A friend observed a pattern where contra dance events seem to be pairing older and younger callers. I looked over my notes for two-caller events in 2025 and saw [1]:
- Two older callers: 33 events
- One of each: 30 events
- Two younger callers: 4 events
Seems pretty clear evidence of pairing, no? But this actually turns out to be what you'd expect to see if organizers ignored age.
With 67 two-caller events there are 134 slots. Of these, 96 (72%) went to older callers and 38 (28%) went to younger ones. So there are four possibilities:
| caller 1 older | caller 1 younger | |
|---|---|---|
| caller 2 older | 72% * 72% | 28% * 72% |
| caller 2 younger | 72% * 28% | 28% * 28% |
This gives us:
- Two older callers: 34 events (72% * 72%), vs 33 observed
- One of each: 27 events (72% * 28% + 28% * 72%), vs 30 observed
- Two younger callers: 5 events (28% * 28%), vs 4 observed
While this is very slightly in the direction you'd expect if organizers preferred to match different-age callers, it's well within what you could get by chance. It looks to me like this is just "two moderately rare events both happening is very rare."
We can compare this to the situation with gender, where you consistently get male-female pairs more often than you'd expect by chance:
The biggest caveat, though, is that this is based on a count of just one year's bookings. If I had more time, I'd like to go back over all the past data and count, but I really don't.
[1] Age is continuous, so this bucketing is somewhat arbitrary. Since
most callers are either baby boomers or millennials, though, I do think
there are two meaningful groups. I also don't know how old almost
anyone actually is, so am just guessing from appearance.
Discuss