Thoughts on the Program Chair Experience

(Revised 16 February 1999)

Comments? E-mail to Harold Brooks


I was the chairman of the Program Committee for the recently held 19th Conference on Severe Local Storms of the American Meteorological Society. I wanted to write down some thoughts on the experience that might help future program committee chairs, particularly for AMS conferences.

Background

The Severe Local Storms (SLS) conferences are held every 2-3 years, usually around October or February. The previous conference was in February 1996 in San Francisco. At that conference, the chair of the STAC committee for SLS (Ed Szoke) approached me about chairing the program committee for the next conference. The location wasn't settled, but Minneapolis was a leading candidate city and the conference was slated for the fall of 1998 (eventually settled on suburban Minneapolis for 14-18 September 1998). After discussing my thoughts on the what I wanted to see done with a conference, we agreed that I would select a program committee and begin preliminary work on planning the conference, while letting the STAC committee know what we were up to. In late summer of 1997, we wrote a Call for Papers, using previous Calls as models. The deadline for abstract submission was 16 March 1998 and we had to inform authors early enough to get their papers in to AMS headquarters by 15 June 1998.

Interaction with Local Arrangements

The local arrangements committee does as little or as much as AMS headquarters' staff or the program committee ask them to do. At minimum, they typically provide students to help with lights and AV during the oral sessions. Other options include what are typically called "spouses'" activities or tours of locations in the areas. In our case, our local arrangements committee helped find the hotel for the conference. We gave them some guidelines--primarily, we wanted a hotel with options for inexpensive food within easy walking distance of the hotel, with the idea that students or other people on a limited budget could eat well without having to have a car to get around. We then had a committee put together a restaurant list--maps of the hotel area and a ballpark estimate of meal cost. We made that list available for the attendees' benefit.

Creating the Program Committee

The two primary duties of the program committee are to "design" the conference (theme, layout, relationship of posters to oral presentations, etc.) and to dispose of the abstracts at the appropriate time. Much of this can be done electronically, but it's hard to see how to do the final layout, including the order of papers without physically being in the same room and being able to shuffle pieces of paper around for a day or two. As a result, it's probably a good idea to have as many members as possible from a small geographic area and that the program chair(s) is comfortable working with everyone in the group. Depending on the conference, different communities should be represented. For instance, for SLS, the large operational and non-university research communities are important. I sought out a forecaster (Steve Weiss), and a university researcher (Jim Moore) and non-university researcher (Jeff Trapp), both with strong ties to the operational community, to form the core of the committee. All had attended multiple AMS conferences and had ideas about what worked well and didn't work well. Later, we added two other members (Jeff Hovis, a forecaster, and Bill Conway, an applied researcher) with less conference experience, but whom we could count on to help, particularly in the abstract disposition process.

The chair has almost absolute autonomy in the selection of the committee. I had to run the names past the STAC committee, but wasn't questioned on any of them. The important thing to remember is that putting together a poor committee leads to one of two options: more work for the chair or a bad conference experience for the attendees, which will reflect badly on the chair's reputation. I had great confidence in my committee and they made the process just about as easy as possible. More than 18 months in advance, we were set to go.

Designing the Conference

Everything the program committee does prior to abstract disposition should be aimed at making the disposition process as painless for the committee as possible. This is a selfish view, but, even when everything goes well, putting together the schedule of things at the conference is not easy. The AMS will provide the committee with a complete list of dates for when things are due, based on the date of the conference. (For instance, papers were due ~3 months before the SLS conference and the abstracts ~3 months before that, with other things falling in place). The dates are a function of when other conferences are and how long it takes to turn things around at the printers, etc., so don't take our experience as universal.

(It is important for the committee to realize and to communicate to authors the effects of improvements in technology on the deadlines. Since the final papers can be turned around much more quickly than in the past, the paper deadline is now closer to the conference than in the past. As a result, the unwritten rule of the past that authors have a couple of weeks after the official deadline for papers before the "real" deadline may not be valid. Also, the growth of electronic abstract submission has shortened the time between submission of the abstract and the final papers. As a result, late abstracts are a nuisance and the committee will have little time to turn around the program.)

We chose to use the planned new Severe Local Storms monograph as the organizing mechanism of the conference. We had other choices, including the 50th Anniversary of the first tornado forecast, but none provided the breadth of subject matter we wanted to include. The severe storms community may be unique in meteorology in having a large core group that are interested in a wide range of topics of how storms are made and what they produce.

We made several other decisions on how to proceed:

  1. Invited talks from the monograph chapter authors
  2. Have a large number of oral presentations from first time (or other new-to-the-field) presenters
  3. No overlapping oral sessions
  4. Oral presentations of 15 minutes (12 for talk, 3 for questions) for submitted abstracts, 30 minutes for invited talks
  5. Panel discussion on topic of importance (eventually this was "Communicating Information on Weather Hazards to the Public")
  6. Individual sessions of no more than 90 minutes, with 30 minute breaks in the morning and afternoon, with 90 minutes for lunch
  7. Posters up for two days, with presenters "guarding" them for one session on one of the two days

(The final idea came from conferences I had attended sponsored by the Department of Energy. The idea allows for attendees to view the posters at their leisure and to catch presenters at times other than the official session to discuss the poster with them. It also allows the presenters to see the posters in their own sessions. Since the sessions were organized around common areas, it would seem natural that presenters would be most interested in the posters in the session in which they were scheduled to be in front of their poster and, as a result, unable to see those other posters. We decided to have the poster sessions at the end of the day, with food available as often as possible.

Failure #1: We didn't communicate with the AMS on the "two-day" poster idea as well as we should have. As a result, there weren't enough poster boards and may not have been enough room in the poster room for all posters to be up for two days. It's a system I think works extremely well and I hope somebody else tries it. We got food for the first night from a corporate partner (Kavouras), but needed to get more help to get food for the other nights. AMS has prices that the various food options cost, so prospective sponsors can be approached.)

Once the general schedule was decided, it was a case of "facing the awful arithmetic", as Lincoln said in a somewhat different context. Assuming that the conference ran from Monday morning until Friday noon, we had time for 18 oral sessions. With one set aside for the panel, that left a total of 17x6=102 15-minute "slots" for oral presentations. Nine chapter authors accepted the invitation to give talks and another was scheduled for the luncheon talk. With two slots per invited presentation, that left us with 84 places for oral presentations. Given that the last SLS conference had had almost 200 papers, it was clear that most of the papers would be posters. We received some criticism for stating in the call for papers that the majority of presentations would be posters, but it was a simple reflection of the anticipated number of submissions and our refusal to consider evening scientific sessions, concurrent sessions, or to have talks shorter than 15 minutes.

The poster/oral situation also reflected my personal view that, as a scientist, I get much more out of making a poster presentation about work in progress than out of making an oral presentation. Not as many people may "hear" me, but those that do are very interested in the topic and the level of interaction is greater. (Review talks, designed to convey the state of the knowledge of the community, are a different matter.) Since I always operate under the assumption that any conference publication is destined to end up as a formal publication, getting quality feedback is much more important than having a large number of pseudo-interested people hear me talk.

On the flip side, as a "senior" scientist with occasional bureaucratic duties, I'm concerned about identifying new people who will contribute to the science and who may be candidates for jobs that I have influence over. Seeing how those people handle themselves in an oral presentation is important in that process and being able to put a face with a new name is important.

We decided, in effect, as far as possible, to limit everyone to one time of standing in front of the audience. "Veterans" would make appearances as session chairs or in the invited talks or in a limited number of submitted presentations. Things didn't work completely that way, particularly when late substitutions for session chairs had to be made, but we decided to give lots of people a little exposure. Recognizing the male dominance of our discipline, we also tried to get as many women as possible included as session chairs and in the oral presentations, to provide positive reinforcement for the growing number of young women in the field.

Receiving and Disposing Abstracts

There were four primary channels by which abstracts were submitted

  1. Mail
  2. FAX
  3. E-mail
  4. Web submission

(Ann McCarthy, in charge of the web site at NSSL, created the submission form for us. Just before our deadline, the AMS opened up a site for abstract submission, but it was too late for our conference. I don't know how well it works, from a program chair's perspective.)

I took all of the abstracts and put them on a restricted access web site, so that all of the program committee could look at them as soon as possible. Scanning mailed and, particularly, FAXed abstracts and trying to correct scanning errors was unpleasant, so encouraging even greater use of electronic submission is a good idea.

The downside of electronic submission is that it allows authors to wait even longer to submit abstracts than they used to. I had hoped to do a preliminary program even before the abstract deadline. That didn't work out. We had 75 abstracts in hand by the end of Thursday, 12 March, just before the Monday, 16 March deadline. We got ~50 on Friday and then ~100 on Monday. There were another ~10 that people had given me titles on and had said they'd have them in a day or two late, mostly because of bureaucratic snafus. Finally, about 5 came in over the next week that were "truly" late and then three more that arrived over the following months that did not even get into the scheduling process but that we allowed into poster sessions.

We scheduled two days (25-26 March) to put the program together (after the 50th Anniversary of the First Tornado Forecast Symposium held in Norman that involved three of the committee heavily) and had our out of town committee member come into town. Before everyone got together, I made a first cut into general subject areas, according to the subjects from the Monograph. With the committee, we started putting together "sessions" of four (with invited talk), or six, papers. We then tried to organize the sessions into places in the various days. We decided to put what we viewed as the "core" subject matter of greatest interest from Monday afternoon through Thursday morning, with special sessions on societal impact and the panel discussion focusing attention on Wednesday. The previously mentioned interests of the severe storms community made that attractive. The conference luncheon was scheduled for Tuesday. Since the subject was numerical modelling, it was designed as a lead-in to two sessions of numerical modelling talks on Tuesday afternoon. (Later logistical considerations moved the luncheon to Wednesday.)

We had the largest number of talks in three areas--tornadoes, numerical modelling of storms, and forecasting techniques. [Fourth was convective high winds, with a large increase over past conferences, perhas as a result of operational deployment of the WSR-88D.] Each of the first three areas got two sessions. We put the first two on Tuesday and the second on Thursday morning. With Wednesday having the special session on societal impacts and the panel discussion, we put our other "hot" topic, convective winds, on Wednesday afternoon and a special session on Wednesday morning on flash floods. There were four abstracts submitted on the Fort Collins flash flood that made a nice, compact session in conjunction with the invited talk.

During the process of settling on an initial line-up for oral presentations, we created some small groups for possible poster sessions (some, such as "tornadoes around the world" were originally intended as oral sessions). We used them as the basis for the poster organization and then put the other abstracts into poster groups, with the goal of dividing them into four approximately equal groups for poster sessions on each of the first four days. Within each "day", we had 4-7 mini-sessions that we intended to to put together within the poster room to focus each area. Our four days could be summarized as "tornado", "numerical modelling", "forecasting techniques & radar-related papers", and "other aspects of severe convection." Since tornadoes and numerical modelling oral papers were on Tuesday, we had to figure out which one would be on Monday and which on Tuesday. We decided to go with tornadoes on Monday afternoon as a lead-in to the Tuesday morning orals. That put numerical modelling on Tuesday. The "other aspects of severe convection" was put on Wednesday, in large part because of the big number of high wind papers and we wanted to put that close to the afternoon orals.

With all of the abstracts laid out on a large table in sessions, we then put them into order for presentation. For posters, we tried to make sure that co-authors on multiple abstracts would have posters next to each other, even across session boundaries. For orals, we were conscious of trying to put appropriate papers next to each other and to try to be aware of outstanding presenters that, as one speaker put it, are "tough acts to follow" by not putting first time presenters right after them. By noon on the second day, we had the conference laid out. Each committee member then went through the posters and found abstracts that they would like to see presented as orals. We then discussed as a group the merits of presentation method for each of those abstracts and with abstracts that would have to be moved from oral to poster. Some of the posters were moved into oral sessions. This process also allowed us to identify a set of several papers that would be candidates for moving into oral sessions on an emergency basis. This turned out to be exceptionally useful when the Northwest Airlines strike led to us losing several oral presentations when authors couldn't attend. In the end, we tried to give authors their requested presentation mode.

After the Program Was Put Together

While putting the program schedule together was not an easy task, the clerical work in the few weeks that followed was tedious, at best. I did not use any of the laboratory's secretarial staff and, as a result, I had three tasks to do:

  1. Inform authors of the disposition of their abstracts
  2. Make a first/corresponding author list (for the AMS mailings)
  3. Type up the program

I accomplished #1 by e-mailing everyone the time of their presentation and their paper number. The second and third I could have made easier for myself with some thought beforehand. (I honestly had not discussed it with the AMS staff and didn't realize the jobs were my responsibility.) If I would have had the exact format ahead of time, I could have made entries for each paper and then just cut-and-pasted them into order once paper numbers had been decided. As it was, the form I had chosen for typing up author names and titles was sufficiently different from the official form that effectively I had to repeat the process I had already gone through. Make sure you get the format ahead of time to save a lot of hassle.

Once authors were informed, it was a matter of listening to needs for changes (people who had to withdraw papers or change authors, etc.), most of which needed to be forwarded to AMS headquarters. We solicited volunteers for the cover ($3000 per side or $5000 for both front and back) and contacted session chairs. A couple of months before the conference, I wrote up a little foreword for the preprint volume. There were only a few details to take care of in the last three months, so it was an easy time.

At the Conference

At the conference, the biggest thing I did was to prepare contingency plans, in case people cancelled or didn't show up. I told several people they might present their poster as an oral presentation and we ended up using all of those. I asked session chairs to write up a brief summary of their sessions, so that I could write a report on the conference for BAMS (that I have so far been derelict in composing). I admit my mind was more concerned with bureaucratic details than science during the conference, so I was happy I had not committed myself to a lot of presentation material. I did chair a panel discussion that went quite well in terms of involving the panel members and not having any of them say anything that was outrageous, although they certainly had comments that created discussion, as I had hoped.

The conference was a relatively easy week, particularly compared to putting the conference together. We had no major disasters and got through rather painlessly. One final lesson that I learned involved the poster room. We had fairly short poster sessions (75 minutes) with the idea that people could stand around and chat afterwards but no one would feel pressured to stay for a long time with no one talking to them. Thus, our poster room was only officially reserved until 6 PM. The final night of posters, the hotel scheduled something in our room for 6:30 PM, since according to their booking schedule, we were done at 6 PM. This created an incredible tear-down scene that would have been a lot funnier to a lot more people if they didn't feel like they were losing out on something. Make sure the rooms are reserved for the entire day, even after evening sessions are done.

Our AMS staff liaisons (Debbie Fleming, registration and bureaucratic matters, and Monica Tolson, handling AV stuff) were wonderful. Debbie kept me informed of who had actually arrived and who had indicated they were ready to give oral presentations. Monica checked all of the AV equipment out and we had very few problems. When things did go wrong, she was on top of it within seconds, so that we didn't fall behind schedule. Unfortunately, it was Debbie's last conference and one of Monica's last as AMS employees. (Debbie was flying straight to Florida from Minneapolis to meet her husband moving there. Monica is now free-lancing and doing some things with AMS conferences. If you can get her she's quite good.)

After the Conference

After the conference, I tried to put most of this together quickly, so that it was fresh in my mind. If more items come up, I'll include them. I'd like to think that this is a useful exercise and that it may help people prepare for the inevitable pitfalls.

I still have a lingering interaction with the next conference, attempting to attract a new program chair. Beyond that, I'm ready to take time off from that sort of thing, but I'd be glad to talk to anyone about the process.

If you've agreed to do this, good luck! If you haven't, don't complain about how somebody runs "their" conference. Sometimes, things aren't always in their control and other times, as Burns said it in The Mouse, "...The best laid schemes o' Mice an' Men, Gang aft agley.... "