24 June 2008

Evolution 2008: Graduate student attention span does not decrease due to conference attendance

Scientific conferences are widely regarded as good opportunities for social networking, the exchange of ideas, and the development of collaborative relationships. However, these events often contain contradictory elements: programmed talks in which scientists present their latest work to their peers, and planned social events in which the same scientists consume alcohol with their peers until the wee hours. The latter is not, traditionally, conducive to the effectiveness of the former, and it is a point of great importance whether the two are compatible at all.

To test the effect of the social aspects of conference attendance on the scientific aspects of the same, I measured a surrogate for the attentiveness of a doctoral student attending the Evolution 2008 conference, held at the University of Minnesota, Minneapolis campus between 21 and 24 June, 2008. Statistical analysis shows that, although the subject's attentiveness varied significantly between days, it did not change systematically as the conference progressed.

The subject, a third-year doctoral student in the biology department of a state land-grant school, attended a mixture of concurrent sessions, which consisted of 15-minute presentations of submitted papers, and symposia, which consisted of 30-minute invited presentations, and took notes on each attended talk (total n = 62). The student was allowed cheap coffee during the day, and pre-paid beer of middling quality ("meso-brew" sensu Godsoe) during scheduled social activities and extracurricular networking, both ad libidum.

The length of notes taken on each talk was measured, in lines of handwritten text, as a proxy for the student's attention span. To compensate for the greater length of the symposium talks, the length of notes taken on these was divided by a factor of 1.5; although the total time of symposium talks was twice that of concurrent session talks, it is understood that the student's attention waned as a roughly linear effect of the length of individual talks, so that he took fewer notes on a single symposium talk than he might have on two consecutive concurrent session talks. A one-way ANOVA was used to test for an effect of conference day on attention span, and a regression analysis was used to see whether there was a trend of increasing or decreasing attention span over time.

Results are displayed in Figure 1; the length of notes for each talk attended are provided as box-and-whisker plots for each day, with the number of talks attended each day (sample size) listed over each box plot. One-way ANOVA found a significant effect of conference day on attention span (p = 0.041); however, no gross trend is visible in the data, and linear regression of attention span on conference day explained very little variance (R-squared = 0.0217). I therefore conclude that, despite forces to the contrary, the subject successfully retained his faculties for the duration of the conference. I am unable to speculate, however, whether conference attendance has been ultimately productive, though the subject has been heard to say it was "a good time."

Thanks to the conference organizing committee! It was all sorts of science-y fun.

4 comments:

  1. Can I make a request? Well, I'm going to anyway. Could you rewrite the scienctific analysis of your experience at the conference like you were talking to a friend of a non-science background?

    I'm glad you enjoyed yourself, and I don't know anyone who want to hang out with a bunch of drunken scientists.

    ReplyDelete
  2. Sorry, Krista - my (apparently unsuccessful) aim was actually to make fun of scientific writing. The really short, clear version is:

    I wondered if I was paying less attention to the talks as the week progressed, so I counted up the lines of notes I took at each talk and did a couple of statistical tests for a trend of fewer notes (less attention paid) later in the week. And I didn't find it. The box-and-whisker plot shows the variation (the largest and smallest values, and the spread between them) in the note-counts, grouped by each day of the conference - you can see they don't really go up or down as the week progressed.

    For what it's worth, I find that drunk scientists are generally pretty fun to hang out with. But I might be a biased observer.

    ReplyDelete
  3. Oh! I kind of thought that's where it was going, but I wasn't sure because you did actually complete a box and whisker plot. My group of nerdy friends obsess less with science and more with science fiction.

    So drunken scientists avoid talking like scientists, right? Wait a second, you actually enjoy figuring out the statistical trends of your own note taking?!

    ReplyDelete
  4. Yeah, if anything, we get more science-nerdy when drunk. I'm locally notorious for having tried to work out a standard deviation by hand after a few too many.

    ReplyDelete