This past Tuesday I guest-taught in my colleague Amanda’s masters-level course on political rhetoric, taught at Asheville’s branch campus of Western Carolina University, where Amanda is a faculty member. In planning for this guest appearance, I struggled for a bit to come up with a topic that would have something of a rhetorical component to it but still draw on ideas I’ve been discussing in my own voting theory course. After some thought, I developed an exercise that would (ideally) help to illustrate the rhetorical impact of numerical data specifically, poll results, on electoral behavior.
I gave each of Amanda’s ten students (and Amanda herself, good sport that she was and always is!) a “voter profile” that consisted of a brief statement of a view on the three candidates in the race, Charmander, Squirtle, and Bulbasaur. For instance, two persons were given the statement “Charmander would be ideal; I’d love to see her rule the gym. But I could live with Bulbasaur, I guess. On the other hand, Squirtle’s election would be the death of the Pokémon Republic!” and one was given “Bulbasaur is the only truly principled Pokémon running this year, but I guess if I had to I’d vote for Squirtle to avoid that horrible Charmander.” There were seven distinct statements in all.
I then asked each person to take 10 points and divide them among the candidates in a manner they felt concordant with the statement they’d been given. We could view such a vote as an instance of range voting, but I meant it merely as a means of getting a larger electorate, effectively allowing each person to act as ten voters rather than one.
After collecting the ballots and tallying the results, I asked the class to vote again, again in a manner concordant with their statements, but not before giving them the following polling data:
- Charmander: 40%
- Squirtle: 35%
- Bulbasaur: 10%
- Undecided: 15%
After tallying these ballots, I requested a third round of voting, after the class viewed new polling data:
- Charmander: 30%
- Squirtle: 30%
- Bulbasaur: 27%
- Undecided: 13%
Though I wasn’t sure how it would go, I suspected that the exposure to initial set of polling data indicating their candidate had a snowball’s chance would cause the Bulbasaur boosters to vote more strategically, rather than sincerely as they might have in the absence of polling data. A second set of polling data, this set showing that Bulbasaur was within reach of the two leading candidates, should cause the Bulbasaur boosters to revert to voting sincerely.
This is exactly what happened, with the following results:
- Charmander: 24 (21.8%)
- Squirtle: 26 (23.6%)
- Bulbasaur: 60 (54.5%)
- Charmander: 35 (31.8%)
- Squirtle: 42 (38.2%)
- Bulbasaur: 33 (30.0%)
- Charmander: 25 (22.7%)
- Squirtle: 28 (25.5%)
- Bulbasaur: 57 (51.8%)
Now, it’s dangerous to draw parallels between a low-stakes academic exercise involving Pokémons and a real-world election with real-world consequences, like the ongoing contentious election for POTUS, but it’s interesting to note that electoral support for a third-party candidate (Gary Johnson, a.k.a. Bulbasaur) with appeal to disgruntled voters in both major parties might depend so heavily on knowledge, or lack thereof, of polling data.
We talked a bit about this phenomenon, as well as the related impact of election returns data, an impact felt particularly strongly on the West Coast, where many folks don’t have a chance to get to the polls until after more easterly states’ races are already called on the basis of a representative sample of exit poll results. Given this effect, is the release of polling data an ethical practice, particularly when that release comes on or near Election Day?
This question gave rise to a host of ethical questions relating to poll results and polling processes and electoral behavior more generally: is one ethically obligated to vote, if one is legally permitted to do so? Is strategic voting an ethical practice? What about buying, selling, or trading one’s vote? It was a rich discussion.
I’m going to run the same exercise in my own class in just a few minutes. I’ll check in afterward with new data!