For Bloomberg (and Quinn), there's a lot riding on the definition of 'fair'
Earlier this week, Democratic consultant Steve Sigmund argued on this site that there has been little evidence in the polls to suggest that City Council Speaker Christine Quinn will suffer among Democratic primary voters because of her association with Mayor Michael Bloomberg.
In response, the dean of the City Hall press corps, New York Post bureau chief David Seifman, pointed a recent poll from NY1-Marist showing the mayor with a "negative" approval rating in every borough except Manhattan.
As is pretty much always the case when results appear to contradict each other (even though, strictly speaking, Sigmund and Seifman are asserting different things), the wording of the poll questions makes a crucial difference in the outcome.
This chart uses Quinnipiac University polls of Bloomberg's job-approval rating over his third term:
Quinnipiac conducted 13 polls between November 2010 and January 2013 in which they asked whether voters "approve or disapprove" of the way Bloomberg has done his job, breaking down responses by political party.
Bloomberg's approval number ebbed in the early-to-middle part of 2011, but it improved in September 2011, then stayed around 50 before climbing up to around 60 late last year.
As Sigmund (who is not formally affiliated with any of the campaigns) pointed out, a Bloomberg ally like Quinn may well suffer if she's perceived to be subservient to him, but the idea that simply being close to Bloomberg is bad politics, even in a Democratic primary, isn't demonstrated by those numbers.
Then there's that NY1-Marist poll.
It is accurate to say that it showed Bloomberg with a negative job-approval rating in the outer boroughs, based on Marist's findings, but it's worth noting that the wording of the vital question in the poll was different.
While Quinnipiac asks whether voters "approve" or "disapprove" of Bloomberg's job performance, Marist gives voters four choices: Excellent, Good, Fair and Poor. In theory two of those are negative and two are positive.
So Marist combines "excellent" and "good" to come up with its approval rating, which for Bloomberg looks like this:
The issue is with the word "fair," which isn't clearly a pejorative.
Not to be like this, but: According to Webster's dictionary, "fair" can mean: "pleasing," of "flawless quality," "superficially pleasing," "not stormy or foul," "free from self-interest, prejudice or favoritism" or "not dark." The tenth definition given says fair means "sufficient but not ample."
I myself would not consider it, in the context of normal conversation, to be a negative judgment on someone to say he or she is doing a fair job.
So here, this following chart is fantastical, because it combines the responses in a way that the pollster never intended. But for what it's worth, this is what Bloomberg's approval rating would look like if the "fair" responses were counted among the approvers.
In a brief interview, Marist's veteran pollster Lee Miringoff defended his survey's wording, saying that giving people four choices "gets the intensity" of voters' feelings.
He admitted that some of the same people who rated Bloomberg "fair" in the Marist poll, if presented with Quinnipiac's options, might say that they "approve" of Bloomberg's tenure.
But, "fair," he said, shows support that is "extremely soft and not deliverable," and therefore, for Marist's purposes, not positive.
So just using the unambiguously positive "excellent" and "good," eight Marist polls taken between January 2011 and February 2013 show Bloomberg up among Democrats from the mid 30s to just shy of 50.
Which isn't great, but isn't quite the stuff backlashes are made of, either.