There has, of course, been lots of commentary about the Globe’s poll, released over the weekend, that shows Charlie Baker within one point of Governor Patrick. As far as I know, however, this detail, which I find quite interesting, has so far escaped commentary. It’s taken from page 36 of the full poll.
| Baker | Patrick | Cahill | Stein | Undecided | |
| Definitely decided (34% of respondents) | 37% | 53% | 4% | 3% | 2% |
| Leaning (21% of respondents) | 48% | 37% | 8% | 4% | 3% |
| Still trying to decide (44% of respondents) | 25% | 20% | 19% | 5% | 28% |
With the important caveat that these smaller subsets necessarily carry larger margins of error, these numbers tells us a couple of things. First, it says that Patrick has closed the deal with a lot more voters than has Charlie Baker, since Patrick is crushing Baker amongst voters who have definitely decided how they’re voting.
Second, and relatedly, it says that a lot of people who now (tentatively) think they are voting for Baker remain persuadable, either by Patrick or by Cahill (or perhaps even Stein). I suspect that a lot of those people are tentatively voting Baker on the somewhat abstract idea that things are tough right now so a change must be a good idea, but when it gets more concrete, they’re just not sold on Baker as the right guy. That’s a big opportunity for the Patrick campaign on which I hope and expect they will capitalize.
Third, it says that polling remains a peculiar business. The total number of “don’t know” responses on the horse-race question is 14%. Yet far more than that – 44% – of the poll’s likely voters say that they “have considered some candidates but are still trying to decide.” How is that different from an “undecided” voter? Or, since those 44% apparently expressed a preference when asked the horse-race question, how is it different from a “leaner”? Furthermore, why do only 28% of the respondents who are “still trying to decide” subsequently respond that they are “undecided” in the horse-race question? That seems oddly low to me; surely, if you say that you’re “still trying to decide” who to vote for, there’s a better-than-even chance that you’re undecided in the race.
Of course, if you parse these numbers too fine, you will drive yourself crazy without actually learning very much. Still, these “level of commitment” numbers strike me as worthy of notice.
Nothing I can’t stand more than people who are indecisive in their indecision!
There are definitely some peculiarities in these numbers!
and the rest of us too. Really our time is better spent on the doors or on the phone or in your case David, at your desk writing inspirational tomes that get the rest of us on the doors and on the phones instead of parsing. I think Doug is right. These close numbers should prompt us to stop analyzing them and get crazy Uncle Clarence out of his recliner and five, count them 5, nephews, grand sons and grand daughters home from college on Columbus day to sit down and fill out their absentee ballot applications. Except the kid from Everett.
<
p>Sorry couldn’t resist.
What would Nate Silver do?
<
p>Having read the details of his underlying math and his approach towards individual polls, I am not ready to get upset at the results of a single poll. There are way too many assumptions in how the poll is gather that create the situation much like the one in 2008, when BO was supposed to upset HRC in New Hampshire. Spikes in polling mean less than long term trends and the long term trend is Charlie Baker staying relatively flat. When it falls back to normal numbers next week, it will not be the result of the public changing its mind but the methodology of this one poll that is probably an outlyer.