Well, we got it mostly right.
By mostly, we are talking 75% accurate. Of the four seats up for grabs in the races for School Committee and the Board of Selectmen, the Avocado correctly predicted 3 out of the 4 winners. Last year, we were 100% accurate, having picked all the winners and the correct order of finish. In addition, we predicted the outcome of the override vote. This year, 64% of our picks this year finished in the order in which we called them. Throwing us off the most were the crazy results of the Selectman’s contest. We still called the first and last place winners correctly in that race. On the School Committee side, Mr. Thielman surprised us with a first-place finish over Ms. Starks, but from Flanagan on down, we called all the rest in the correct order. Multiple simulations produced widely different results in the Selectman’s race and we went with the results that were based on a much higher than predicted turnout and total number of votes cast than what we actually observed on Tuesday. This overestimate of turnout also caused us to overshoot the vote totals for each of the candidates. Below, we will provide more post-election analysis, but first a few more numbers…
Over the past three elections (including the override vote last year), the Avocado can still boast an 88% accuracy in its predictions. When it comes to the correct order of finish, after aggregating all of the races, the Avocado’s record is 75% accuracy. Still, when you examine the data more closely, despite the craziness in the Selectman’s race, there are many other things that we actually got right – or otherwise very close to right.
What’s more, last year, we came very close to predicting the actual vote totals for many of the candidates. This year, though our data were thrown off by overshooting the estimated voter turnout, we still came very close to predicting the actual vote totals for two of the candidates in the School Committee race (Mr. O’Brien and Mr. Jackson). Though we had picked Jeff Thielman to finish 2nd, we were actually within 228 votes of his actual total. We were within 2% of calling the correct percentage of the vote for Ms. Starks, even though we had picked her to be first and she actually came in 2nd. We were within 2% of calling the correct percentage of the vote for Mr. Flanagan.
In the Selectman’s contest, we correctly predicted that the 2nd place finisher would eke out a win. Though we got the candidates wrong, our margin was very close. Byrne, the 2nd place winner, finished just 99 votes ahead of 3rd place finisher Joe Curran. We predicted that the margin would be 74 votes between 2nd and 3rd place. That’s an error of just 25 votes. For the middle of the pack (2nd, 3rd, and 4th) our predictions were off by just 1%-2% of the actual percentages garnered by those candidates.
Last year, we were very close to predicting the actual turnout, which helped us get as accurate as we did then. The actual turnout this year was around 7,441. This was similar to last year’s turnout (7,466) and actually a hair smalller, but we had banked on a turnout of at least 7,700. Though we were off by just 4% in our estimate, that was more than enough to skew the data. Factoring in just 100 extra voters, let alone 300, would have been enough to throw everything off kilter.
Below, we take a closer look at the predicted versus actual results. In bold, we have highlighted the candidates we correctly picked to win. The candidates we picked in the right order of finish are in italics. We also bolded any numbers or percentages on which our predictions were close to the actual results for that candidate.
——————————————————————————————————————-
SELECTMEN
PREDICTED ACTUAL
Joseph Curro 3,943 23% Joseph Curro 3,417 27%
Bob Tosi 3,496 21% Steve Byrne 2,805 22% *
Steve Byrne 3,422 20% Joseph Curran 2,706 21%
Joseph Curran 3,351 20% Bob Tosi 2,447 19%
Maria Romano 2,773 16% Maria Romano 1,431 11%
* The predicted margin between the 2nd and 3rd place finisher was also very close to the actual vote (even though we picked the wrong candidates in those spots).
———————————————————————————————————————————————————-
SCHOOL COMMITTEE
PREDICTED ACTUAL
Cindy Starks: 4,496 29% Jeff Thielman 3,426 28%
Jeff Thielman: 3,654 23% Cindy Starks 3,284 27%
Jim Flanagan: 3,360 22% Jim Flanagan 2,433 20%
A. Matthew Pallett: 2,804 18% Matthew Pallett 1,402 12%
Andrew O’Brien: 901 6% Andrew O’Brien 1,049 9%
Ian Jackson: 356 2% Ian Jackson 446 6%
—————————————————————————————————————————————————
IMPRESSIONS:
So what do all these results mean?
The short answer is, we have absolutely no idea. We have followed with great interest some excellent discussions on the Arlington List about the election results. Many interesting hypotheses were formulated. We will break down our own impressions according to different aspects of this election that have piqued the most curiosity.
How Much Did Bullet Voting Impact the Race?
There has been much discussion over how much “bullet” voting may have influenced this election. Bullet voting refers to the practice employed by some voters whereupon they choose just one candidate in a race for which they have two or more votes to give. For example, in both the Selectman’s and School Committee races, voters could choose to vote for any two candidates among the five and six that were seeking seats on those boards, respectively. Voters who bullet vote would choose to vote for just one candidate instead of two. Speculating as to why a given voter would choose to cast a bullet vote is just that – speculation. Even if we could examine each bulleted ballot individually, that still would not tell us anything, other than how many people bullet voted. Some voters may only like or know one of the candidates. For others, they may simply be undecided on their other choice. Still others may think one of the candidates is a shoo-in and therefore does not need their vote. The truth is, we may never really know.
If bullet voting was rampant, our conjecture is that Mr. Tosi may not have received enough 2nd votes from both Mr. Curro and Mr. Curran voters. If that is the case, this may have helped both Mr. Byrne and Mr. Curran surpass Tosi. Mr. Byrne may have received just enough of Mr. Curro’s second votes along with enough of his own supporters to edge out Mr. Curran – but this is only a theory.
How Much Impact Did Precincts 8, 10, and 12 Have On The Election?
Often, much is made about the impact of Precincts 8, 10, and 12. This is thought by some to represent the center of gravity for the movers and shakers in town. In recent elections, the numbers that turn out in those precincts have been impressive, to say the least. The elevated numbers in these precincts, some argue, help maintain the “status quo” as they call it, and they assume that voters in these precincts tend to vote overwhelmingly for the incumbents or those who are otherwise favored by the kingmakers. When we look at this year’s numbers, we see that between 25%-26% of the votes for Joe Curro, Jeff Thielman, and Cindy Starks came from just those three precincts.
Some have argued in the past that those three precincts alone skew the results in favor of incumbents. However, we feel that argument lacks supporting evidence. While it is clear that a good 25% of the votes for the incumbents and Joe Curro came from those precincts, there is no evidence that those three precincts alone made the difference in the outcome. Consider two scenarios. If you were to completely eliminate those three precints, Thielman and Starks still end up winning the election handily, topping Flanagan by some 500-600 votes. If you played this game a different way, kept 8, 10, and 12 in the picture, but pared down Thielman and Starks’ votes such that they were equal to those of Flanagan in those precincts, once again Thielman and Starks still win the general election with ease (though by a slightly smaller margin of about 400-500 votes). We are not going to go back and compare this to other elections, as that is largely an apples to oranges proposition. Suffice to say, that precincts 8, 10, and 12 alone cannot elect or defeat anyone. The elevated numbers could just as easily be attributed to the fact that there are more registered voters or more voters who actually turn out in those precincts. Not to be outdone, though he came in third in the race and in those precincts, Flanagan managed to make some inroads, having spent a lot of time campaigning there. Flanagan’s highest vote total came in Precinct 12, suggesting again that the trends we see in 8, 10, and 12 may simply have more to do with turnout there than with any particular slant by the electorate towards one candidate or another.
The number of candidates in the School Committee race may have also contributed to Flanagan’s loss. It was relayed to the Avocado that Mr. O’Brien’s only purpose in running was to take votes away from Flanagan and Pallett. If true, Mr. O’Briens’ 1,000-plus votes may have served to dilute the maximum number of votes Flanagan could have obtained. Mr. Jackson’s votes may have had a small impact, but he registered less than half of Mr. O’Brien’s total, which leads us to believe that Mr. Jackson may not have been as much of a factor in the race at all. Interestingly, Mr. Pallett did not appear to benefit from too many of Mr. Flanagan’s second votes. This may be because Flanagan received some 2nd votes from Ms. Starks or Mr. Thielman or because some may have simply bullet-voted for Flanagan. Again, there is truly no way to know for certain what happened.
Another piece of evidence that works against the 8, 10, and 12 conspiracy theory is that Steve Byrne did not exactly impress in those precincts. Byrne, a darling of the Curro crowd, just barely edged out the other Joe (Curran) in 10 and 12. He was only a little better than Tosi in those precincts, too. Curro was due to score big no matter what. However, those three precincts really did Byrne no special favors. We did not think Byrne would benefit from Curro’s coattails all that much. The data suggest this could be true, but it is not conclusive. Our hypothesis is that Byrne had to get a fair amount of support from Curro voters, otherwise it is doubtful he would have won. Still, Curro trumped Byrne by well over 600 votes and for his part, Byrne just scraped by Curran with a 99 vote margin. If you simply eliminated the margin of victory Byrne claimed over Curran in Precinct 8 (let alone 10 and 12), then we would have had a Selectman Curran (by just 12 votes). Again, this is just further data that would refute the claim that 8, 10, and 12 are significantly slanted towards incumbents or quasi-incumbents.
Name Recognition
How much did name recognition benefit or hurt candidates? Again, much of this is purely speculative. Though, it is interesting to observe that of the top 2 finishers in both the School Committee race and the Selectman’s race, those names are fairly well known in town. Jeff Thielman has been on the School Committee for nine years and was running for a fourth term. Cindy Starks has served for three years and ran for a second term. As an aside, why is it that nobody ever gets Ms. Starks’ name right? It seems that in both print and broadcast media, everyone calls her Cindy Stark. Do they not know that her surname shares its name with a town in Maine?
Joe Curro has been an elected official for four years and he has run in three successful campaigns. Joe Curran was also an elected official for four years and ran in four races (with a .500 record; though one race was uncontested). With all that has been made over the two Joes and Joe with the “O” versus Joe without the “O”, and not to mention some of the more memorable moments when both Joes served on the School Committee, it is not difficult to conclude that both men brought serious star power to the race in terms of name recognition. Mr. Byrne, though he is young and never ran for major office before, has a name that is familiar to many in town. If his name was Steve Baker, the outcome might have been different, but not necessarily, because even Steve Baker, with no name recognition, may have benefitted from Joe Curro’s support. What gets difficult is separating the variables. It is foolhardy to consider the impact of name recognition on the race when you must also factor in the effect of 2nd votes from another candidate, not to mention several other variables that may have influenced the outcome. Bob Tosi may not have had the same name recogntion that the others did, and he finished fourth. Again, there are too many other variables that may have accounted for his fourth place finish. Maria Romano has run in three consecutive races. She has finished last in all three. One could argue that her name recognition made no difference in her case, but then again, it may have had a negative impact. Name recogniton can be a bonus for a candidate…except when it is not.
Money
Another factor to muse over is “money”. Bob Sprague, on his YourArlington site, posted a piece that disclosed each candidate’s campaign coffers. Steve Byrne led the pack and reportedly collected over $17,600 in receipts and spent about $5,700. Joe Curran was next with about $11,300 in receipts and $8,132 in expenses. Curro was not far behind him with $10,000 collected and $6,000 spent, respectively. Tosi took in $6,600 and spent nearly all of that amount. Romano collected over $2,500 but apparently spent nothing. The game people like to play is how much did each vote cost or cost-per-vote. So, we’ll bite.
If we play along and do the math, then Curro spent $1.75 per vote. Byrne spent roughly $2.03 per vote received. Curran spent about $3.00 per vote. Byrne raised more than Curran, yet Curran outspent Byrne and lost by almost 100 votes. Tosi spent $2.67 per vote. Romano spent $0 per vote. She may have spent nothing to get 1,400 votes, but she also came in last. In the Selectman’s race, if and only if we could eliminate all other variables (which we can’t), one might conclude that Mr. Curro got the most bang for his campaign buck. Spending just over two bucks per vote, Byrne got a decent bang for the buck, too. Curran spent the most and still didn’t win. Tosi spent the next highest amount and finished behind Curran. Curro finished first and yet spent only a tiny bit more than Tosi, who finished fourth. In sum, as far as the Selectman’s race is concerned, if you focus on money spent, you may be able to figure who got the most bang for their buck, but there is no evidence to suggest that the amount candidates raised or spent greatly altered the outcome of the race (at least among those finishing 1-4).
What about the School Committee race? Thielman spent $4,000 and received over 3,400 votes. His cost per vote was $1.16. Matt Pallett was the next biggest spender, spending $3,800. His cost per vote was $2.71. Cindy Stark or Cindy Starks spent $2,407 for a cost per vote of just $0.73. Flanagan spent almost $2,400 and received over 2,400 votes, making his cost per vote almost a $1.00 even. Jackson spent just 16 cents on the dollar for each vote he received. O’Brien received over 1,000 votes though he spent nothing. It can be argued that Cindy Starks got the most bang for her buck given that a) she won, and b) she spent less than Thielman per vote, who also won. Of those spending any money at all, Jackson spent the least, but has last place to show for it. With a pattern of one vote per dollar, one could postulate that if Mr. Flanagan spent $3,500 that this might have helped him compete better against Thielman and Starks. But again, one can draw no definitive conclusions.
In both races, money may have been a factor. It may have helped Mr. Byrne and Mr. Curran make the contest for 2nd place competitive. It may have helped Mr. Curro solidify what was thought by many to be an easy first-place finish. Money advantages may also have helped Mr. Thielman. However, as can be seen, spending the most does not guarantee you a win. What might be an easier conclusion to draw is the fact that those who spent no money or anything less than $2,000 did not win – or even come close.
Ballot Order
How much might ballot order have contributed to the election results? In the Selectman’s race, the order was as follows: Curran, Tosi, Romano, Curro, and Byrne. Curro and Byrne were fourth and fifth on the ballot, but managed to finish 1st and 2nd, respectively. Romano was in the middle of the pack and finished last. Curran’s first place positioning on the ballot did not help him get first place in the election – nor did it gain him 2nd place. In the Selectman’s race then, there is no evidence that ballot order made a difference. We will say that Joe with the “O” gave people the most vocal reminders about where they could find him on the ballot. We could go back and pore over results in previous elections, but we just simply do not have the time. Plus, as we stated before, we do not want to mix apples with oranges.
In the School Committee race, ballot order would at first appear to have paid off for Mr. Thielman. The Avocado may have done well to consider that factor a bit more carefully before issuing its predictions. Andrew O’Brien and Ian Jackson were second and third on the ballot, but they finished 5th and 6th. Flanagan was fourth but he finished 3rd. Starks was in the penultimate position on the ballot but managed a strong 2nd place finish. Mr. Pallett was last on the ballot but he did not finish last. As with the Selectman’s race, we can draw no definitive conclusions as to whether ballot order had any major impact. It may have had a negligible impact on Jeff Thielman’s vote totals (he did win the most votes of all the candidates in either race – even more than the very popular Joe Curro).
———————————————————————————————————————————————————
When considering all the above potential factors, remember to consider this:
In 2010, Kurt Fusaris was a first-time candidate for School Committee. He was last on the ballot, had no name recognition, had no big name endorsers, and had no big-time running-mates. He was up against two well-financed candidates with solid support from the powers-that-be in town. He raised just $125, spending a mere 7 cents per vote, and yet he garnered almost 1,700 votes. He finished 3rd in a 4 person-race where the top two finishers were elected.
————————————————————————————————————————————————–
CONCLUSIONS:
Some lingering questions are…
1) Did the likes of Joe Curro and Cindy Starks lose support as compared to past races?
2) Does this election prove that once again, the pro-incumbent attitude in town is stronger than ever?
3) Is there a wall or an invisible ceiling with respect to the numbers that will turn out – with the day of the election, the number of candidates running, the issues, and the general interest in the race notwithstanding?
To tackle the first question, again, we do not have the time to meticulously examine all the data from past elections. That said, at first glance, Joe Curro and Cindy Starks did receive fewer votes than they did in their last elections. In Joe’s case, it’s made trickier because he was running for a different office this time. In addition, he was running against some strong opposition in this election. Though some may have figured him to be a shoo-in and chose not to vote for him, that does not mean that they don’t still support him. Also, some may have chosen not to vote for him because he didn’t resign from his School Committee seat. All things considered, Curro received over 3,400 votes. One could argue that this is fairly representative of his solid base of supporters. Anything more may just be gravy. A good chunk of the 1,300 more votes he received in his last School Committee race may have been votes received from people who were not necessarily part of his “base”, but who were instead inclined to throw their support behind him because they did not like the other candidates or by virtue of the fact that he was an incumbent. We would therefore disagree with the argument that Mr. Curro has “lost” support. The same goes for Ms. Starks. Her support this time around is fairly consistent with the amount of support she received last time. Mr. Thielman received even greater support this time than he did last time around. One could argue that he’s become even more popular, not less so.
As to the second question, much has been said about Arlington being a pro-incumbent town. For this, we will look back to past elections. In the past five Annual Town Elections, incumbents were an impressive 15 for 15 in contested races. That’s not even counting the uncontested races. When new faces do join the Board or the School Committee, it’s typically only because there is a truly “open” seat. See Cindy Starks and Joe Curro in 2009, Jud Pierce and Kirsi Allison-Ampe in 2010, Bill Hayner and Dan Dunn in 2011, Steve Byrne in 2012. None of those individuals had to knock anyone out of office to gain election. All this too may not be absolutely conclusive, but it does lend strong support for the pro-incumbent hypothesis.
The third question speaks to whether there is an invisible ceiling to the numbers that will turn out for town elections. The past two Annual Town Elections averaged about 7,400 voters, even though there was a perceived greater interest in this year’s race and the fact that one election was held on a Saturday and the other on a Tuesday. Looking back, the numbers have fluctuated over the years and again, it would be difficult to pinpoint an exact ceiling. Honestly, the bigger question we should all be asking is: if there is a ceiling, how do we break through it and get more people to vote in the town elections?
A LOOK AHEAD:
Looking ahead to 2013, Jud Pierce and Kirsi Allison-Ampe will be facing expiring terms. Also, due to Joe Curro’s election to the Board of Selectman, there will be one additional seat up on the School Committee. Cindy Starks has gone on record as saying that she would rather not have Mr. Curro’s replacement on the Committee seek re-election to that seat next year. She does not want that individual to have the advantage of incumbency? As per our discussion about incumbency, this strikes us as just a bit ironic because Ms. Starks didn’t seem to mind the advantages incumbency brought her this year. We didn’t quite hear, “Shucks, I really wish I wasn’t the incumbent”. The other funny thing is that one can only wonder how quickly she would change her tune if the person they end up selecting to fill Joe Curro’s seat is someone she happens to really like a lot! We love Ms. Starks personally, even though we did not endorse her. However, it’s little things like that which at the very least, are just a bit curious. If the person they select is someone we really like too, we think they should have every right to seek the seat in their own right. If the person selected chooses not to run again, that should be their decision alone. In full disclosure, the Avocado will endorse Jud Pierce for re-election next year. Yeah, it’s a bit early we know, but we have to put it out there.
On the Board of Selectman, it will be Kevin Greeley’s seat that is up. It will be interesting to see if he decides to go for one more term. Now that Steve Byrne has been elected, this means that Mr. Greeley is no longer the “youngest” on the Board 