Question
What is the optimal length of a single lecture?
Answer
I forget the source, but I recall one study that found 22 minutes to be the optimal time at which point the majority of people start losing the narrative.
It's hard to measure because "lecture" is an arbitrary administrative unit that bears no relation to what I call "narrative time." You could have a single long lecture that lasted an entire day but was structured right, with a few breaks. Or you could have what are nominally many separate lectures, but feel like one long interminable drag (this is often true of full-day "strategy session" type meetings in companies).
It depends partly on the audience, but above about 20 adults, you usually get the same mix of learning styles/temperaments, so it ends up being much more strongly dependent on the lecturer and his/her style. When I was teaching regularly, I found through trial and error (I was teaching a TuThu class with 90 minute sessions) that the best way was to do 2 pieces of about 40 minutes each, each further broken down into about 25 minutes of me talking and 15 minutes of some sort of interactive exercise. The next best thing was to do a to do a 50-30 split of lecturing/workshopping.
The content matters a lot too. When you are presenting something that is long, but needs to be long due to its own internal structure (for example, deriving a major proof live), you can talk non-stop for 90 minutes and still hold people's attention. In fact the ability to follow along is even a test of aptitude for the material. I had an excellent professor (Jessy Grizzle at the University of Michigan) for my graduate nonlinear systems class, and I still fondly remember a marathon 2-lecture sequence he did, deriving the basic existence proof for solutions to nonlinear ODEs. I was absorbed all the way through, as was the rest of the class. It was all classic blackboard work, no powerpoint, no clever gimmicks. In a way he merely got out of the way of material that was so beautiful, it practically taught itself.
But if you are dealing with much more interpretative material, like a historical argument, a business plan, or a session on engineering design, you can't do that. You have to break things up more.
I am wandering a bit from "lecture" strictly speaking now, but it is also interesting to ponder business presentations.
My formula for how much time you can hold people's attention in such situations is this: the more money and power there is in the audience, waiting to be committed, the less time you have. The more money and power there is, already committed, the more time you have. At the two extremes, where you are either purely asking for more, or purely presenting what you've already done with what you were given, the limits are 10 minutes and 90 minutes respectively. So a first pitch to a VC or corporate VP is 10 minutes, but if they've already put money on you and you are reporting, they'll take as long as 90 minutes to listen to you.
So the formula is:
[math]T_{\max}=10\alpha+90(1-\alpha), \alpha=M_1/(M_1+M_2)[/math]
where [math]M_1, M_2[/math] represent the money you've taken and the money you're asking for, respectively. The logic of the formula is that people are willing to put in more time to protect money they've already invested. You could substitute some abstract measure function of money and other forms of power too.
Since people asking for money/influence are generally lower status than people giving it, this is the reverse of teaching situations. Nominally, students pay professors to teach, but professors have a higher status, because they control a much more vital currency, grades. You could apply the formula above to a class where students don't care about grades, and the reverse formula where they do (i.e., switch 10 and 90 around).
It's hard to measure because "lecture" is an arbitrary administrative unit that bears no relation to what I call "narrative time." You could have a single long lecture that lasted an entire day but was structured right, with a few breaks. Or you could have what are nominally many separate lectures, but feel like one long interminable drag (this is often true of full-day "strategy session" type meetings in companies).
It depends partly on the audience, but above about 20 adults, you usually get the same mix of learning styles/temperaments, so it ends up being much more strongly dependent on the lecturer and his/her style. When I was teaching regularly, I found through trial and error (I was teaching a TuThu class with 90 minute sessions) that the best way was to do 2 pieces of about 40 minutes each, each further broken down into about 25 minutes of me talking and 15 minutes of some sort of interactive exercise. The next best thing was to do a to do a 50-30 split of lecturing/workshopping.
The content matters a lot too. When you are presenting something that is long, but needs to be long due to its own internal structure (for example, deriving a major proof live), you can talk non-stop for 90 minutes and still hold people's attention. In fact the ability to follow along is even a test of aptitude for the material. I had an excellent professor (Jessy Grizzle at the University of Michigan) for my graduate nonlinear systems class, and I still fondly remember a marathon 2-lecture sequence he did, deriving the basic existence proof for solutions to nonlinear ODEs. I was absorbed all the way through, as was the rest of the class. It was all classic blackboard work, no powerpoint, no clever gimmicks. In a way he merely got out of the way of material that was so beautiful, it practically taught itself.
But if you are dealing with much more interpretative material, like a historical argument, a business plan, or a session on engineering design, you can't do that. You have to break things up more.
I am wandering a bit from "lecture" strictly speaking now, but it is also interesting to ponder business presentations.
My formula for how much time you can hold people's attention in such situations is this: the more money and power there is in the audience, waiting to be committed, the less time you have. The more money and power there is, already committed, the more time you have. At the two extremes, where you are either purely asking for more, or purely presenting what you've already done with what you were given, the limits are 10 minutes and 90 minutes respectively. So a first pitch to a VC or corporate VP is 10 minutes, but if they've already put money on you and you are reporting, they'll take as long as 90 minutes to listen to you.
So the formula is:
[math]T_{\max}=10\alpha+90(1-\alpha), \alpha=M_1/(M_1+M_2)[/math]
where [math]M_1, M_2[/math] represent the money you've taken and the money you're asking for, respectively. The logic of the formula is that people are willing to put in more time to protect money they've already invested. You could substitute some abstract measure function of money and other forms of power too.
Since people asking for money/influence are generally lower status than people giving it, this is the reverse of teaching situations. Nominally, students pay professors to teach, but professors have a higher status, because they control a much more vital currency, grades. You could apply the formula above to a class where students don't care about grades, and the reverse formula where they do (i.e., switch 10 and 90 around).