I studied patterns in laughter during stand-up comedy performances while at Queen Mary University of London. To my knowledge this is the first time laughter has been quantitatively studied across performances of the same material.
I used ELAN to annotate performance audio and Python to analyse and visualise the data.
Laughter bouts are significantly longer when more people are laughing, with more whole audience laughter at in the introductory section of the show.
While the amount of laughter per performance varied, laughter occurred frequently in very short bursts. In one show studied, laughter occurred on average every 4—5 seconds: laughter is peppered throughout rather than a response to regular punchlines with a long build-up.
Laughter bout density categorised by size of group laughing in a 40 minute performance. Solo (1 person), Babble (2-5 people), Audience (5+ group laughter). Audience Laughter is most frequent at the beginning and end of the performance.
Because of its frequency, the absence of laughter is notable. Long gaps between laughter occurred during the same material in different performances, suggesting the comedian has control when withholding laughter.
Some of these findings were presented in a short article at the 5th Laughter Workshop (2018).
Gaps between laughs in five live performances of the same material show the relative stability of long gaps between laughs at the beginning and halfway through the segment, particularly in the last four shows. Black dots indicate time since previous laugh ended, while the green dots show performer self-laughter.