, in which case, yes, it’d be better to be given notes in advance. But otherwise, I use note-taking as part of my own thinking and listening process during class or meetings, even if the lecture slides are already in front of me. I make notes such as: What points are most important and relevant to me? What tangential questions do I have, to be asked later? What gaps or connections do I see, when it’s all laid out on a page at once, which I might otherwise miss during the linear sequence of the talk? Writing is a tool that helps me process the lecture in real time.

We discussed what this means for “scaffolded” notes, where the instructor gives students notes in advance with a few blanks in them, to be filled in during class. If those blanks are just things to be transcribed exactly (say, an important math formula), then maybe this incentivizes class attendance but doesn’t necessarily help with learning otherwise. Every person would write the same thing. On the other hand, it can be more helpful if the blanks are for class discussion: you ask a broad question and write various students’ responses on the board. Then the blank space in the notes really becomes an opportunity for each student to synthesize those responses and process the discussion in their own terms.

Bowen also advises pausing for 1-2 minutes several times during each lecture, letting students digest the information and “compare and clarify their notes.” Someone pointed out this is especially helpful for students whose first language is not English (or whatever the class is taught in), who will need extra time to process.

Some other challenges in AL that came up or that I’ve run into:

  • Where do you find a bank of good questions to ask and discuss or activities to do—especially low-bar-to-entry questions? I haven’t seen such a question bank for Statistics yet.
  • How do you balance time spent on pair/group discussions? Often I see some groups are done and bored, while others are still having a lively discussion. When to cut them off and bring it back to the whole class?
  • How to effectively do outside service-learning projects that interact with the community? Are there rules of engagement for volunteer projects with sponsors outside the university? Is there some equivalent of IRBs for this?
  • How does AL work in distance learning, when you never have in-class face-to-face meetings?

Scientific evidence that AL works

We read a couple of articles exploring the evidence for success of AL, at least in STEM fields.

First, Smith et al. (2009) was a study of peer discussion in a genetics course. If you ask a multiple-choice question; have students vote; ask them to discuss their answers with a partner; and have them vote again (before you’ve told them the right answer)… then the proportion of correct answers tends to go up. But is this because they really learned from the discussion? Or just because they got a more-knowledgeable friend to tell them the right answer?

Smith et al. tested this out by adding a second “isomorphic” question (same topic, similar but distinct question). So students voted (with clickers) three times: Q1, Q1ad (“after discussion”), and Q2. The instructor didn’t tell them any correct answers, or show any class vote results, until after they answered Q2. It’s not surprising that Q1ad was correct more often than Q1; but Smith et al. found that Q2 was also correct more often. Presumably the students really learned something from the discussion of Q1, and were not merely being told the right answer.

There was a really nice quote from one of the students: “Often when talking through the questions, the group can figure out the questions without originally knowing the answer, and the answer almost sticks better that way because we talked through it instead of just hearing the answer.”

On the other hand, students do tend to sit next to similar students, so the answers around you might not be all that diverse. Could it be even better if you enforce random seat assignment, or get them to stand and shuffle around between questions, or somehow use cold-calling throughout the room?

The other study we read, by Freeman et al. (2014), was a meta-analysis of 225 studies that each compared traditional lecture to some flavor of AL. I found this a little more problematic because they averaged over so many different variants of AL, from occasional clicker questions or worksheets to full-on studio classes. They also didn’t really describe whether the “traditional lecture” controls in each study were all that comparable. Was it usually a comparison of AL vs. really good lectures? Average lectures? Mediocre lectures?

Still, most individual studies showed improved test scores and reduced failure rates under AL (though a few did estimate AL to be worse); and the average across all studies (whatever that means for such a heterogeneous set) was a sizeable improvement under AL. I also like that the meta-study authors took care to translate their averaged, normalized effect-sizes back into meaningful terms.

I haven’t seen many studies of AL in teaching the field of Statistics, but Freeman et al. mentioned they found a few. Here’s what I could see in their online supplemental info:

  • Zieffler et al. (2008), “What Does Research Suggest About the Teaching and Learning of Introductory Statistics at the College Level? A Review of the Literature.”
  • Giraud (1997), “Cooperative learning and statistics instruction.”
  • Keeler & Steinhorst (1994), “Cooperative learning in statistics.”

They might be worth a read.

In defense of lecture

Finally, we read Worthen’s New York Times op-ed, “Lecture Me. Really.” Worthen takes a strong stance against what she calls the “active learning craze.” She also seems to conflate it with “a broader crisis of confidence in the humanities,” given that much AL research seems to come from STEM fields.

I agree with Worthen that traditional lectures hone valuable skills: concentration, attention, “absorbing a long, complex argument,” listening “with an analytical ear,” thoughtful note-taking, etc.

On the other hand… Maybe [paying attention as someone talks for 60 minutes nonstop] is indeed best practiced in traditional-lecture format—but are you going to assess that skill too, in your assignments and exams? Is that skill more important than the content of those lectures? What about students whose first language is not yours, or with learning disabilities, students who legitimately cannot keep up with your hour of nonstop talking?

Otherwise, many AL techniques hone these same skills too, and so do all STEM fields. Does lecture have a monopoly on practicing careful attention? (You can’t be haphazard at the bench in the lab sciences.) Do the humanities have a monopoly on teaching reasoning and synthesis? (You can’t create a mathematical proof without knowing exactly what is and isn’t relevant.)

So, I was disappointed that Worthen didn’t have a better defense of lectures or the humanities. Both are valuable, and they don’t need to be pitted against their “competitors.” I want to teach in a world where it’s OK to choose between lecture and AL as appropriate for my instructional goals. Likewise, I want both STEM and the humanities to be valued highly.

Yet, as folks in the seminar noted, many faculty do still see AL as coddling lazy students. This makes it hard to go on the teaching job market and present yourself as a proponent of AL without raising the hackles of traditional-lecture fans. Will you lose a potential job because you think AL is a useful part of your teaching toolkit?

On the other hand, the gap between AL and traditional classes isn’t always that big. For example, some folks pointed out that in English lit classes, traditional teaching is very much a case of AL. You read a book or chapter outside of class, then come and discuss it together—obviously. How else would you even teach literature?

Derek Bruff has a blog post about the many defense-of-lecture articles out there, and how very few are actually defending “continuous exposition by the teacher” (as Freeman et al. defined traditional lecture). Most of these authors use lectures interspersed with questions, discussion, and projects. This is no different from teaching via AL interspersed with some periods of talking for exposition, motivation, or summary. We should be careful not to talk past one another. (Worthen might be one of the few who really do seem to be defending straight-up talking, with no questions or activities or breaks.)

2 responses to “Seminar on Active Learning pedagogy

  1. My sense is that one of the blocking obstacles in the path of introducing new teaching methodologies is inertia. If lecture notes already exist for a course, the easiest approach is to simply deliver the lecture from the notes.
    On the other hand, spending time on reworking the course for Active Learning seems like it will take a great deal of effort and time. Is that true? Could you provide a ballpark figure for the amount of work required to modify a course for Active Learning? Does it pass a cost-benefit analysis?

    • Well, you could spend literally 0 prep time and 6 minutes of class time on “the pause procedure”—just take a 2-minute break every 15 minutes or so. It gives students a chance to assimilate the info, helps ESL students catch up to what you’ve been saying, etc. At least in that one study, there were substantive improvements on test scores, and also on the number of correct things students listed in free-recall after each class. I think that passes a cost-benefit analysis.

      Past that, it really depends on how far down the rabbit hole you want to go. I’ll try to address prep time more when I post notes from the 2nd session (soon).

      Update: Dan Meyer has great advice that can be done with very little prep time, but (I believe) substantial benefit:
      “…before any explanation, create conditions that prepare students to learn from that explanation. These for example.
      Some examples in a stats context: Before you tell students that standard errors decrease with greater sample size, or before you plot this relationship, ask them to conjecture about and/or sketch the relationship. Before you tell students that lasso fits with higher lambda values give sparser models, ask them to conjecture about lambda’s effect.
      You could spend a lot of time on these, making them full-blown class activities. But even if you *just* ask the question and give the students 60 seconds to think about it, then continue with your lecture—that’s already a valuable improvement right there.