A few items on the election to go through before we turn our attention to Town Meeting. First, I want to pass on a surprising finding about ballot position as it relates to success in the election. The conventional wisdom is that the higher your name appears on the ballot, the better you will do. Some towns (like Billerica) randomly assign position on the ballot (rather than alphabetically) to make sure Mr. Abbott does not have a ballot position advantage over Mr. Zucker, for example.
But as it turns out, ballot position has no affect on the outcome of the election at all.
I plotted the ranks of the Town Meeting Representative voting results for each precinct against the position of the ballot, and it looks like this:
If there was a strong correlation, all the dots would fall nearly on a steep, upward-slanting line with a R-squared value near 1.00. Instead, it looks like the dots were sprayed out of a shotgun and the R-squared number is less than 0.02.
So the idea that voters simply check off the first 21 names on the ballot simply isn’t true. Furthermore, by and large voters are “bullet voting” Town Meeting Representatives. Look at the number of votes per ballot, by precinct:
| Precinct | Ballots Cast | Registered Voters | Total Votes Cast | Candidates | Write-ins Elected | Seats | Turnout | Votes per Ballot |
| 1 | 240 | 2606 | 2846 | 24 | 23 | 9.2% | 11.9 | |
| 2 | 320 | 2368 | 3587 | 20 | 2 | 22 | 13.5% | 11.2 |
| 3 | 241 | 2213 | 2498 | 17 | 21 | 10.9% | 10.4 | |
| 4 | 193 | 1925 | 1469 | 14 | 23 | 10.0% | 7.6 | |
| 5 | 339 | 2471 | 3773 | 26 | 21 | 13.7% | 11.1 | |
| 6 | 216 | 1850 | 1907 | 18 | 23 | 11.7% | 8.8 | |
| 7 | 194 | 2214 | 1672 | 14 | 1 | 21 | 8.8% | 8.6 |
| 8 | 294 | 2552 | 3102 | 24 | 22 | 11.5% | 10.6 | |
| 9 | 196 | 2257 | 1738 | 16 | 22 | 8.7% | 8.9 | |
| 10 | 230 | 2395 | 1810 | 12 | 21 | 9.6% | 7.9 | |
| 11 | 227 | 2178 | 2120 | 18 | 21 | 10.4% | 9.3 | |
| Totals | 2690 | 25029 | 26522 | 203 | 3 | 240 | 10.7% | 9.9 |
The average voter is filling in the oval for about half of the number of candidates. This suggests that voters are putting some level of thought — even some rudimentary strategy — into who they are voting for in Town Meeting. This is great news.
I realize this might not be a deep level of thought — possibly people are voting for names they know, streets they know (each candidate’s address is listed on on the ballot), or even names that sound like ethnically like their own. But it’s better than no thought at all, and it may mean that voters have the capacity to send the best people to TM if they are given a real choice.
A few other takeaways from this chart: You can see the turnout per precinct. It’s no surprise that the precinct with the most TMR candidates (#5), had the highest turnout (13.7%) another with a TM race (#8) came in third (11.5%) in terms of turnout. On the other hand, the third precinct with a race (#1) drew well below the town-wide average (9.2%).
You might also assume that turnout neighborhoods in which people were running for contest higher office, such as selectman and moderator, would be better because those candidates would bring their neighbors out to the polls. But town-wide top vote-getters Gil Moreira (#3, 10.9%), Andrew Deslaurier (#11, 10.4%), and Michael Moore (#4, 10.0%), didn’t really do that. In fact Deslaurier came in distant third in his own TMR race. (Even I beat him, 142 to 126, albeit in two different precincts.)
So the question remains as to why turnout numbers are so depressingly poor in Billerica. Later this week I will compile turnout numbers for all the towns I have previously researched to see if we can find any other with a problem as severe.
A final note: Three people won by write-in, meaning that there will be 42 empty seats headed into the caucuses this Thursday.
