kitchen table math, the sequel: bring back recitation

Tuesday, April 12, 2011

bring back recitation

I was just reading a short article in Education Week about Bill Gates' "Gold Star teacher" plan when I came across a letter criticizing the concept.

First, here's Gates:
Bill Gates closed the National Governors Association's 2011 winter meeting last week by urging the governors to consider increasing the class sizes of the best teachers.

Under the Microsoft founder's model, a school's most effective teachers would be given an additional four or five students. Less effective teachers could then work with smaller classes and receive professional development.

A 2008 study supported by the Bill and Melinda Gates Foundation determined that 83 percent of teachers would support increasing their class sizes for additional compensation. (The foundation provides grant support to Editorial Projects in Education, which publishes Education Week.) In 2009, a Goldwater Institute report argued for tying teacher effectiveness to a higher pupil-teacher ratio and a higher salary.

The endorsement by Mr. Gates now could push the proposal further into the mainstream, given the level of support shown at the NGA meeting.

Gates to NGA: Tie Class Sizes to Teachers' Skills
Education Week
Published Online: March 8, 2011
Published in Print: March 9, 2011, as Gates to NGA: Tie Class Sizes to Teachers' Skill
And here's the reaction from a letter writer:
What makes a teacher of young learners effective is his or her ability to work with individuals in ways that are appropriate to their needs. During whole-group lessons, such teachers move around their classrooms, spotting those who are having difficulty and taking the time to give a little help and encouragement. Later, when planning future lessons, they include modifications for the range of abilities in their classrooms and figure out ways to have most students working on their own or with a partner, so they can meet with small groups.

It is only the least-competent teachers who stand in front of their classrooms and give the same instruction to all, blind to the boredom of those who already know the material, the confusion of those who aren’t ready for it, and the tuned-out state of the few who don’t care.

Although the notion of getting extra pay for taking on more students might have seemed attractive to most of the teachers responding to a survey funded by the Bill and Melinda Gates Foundation in 2008, the situation at that time was only hypothetical.

Today thousands of teachers all over the country have classes of 30 and up. I wager that neither Bill Gates nor the governors who agree with him could keep order in such classrooms, much less teach anybody anything.
Linking Pay and Class Size Hurts Teaching Quality
And here is Doug Lemov:
[Many] to most of the top-performing urban charter schools of which I'm aware buck the otherwise orthodox belief in heterogeneous classroom grouping and solve this problem by homogeneously grouping classes.

Teach Like a Champion
p. 256
With homogeneous grouping, the teacher is always teaching to the level of the entire class because the entire class is on the same level.

Also: whole-group instruction does not mean whole-group lecture. Whole-group instruction means  "Call and Response," "Pepper," "Cold Call," "Wait Time," "Everybody Writes," etc. In the well-taught homogeneously grouped classes Lemov describes, the situation is pretty close to 100% of students learning from the teacher 100% of the time because 100% of students are directly engaged with the teacher for 100% of the class. That's the goal.

How much time are students directly engaged with the teacher in a heterogeneous class?

Not much.

Say class time is 50 minutes and you've got 20 kids.
  • 12 minutes for the mini lesson
  • 2 or 3 minutes for transition-time (sit on the floor to observe mini-lesson; re-group for partner-work; sit on floor again for mini-lesson; etc.)
  • 35 minutes for individual time with teacher
That's maybe 3 minutes of direct instructional time with the teacher per each two-child pair and another 12 minutes of time directly engaged with the teacher during the mini-lesson if the mini-lesson happens to be pitched to the child's level.

If the mini-lesson is not pitched to the child's level, then 3 minutes max.

One of these days I'll have to write up my notes from the 5th grade writing workshop I observed. I think it's fair to say that the two boys I was sitting closest to learned nothing at all for the entire class period. Learned nothing and practiced nothing.

I don't know whether the other kids were engaged in productive partner work.


Catherine Johnson said...

Having watched a teacher float around a classroom engaging in two-on-one interactions, I find it slightly staggering that no one has actually stopped to ask how much time kids are spending working on their own these days.

Catherine Johnson said...

I still remember the syllable recitations we did in 2nd grade.

They were fun.

Catherine Johnson said...

During whole-group lessons, such teachers move around their classrooms, spotting those who are having difficulty and taking the time to give a little help and encouragement.

What happens with the kids who aren't 'having difficulty'?

Does the teacher talk to them?

Catherine Johnson said...

This is classroom teaching as Extra Help.

Allison said...

In the last few months, I visited over a dozen elementary schools. Mostly I visited kindergartens, but whenever possible, I visited the 1st, 2nd, 3rd and 4th grades as well.

Over and over I saw schools where "math class" was the same template: children doing activities from Everyday Math on their own in chaotic, loud classrooms where students didn't have individual desks but had to sit at group tables (sometimes putting up their books and folders to act as little cubicle walls) while they waited for a teacher or an aide to interact with them. Uniformly, I saw half a dozen kids doing nothing at all in those times; another half a dozen chatting or playing but obviously not doing anything, and a precious few trying to block out the stimulus. Some read cheap fiction books.

No one could have learned anything in such a room even before you find out that the task at hand is some bizarre manipulative task in Everyday Math that had no goal or explained purpose anyway.

The teacher didn't spend more time with those having trouble it seemed, either, because those having trouble weren't even bothering to do the activity.

ChemProf said...

I wonder this too, when students are working on their own especially in elementary grades. In my college class, I often will work a model, then have the students work a related but slightly different example. I generally count on my students to call me over to check their work or ask questions, but the ones who are really struggling can be identified easily -- they are the ones sitting back and waiting until we work through the example as a class because they don't know where to start! If that's what some college students do, I can't imagine this works well with kids. Plus, I'm not dealing with disciplinary issues on top of the task at hand.

Bonnie said...

Just a counterpoint - I am a university professor who has switched over to the model that you are complaining about. Why? because my students weren't learning anything from the old fashioned "lecture plus out of class assignment"!! Now, I teach computer science, which is really an engineering field. We want our students to learn how to solve problems and design solutions. Learning to solve problems is inherently a hands-on task. I always tell my students that computer science is not a spectator sport. In the old days, we would lecture, and assign a programming project. Two weeks later, the project would be due. Typically, about half the class would either hand nothing in, or would hand something in that was so rife with mistakes and misconceptions that there was simply no hope. Eventually, these students would drop, leading to the problem that the introductory computer science course had one of the highest drop/failure rates on campus.
So we moved to "closed labs" - that is, in class structured exercises. I could walk around the room, and see exactly where students were going wrong. If lots of students were having the same problem, I could stop and do another example for the whole class. If a couple of students were really lost on a topic such as a linear search, I could walk through it visually with them. It was so much better working this way because I knew right away where the problem areas were. I will never go back to a 100% lecture model.

The other point to make is that while I do support ability-based grouping, the idea that you are going to end up with 30 kids all at the exact same level on the same day is simply ridiculous. My son, who is pretty far ahead of his grade level, plays with other kids who are at a similar level. I can see that each of these kids has particular strengths and weaknesses. Some are better writers and some are more advanced in math. Some are advanced in math but have particular trouble with one kind of concept. There is no way you are going to get 30 students who are so homogenous that the teacher never needs to differentiate.

SteveH said...

"There is no way you are going to get 30 students who are so homogenous that the teacher never needs to differentiate."

This is a matter of degree. At what point does it become a real problem? When my son was in fifth grade, the teacher had to work extra for kids who didn't know their times table and other basics. She could do nothing for the kids who were prepared for more.

Our K-8 schools use a full-inclusion model. They assume that whatever they do for differentiation is good enough. The problem is that this approach makes it easy for teachers to blame the kids; "They will learn when they are ready." As you go up the grades, the range grows larger until, finally, most schools divide kids into different levels for math starting in 7th grade. The issue is not whether you divide kids into homogeneous groups, but when. With our full-inclusion schools, they only offer platitudes.

SteveH said...

"Typically, about half the class would either hand nothing in, or would hand something in that was so rife with mistakes and misconceptions that there was simply no hope."

You lower expectations in the hope of getting a little bit more learning? What about the kids who could handle a more rigorous class? So, classes have become a kind of "office hours", except that you don't necessarily get individual attention? The need to teach this way is a really bad sign.

Catherine Johnson said...

Hi Bonnie -

Just a quick point - the teachers in Teach Like a Champion don't lecture.

I don't know how to describe what they do exactly; it's intensely interactive.

LOTS of questioning by the teacher, using techniques such as (Cold Call) that require every student to be ready to answer each question.

"Class recitation" is a means of having every student in the class answer every question.

Not sure how long the teacher ever talks or demonstrates in these classrooms.

Allison said...


The fallacy of the excluded middle is when we go all-or-nothing. There's a world of difference between some lecture coupled with recitation section plus lab time plus extra time in lab and "100% lecture" plus unguided lab time.

But what you said smacks of thinking "MY students aren't learning anything from MY lectures; it must be my FORMAT!" instead of "MY students aren't learning anything from MY lectures; how can I improve MY content and delivery in this format? How will that fit the rest of the course design?"

SICP is the best designed course in computer science; the course, taught at dozens of universities, if not hundreds, is certainly not equally effective in all of them, but at the top two, MIT and UC Berkeley, the lectures are BRILLIANT. And they are a necessary piece of the whole. The "whole" is 3 hours of lecture a week, 2 hours of TA/recitation instruction, 2 more hours of lab time with TAs present, plus 5-15 hours a week of homework which includes time supposed to be spent READING the BOOK in addition to both problems requiring thought and problems requiring coding. Remove the lecture and the whole thing would fall apart--you simply can't introduce the concept of the lambda operator or metacircular evaluation in a lab setting. Likewise, you can't successfully learn the material without thinking, and for many kids, you can't really see what to do without a lot of trial and error. And even in the lectures, it's not 100% the teacher speaking. There's time spent showing the interpreter, time spent using clickers to get feedback from the hall, time spent taking questions, time spent doing examples on the board. There's a lot of value in working hard to create good lectures. Ignoring them hurts students.

Parker said...

"Just a quick point - the teachers in Teach Like a Champion don't lecture."

Good point. We often use the term "lecture" to mean any time the teacher is standing in front of the class talking. This can be a teacher simply talking with no interaction or a highly interactive session like Socratic questioning.

This often makes it confusing to discuss this issue because every one brings their own interpretation as to what "lecture" means.

Bonnie said...

MIT doesn't use SICP any more. And Hal Abelson has moved on to mobile phone apps - he is behind AppInventor - and the whole point of that is to get students engaged by doing hands on development.

You will notice that there are 2 hours of lab time built into that course. I guarantee you that 90% of the learning is happening in those 2 hours, and that the students are very often not even showing up to the lectures.

The question-answer format is nice in many disciplines, but computer science is not a question-answer field. Would you teach a studio art class via question-answer, or in fact, any kind of lecture? I doubt it.

Allison said...


I taught the course at Cal. Taught it to 200 registered students and my ave lecture attendance was over 150, usually 180. There were far more than 2 hours of lab time built in to the class, and you can't say "90% of the learning" is taken place in "those two hours" or even the 10 hours of homework a week. You simply have no evidence for that. The learning that takes place in doing the assignments wouldn't work without doing the book's homework, without attending TA office hours and recitation section, without hearing the lecture. It's a package deal.

Alternatively, I TA'd a typical programming course at UC Santa Cruz where TAs were just helping in lab. There was no place to teach the material in an overview, no place to correct massive misunderstandings, no way to query students well enough to understand why the had the misconceptions they had. So you could correct their code but they'd make the same mistake next time, and on the midterm, too.

I don't know what you mean by a "question-answer" format in any field. It's just a straw man, because no one hear posed that any field was simply "q-a". The defn of CS you use simply doesn't match up to the breadth of what CS means. It's a lot more than programming. Thinking isn't just "hands on".

Studio art classes are garbage unless they teach technique. That means you sit there and get told explicitly what you are supposed to do, what you are attempting to do, what the goal is. If you're just left there to do figure it out for yourself, it's worthless. Sadly, most studio art courses are that ineffective. Wasted time isn't the same as learning.

Bonnie said...

Here are a few comments on Allison's post
1. The division of the course into lecture, recitation, and lab is mainly done at large universities, so that hundreds of students can be herded through a course without overly taxing the faculty, whose main job is to get grants, not to teach. You won't see courses like this at small liberal arts colleges, even the elite ones. More typically, you will see small classes of at most 25 students, and one professor.
2. I am mainly focusing on programming courses, especially the first ones, because that is where the problem is. For at least 20 years, we have lived with 50% drop/failure rates in those first courses. It has become untenable, and as computer science enrollment shrank dramatically in the last 10 years, we have been pressured to DO SOMETHING, or else (and yes, computer science departments have been closed, especially in the last year or so). Really, a 50% drop/failure rate is ridiculous, and evidence that our traditional methods were not working.
3. I am not saying that we should get rid of all teacher-led explanation. We clearly need that. But those 90 minute, 200-student lectures are just not effective. I can remember back in my days as a student, taking CS1. I had Deitel as the professor, about as engaging a professor as you can imagine. He was a very lively lecturer, and he peppered the class with questions. But what were the majority of students doing in his class? Reading the newspaper. Seriously.
4. In class questioning, and "discussions" tend to lend themselves best to learning small pieces of information at a time. So, I can introduce a particular kind of loop, say, a linear search, and ask them questions - when will this stop? If we are looking for a particular value, and it occurs twice in the list, which one will be returned? How many cells, on average, will we visit when trying to find something? That is fine, and students get something out of it. However, the students who are in trouble almost never answer, because they are just too busy trying to process the idea, and many of my students are from cultures that do not encourage speaking in class, or English is not a first language. So huge swaths of the class are left out of the conversation. Worse yet, the students will all still fall apart when they have to write their OWN linear search as part of a larger program. They can't seem to make the leap from seeing something presented to actually doing it themselves. That is the fundamental problem in teaching students to program. And yes, we can just wave our hands and say, if you can't do it as out of class homework, then you deserve to be part of the 50% who don't make it. But if we can do things differently, and reach people from that 50%, I think we should do it. The survival of our departments may be at stake.
5. If you go to the computer science education conferences, you will see lots of people discussing active learning. The NSF is funding it. Take a look at this
and this (from MIT, which is also moving away from lecture courses)