Difference between revisions of "How to Run Online Talks"
Line 7: | Line 7: | ||
---- | ---- | ||
"Since I'm an economist, I'll dispense with the sometime-hypocritical, sometime-sincere, but always nauseating flattery of saying what a great paper this was and how utterly brilliant an author. I'll get right down what's good in it and what's bad, especially what's bad, since that's where the author needs help." | "Since I'm an economist, I'll dispense with the sometime-hypocritical, sometime-sincere, but always nauseating flattery of saying what a great paper this was and how utterly brilliant an author. I'll get right down what's good in it and what's bad, especially what's bad, since that's where the author needs help." | ||
+ | |||
+ | I'm coming back to the idea of revising my <a href="http://www.rasmusen.org/GI/reader/writing.pdf" rel="noopener noreferrer" target="_blank">Aphorisms on Writing, Speaking, and Listening</a>. In particular, though good writing never changes, technology does. I need to talk more about how to use software in writing articles. I need to talk about programming style and posting data and code on the Web. I need to talk about online seminars. Professor Judy Chevalier at Yale was<a href="https://twitter.com/judy_chevalier/status/1260646267346014208" rel="noopener noreferrer" target="_blank"> asking for advice on that on Twitter.</a> Here are some ideas, particularly in light of Porfoessor Brunnemeir's Jeremy Stein workshop Monday. I'll keep adding to this post over time. One thing I need to add are my notes from the Toronto-Yale Aug. 28, 2020 talk my co-author Mark Ramseyer gave on our ostracism paper with attendance of about 900 people; I took a lot of notes on the mechanics. | ||
+ | |||
+ | A fundamental principle is to make the look-and-feel as close to that of a real seminar as possible. That means seeing people's faces, whether they're talking or not, hearing background noises such as chairs shuffling, seeing whose hand is up, allowing interruptions where appropriate, and having a pile of papers for people to grab as they come in. This is what e-books ought to do, too, and it took forever for them to do this--- in fact, they still fail. | ||
+ | |||
+ | There are two styles I'm seeing. One is the "webinar", where only the moderator and speaker appear and can communicate with the audience. This is a really bad format, reducing the seminar to pretty much the level of a You-Tube video, but less organized and polished. So it actually is a lot worse than a You-tube video. The webinar is favored by organizers who like tight centralized control and think of the audience as being like undergraduates, and below the level of children, who if not heard, at least are to be seen. | ||
+ | |||
+ | The other style is the Zoom session proper, like the zoom classes I've taught, where as many people as possible have video and audio, and there is expectation of true interaction. | ||
+ | |||
+ | There actually is a good case for having a number of styles. I writing this, though, I see we must talk about presentations in general, and their various purposes, before getting to the special features of online presentations. | ||
+ | |||
+ | . | ||
+ | Style A: The Lecture. If the point is for the audience to just listen to what someone has to say, then the lecture should be pre-recorded and posted as a web link. There is no point to it being live, and good reason for it not to be. It should be recorded in 15-minute chunks as distinct, titled, sections. It should be recorded more than once, and only the best "take" kept. Academics should try to live up to the standards of you-tube cooking demonstrations, which are much more polished, intelligent, lively, and professional than our video lectures, many of which are quite pitiful. <a href="https://twitter.com/akhivae/status/1263486392836947970" rel="noopener noreferrer" target="_blank"> As someone said </a>about college plans for Fall 2020," What do you mean you don't want to spend $5000 a semester watching really bad versions of Khan Academy lectures over Zoom?" | ||
+ | |||
+ | . | ||
+ | Style B: The seminar. If the point is to have an economics-style "seminar" or "workshop", the main objective is to get interaction between speaker and audience. It is best, in fact, not to think in terms of "speaker", but of "main speaker". | ||
+ | |||
+ | When I visited Chicago in 1989, I found that at the law-and-economics workshop at the law school, notorious for its liveliness, the "main speaker" ordinarily spoke fewer words than either David Friedman or John Lott, two regular attendees. That was good, though it is important with those two people to tell them to stop talking if they talk too much--- which, they would willingly do--- if told, though sometimes not otherwise. When I was at UCLA, it would happen in Industrial organization seminars by PhD students that the main speaker stood silent and confused as Harold Demsetz and Ben Klein argued over what the paper was about. That was good, because often PhD students don't know what their own papers are about, and don't have much interesting to say, whereas Demsetz and Klein were always worth listening to. | ||
+ | |||
+ | Our own Indiana University Business Economics and Public Policy workshop is in this "Chicago style" too, though less extreme. The point is to learn things and improve the paper being presented, not to give the speaker a chance to show off (unless it's a job talk). If learning and improving are the objectives, then the audience questions and comments are far more important than the speaker's prepared talk. It isn't particularly important for the speaker to get through all his slides, or even to get to the end of his talk. When I was a student at MIT, Peter Temin told us that a talk has three objectives: 1. Teach something to the audience. 2. Learn something from the audience to make it a better paper for you and the world. 3. Show off how smart you are. He said that (3) is appropriate and important for job talks, but only for job talks. | ||
+ | |||
+ | Questions are not at the end, but sprinkled throughout. I will write on that later, since it is so important and so often done badly. Law schools, the Ostrom Workshop, and Liberty Fund conferences are very bad in this dimension. The queue is an abomination. | ||
+ | |||
+ | . | ||
+ | Style C. The Lecture with Questions at the End. This is like conference presentations and titled Lectures, e.g. the Clarendon Lectures. It combines a straight lecture with questions from the audience at the end. This is what the Webinar tries to do, but fails at every time I've been to one. | ||
+ | |||
+ | . | ||
+ | Style D. The Large Seminar. This perhaps is unique to online. It is appropriate when a scholar is presenting to other scholars, but there are going to be lots of non-scholars present. Or even perhaps just when there are 100 scholars present. The difference from style B is that there are 200 in the audience rather than 20, and many members of the audience don't know how to behave in a seminar and are shameless about it (unlike grad students, who must learn how to behave but are scared stiff about their ignorance), or can behave well but are so ignorant that their questions and comments are a waste of time. | ||
+ | |||
+ | On to my unorganized list of tips. | ||
+ | |||
+ | 1. Have a list of who's attending--- I was curious. Allow opt-out from the list. | ||
+ | |||
+ | 2. If you can show faces of audience, do that (e.g. , if fewer thn 30 people). Do that even if you mute them. | ||
+ | |||
+ | 3. Maybe have the moderator be like Snerdly on Rush Limbaugh and choose questions and turn the audience members on to read them, after coaching for conciseness. In genral, try for the live feel. If it's a small seminar, though, just let everybody chip in when they want to, as in regular seminars. | ||
+ | |||
+ | 4. If there is a large queue of question, use the Zoom thumbs up feature and tell audience in advance to thumbs up questions they like. | ||
+ | |||
+ | 5. Explain the technology and rules to audience carefully at the start-- e.g., that thumbs up feature for seeing what the audiecne is interestd in. | ||
+ | |||
+ | 6. At Princeton, 90% of questions were boring and mediocre, so you definintely need a moderator to filter a large seminar, with nonscholars present. | ||
+ | |||
+ | |||
+ | 7. Encourage audience to use the chat box to help with citations, pinning down facts (e.g. population of France, if that comes up and the speakers are uncertain). | ||
+ | |||
+ | 8. Have a grad student or colleague help the main moderator keep track of chat and questions and technical difficulties, and for backup if the main moderator's computer goes dead. | ||
+ | |||
+ | 9. Post the slides on the web, and the paper. Put the addresses in the chat box so the audience can get them in the middle of the seminar. | ||
+ | |||
+ | 10. Encourage people to raise their hands to get the speaker or moderator's attention, and to wave their hands, because the movement is easier to see. | ||
+ | |||
+ | 11. Allow the audience to share. | ||
+ | |||
+ | 12. Be prepared for zoombombing, but use ex post rather than ex ante solutions. Don't require registration and a password and other thigns that discourage participation and shut down discussion. But think in advance of all the ways an evildoer might try to sabotage you, adn think of ways to shut them off. For example, know how to turn off people's video and sound, how to expel them and prevent reappearance from the same computer, and how to stop them from sharing their screen; and know how to do these things instantly, not after 10 seconds of fumbling. | ||
+ | |||
+ | 13. Open the meeting 15 minutes early, so early arrivers know it's going to happen. Use a Zoom waiting room, or just turn on the organizer's video even if he isn't there yet and it just shows an empty chair. Use Screen Share to say what's going to happen next, give links to the paper and slides, set out the ground rules, and suchlike. Some pleasant music--- Vivaldi perhaps--- would help. | ||
+ | |||
+ | 14a. In your email and webpage announcements, give the zoom link prominently (not hidden on a page of useless garbage about what to do if you're stuck in a jungle with only a pocketknife and a cellphone), together with the other relevant info: Time, Geographic Place, speaker, and title. | ||
+ | |||
+ | 14b. Also: give the Time for different time zones. In particular, list the starting time for New York, London, Los Angeles, and Peking. Maybe Melbourne, Berlin, Jerusalem, and Delhi too, but there's always a tradeoff when adding info. The main thing is to have at least one US time and at least one European time. | ||
+ | |||
+ | |||
+ | Do not add security to Zoom meeting unless you really ened it. | ||
+ | |||
+ | webinar advice--- do sen it out this weekend. | ||
+ | 1. send out code emails over and over. Espcially, right before. In the 5 minutes here you look up which physical room the seminar is in. We pasted the visitor schedule to the secretary’s desk for in-person seminars. | ||
+ | |||
+ | 2. Post a link to the ground rules where participants can see it. Zoom doesn’t allow that, alas. Have a webpage for each sminar where stuff like that can go,a dn post it on each speaker’s screen at the bottom. Or, tell audience how they can set it up on their desktop, with one window for browser, one for zoom video, one for zoom chat, no screen maxing that covers up all but one screen. Have a screenshot for them of a good auidence member. | ||
+ | |||
+ | 3. I need to send my stuff to the Zoom people. Easy improvements to their software. | ||
+ | |||
+ | 4. Tell peopel to get their faces close enough tot he webcam so their face dominates teh video, especially if the screen is going to have 10 people showing at once so each person is small. | ||
+ | |||
+ | <a href="https://www.colorado.edu/business/finance-cloud" rel="noopener noreferrer" target="_blank">A Virtual Finance Conference</a> | ||
+ | |||
+ | |||
+ | |||
+ | |||
+ | In zoom chat, a single sepaker must put in blank lines to seaprate his variousmessage suless onter speaker is in between. | ||
+ | |||
+ | . | ||
+ | <em>County Council. </em> | ||
+ | 1. If smoene wants to speak, require video. Open up chat. List attenders. Have a way questioners can test their audio and video--- a separate zoom session address with a clerk. Give the list of people who want to ask questions. | ||
+ | Required open meetings: should require people to be there visibly,tho they can wear masks (and use fake accounts). Allow chatting of people on the chat box, like in the hallway. You aren't using the chat box anyway. | ||
+ | |||
+ | |||
+ | An amazing number of peopel who think that their boss owes it to them to let them take care of their famileis while being paid to work. See <a href="https://twitter.com/drbarryrubin/status/1282324871960244224" rel="noopener noreferrer" target="_blank">this Twitter thread. </a> One commetn said: | ||
+ | <blockquote> | ||
+ | . | ||
+ | 1 - set meetings to run from five-past to five-to, never hour to hour. That way everyone gets a break | ||
+ | between. | ||
+ | . | ||
+ | 2 - camera on while you're talking, if possible. Not always possible for privacy & practicality. | ||
+ | . | ||
+ | 3 - really??? You've never had refreshments in a meeting? 1/n</blockquote> | ||
+ | |||
+ | |||
+ | |||
+ | |||
+ | <blockquote> | ||
+ | <a href="https://twitter.com/smsaideman/status/1303792901395288066" rel="noopener noreferrer" target="_blank">Steve Saideman</a> | ||
+ | @smsaideman | ||
+ | important | ||
+ | @APSAtweets | ||
+ | strategy: if your panel system isn't working, post a link to a zoom and have the entire panel and audience move with you. This is what the civ-mil panel did. And the massive audience (really!) followed along. | ||
+ | 4:30 PM · Sep 9, 2020·TweetDeck</blockquote> | ||
+ | |||
+ | |||
+ | |||
---- | ---- |
Revision as of 06:13, 28 September 2020
Notes
"Since I'm an economist, I'll dispense with the sometime-hypocritical, sometime-sincere, but always nauseating flattery of saying what a great paper this was and how utterly brilliant an author. I'll get right down what's good in it and what's bad, especially what's bad, since that's where the author needs help."
I'm coming back to the idea of revising my <a href="http://www.rasmusen.org/GI/reader/writing.pdf" rel="noopener noreferrer" target="_blank">Aphorisms on Writing, Speaking, and Listening</a>. In particular, though good writing never changes, technology does. I need to talk more about how to use software in writing articles. I need to talk about programming style and posting data and code on the Web. I need to talk about online seminars. Professor Judy Chevalier at Yale was<a href="https://twitter.com/judy_chevalier/status/1260646267346014208" rel="noopener noreferrer" target="_blank"> asking for advice on that on Twitter.</a> Here are some ideas, particularly in light of Porfoessor Brunnemeir's Jeremy Stein workshop Monday. I'll keep adding to this post over time. One thing I need to add are my notes from the Toronto-Yale Aug. 28, 2020 talk my co-author Mark Ramseyer gave on our ostracism paper with attendance of about 900 people; I took a lot of notes on the mechanics.
A fundamental principle is to make the look-and-feel as close to that of a real seminar as possible. That means seeing people's faces, whether they're talking or not, hearing background noises such as chairs shuffling, seeing whose hand is up, allowing interruptions where appropriate, and having a pile of papers for people to grab as they come in. This is what e-books ought to do, too, and it took forever for them to do this--- in fact, they still fail.
There are two styles I'm seeing. One is the "webinar", where only the moderator and speaker appear and can communicate with the audience. This is a really bad format, reducing the seminar to pretty much the level of a You-Tube video, but less organized and polished. So it actually is a lot worse than a You-tube video. The webinar is favored by organizers who like tight centralized control and think of the audience as being like undergraduates, and below the level of children, who if not heard, at least are to be seen.
The other style is the Zoom session proper, like the zoom classes I've taught, where as many people as possible have video and audio, and there is expectation of true interaction.
There actually is a good case for having a number of styles. I writing this, though, I see we must talk about presentations in general, and their various purposes, before getting to the special features of online presentations.
. Style A: The Lecture. If the point is for the audience to just listen to what someone has to say, then the lecture should be pre-recorded and posted as a web link. There is no point to it being live, and good reason for it not to be. It should be recorded in 15-minute chunks as distinct, titled, sections. It should be recorded more than once, and only the best "take" kept. Academics should try to live up to the standards of you-tube cooking demonstrations, which are much more polished, intelligent, lively, and professional than our video lectures, many of which are quite pitiful. <a href="https://twitter.com/akhivae/status/1263486392836947970" rel="noopener noreferrer" target="_blank"> As someone said </a>about college plans for Fall 2020," What do you mean you don't want to spend $5000 a semester watching really bad versions of Khan Academy lectures over Zoom?"
. Style B: The seminar. If the point is to have an economics-style "seminar" or "workshop", the main objective is to get interaction between speaker and audience. It is best, in fact, not to think in terms of "speaker", but of "main speaker".
When I visited Chicago in 1989, I found that at the law-and-economics workshop at the law school, notorious for its liveliness, the "main speaker" ordinarily spoke fewer words than either David Friedman or John Lott, two regular attendees. That was good, though it is important with those two people to tell them to stop talking if they talk too much--- which, they would willingly do--- if told, though sometimes not otherwise. When I was at UCLA, it would happen in Industrial organization seminars by PhD students that the main speaker stood silent and confused as Harold Demsetz and Ben Klein argued over what the paper was about. That was good, because often PhD students don't know what their own papers are about, and don't have much interesting to say, whereas Demsetz and Klein were always worth listening to.
Our own Indiana University Business Economics and Public Policy workshop is in this "Chicago style" too, though less extreme. The point is to learn things and improve the paper being presented, not to give the speaker a chance to show off (unless it's a job talk). If learning and improving are the objectives, then the audience questions and comments are far more important than the speaker's prepared talk. It isn't particularly important for the speaker to get through all his slides, or even to get to the end of his talk. When I was a student at MIT, Peter Temin told us that a talk has three objectives: 1. Teach something to the audience. 2. Learn something from the audience to make it a better paper for you and the world. 3. Show off how smart you are. He said that (3) is appropriate and important for job talks, but only for job talks. Questions are not at the end, but sprinkled throughout. I will write on that later, since it is so important and so often done badly. Law schools, the Ostrom Workshop, and Liberty Fund conferences are very bad in this dimension. The queue is an abomination.
. Style C. The Lecture with Questions at the End. This is like conference presentations and titled Lectures, e.g. the Clarendon Lectures. It combines a straight lecture with questions from the audience at the end. This is what the Webinar tries to do, but fails at every time I've been to one.
. Style D. The Large Seminar. This perhaps is unique to online. It is appropriate when a scholar is presenting to other scholars, but there are going to be lots of non-scholars present. Or even perhaps just when there are 100 scholars present. The difference from style B is that there are 200 in the audience rather than 20, and many members of the audience don't know how to behave in a seminar and are shameless about it (unlike grad students, who must learn how to behave but are scared stiff about their ignorance), or can behave well but are so ignorant that their questions and comments are a waste of time.
On to my unorganized list of tips.
1. Have a list of who's attending--- I was curious. Allow opt-out from the list.
2. If you can show faces of audience, do that (e.g. , if fewer thn 30 people). Do that even if you mute them.
3. Maybe have the moderator be like Snerdly on Rush Limbaugh and choose questions and turn the audience members on to read them, after coaching for conciseness. In genral, try for the live feel. If it's a small seminar, though, just let everybody chip in when they want to, as in regular seminars.
4. If there is a large queue of question, use the Zoom thumbs up feature and tell audience in advance to thumbs up questions they like.
5. Explain the technology and rules to audience carefully at the start-- e.g., that thumbs up feature for seeing what the audiecne is interestd in.
6. At Princeton, 90% of questions were boring and mediocre, so you definintely need a moderator to filter a large seminar, with nonscholars present.
7. Encourage audience to use the chat box to help with citations, pinning down facts (e.g. population of France, if that comes up and the speakers are uncertain).
8. Have a grad student or colleague help the main moderator keep track of chat and questions and technical difficulties, and for backup if the main moderator's computer goes dead.
9. Post the slides on the web, and the paper. Put the addresses in the chat box so the audience can get them in the middle of the seminar.
10. Encourage people to raise their hands to get the speaker or moderator's attention, and to wave their hands, because the movement is easier to see.
11. Allow the audience to share.
12. Be prepared for zoombombing, but use ex post rather than ex ante solutions. Don't require registration and a password and other thigns that discourage participation and shut down discussion. But think in advance of all the ways an evildoer might try to sabotage you, adn think of ways to shut them off. For example, know how to turn off people's video and sound, how to expel them and prevent reappearance from the same computer, and how to stop them from sharing their screen; and know how to do these things instantly, not after 10 seconds of fumbling.
13. Open the meeting 15 minutes early, so early arrivers know it's going to happen. Use a Zoom waiting room, or just turn on the organizer's video even if he isn't there yet and it just shows an empty chair. Use Screen Share to say what's going to happen next, give links to the paper and slides, set out the ground rules, and suchlike. Some pleasant music--- Vivaldi perhaps--- would help.
14a. In your email and webpage announcements, give the zoom link prominently (not hidden on a page of useless garbage about what to do if you're stuck in a jungle with only a pocketknife and a cellphone), together with the other relevant info: Time, Geographic Place, speaker, and title.
14b. Also: give the Time for different time zones. In particular, list the starting time for New York, London, Los Angeles, and Peking. Maybe Melbourne, Berlin, Jerusalem, and Delhi too, but there's always a tradeoff when adding info. The main thing is to have at least one US time and at least one European time.
Do not add security to Zoom meeting unless you really ened it.
webinar advice--- do sen it out this weekend.
1. send out code emails over and over. Espcially, right before. In the 5 minutes here you look up which physical room the seminar is in. We pasted the visitor schedule to the secretary’s desk for in-person seminars.
2. Post a link to the ground rules where participants can see it. Zoom doesn’t allow that, alas. Have a webpage for each sminar where stuff like that can go,a dn post it on each speaker’s screen at the bottom. Or, tell audience how they can set it up on their desktop, with one window for browser, one for zoom video, one for zoom chat, no screen maxing that covers up all but one screen. Have a screenshot for them of a good auidence member.
3. I need to send my stuff to the Zoom people. Easy improvements to their software.
4. Tell peopel to get their faces close enough tot he webcam so their face dominates teh video, especially if the screen is going to have 10 people showing at once so each person is small.
<a href="https://www.colorado.edu/business/finance-cloud" rel="noopener noreferrer" target="_blank">A Virtual Finance Conference</a>
In zoom chat, a single sepaker must put in blank lines to seaprate his variousmessage suless onter speaker is in between.
. County Council.
1. If smoene wants to speak, require video. Open up chat. List attenders. Have a way questioners can test their audio and video--- a separate zoom session address with a clerk. Give the list of people who want to ask questions. Required open meetings: should require people to be there visibly,tho they can wear masks (and use fake accounts). Allow chatting of people on the chat box, like in the hallway. You aren't using the chat box anyway.
An amazing number of peopel who think that their boss owes it to them to let them take care of their famileis while being paid to work. See <a href="https://twitter.com/drbarryrubin/status/1282324871960244224" rel="noopener noreferrer" target="_blank">this Twitter thread. </a> One commetn said:
. 1 - set meetings to run from five-past to five-to, never hour to hour. That way everyone gets a break between. . 2 - camera on while you're talking, if possible. Not always possible for privacy & practicality. .
3 - really??? You've never had refreshments in a meeting? 1/n
<a href="https://twitter.com/smsaideman/status/1303792901395288066" rel="noopener noreferrer" target="_blank">Steve Saideman</a> @smsaideman important @APSAtweets strategy: if your panel system isn't working, post a link to a zoom and have the entire panel and audience move with you. This is what the civ-mil panel did. And the massive audience (really!) followed along.
4:30 PM · Sep 9, 2020·TweetDeck
Zoom should show you in Gallery EXACTLY what peopel's creens look like, rather than everybody having a differen torder of people's pictuers. I want to know whether people are seing my photo right next to teh at of the person who is sepaking, os they're noticing me.
WHat is the best way to tell someoen that they need to unmute themselves? (photo of me with ears). A new software command? We cannot rely on the moderatior along--too busy.
You don't need to mute yourself while you're typing notes.