The Washington Post's Greg Sargent got ahold of a
worrying new memo from Progress Ohio, a labor-backed organization which is bearish on polls showing that SB5, the anti-collective bargaining bill, is headed for a wide defeat at the ballot box next month. As Sargent notes, these kinds of memos are "sometimes leaked to the press to game expectations and goose turnout," but Progress Ohio is claiming they
didn't intend for it to become public. That means their pessimism may not just be for show.
Indeed, I've been concerned for some time myself, given how difficult it is to poll ballot measures in general, and how badly off the polling was for a package of electoral reforms in Ohio in 2005. (Polls showed they could pass, but all four went down to wide defeats). And that's exactly the sort of thing the memo focuses on:
“Those predicting a blowout for our side are basing their analysis on flawed public polling samples." […]
“Modeling turnout for an off year ballot initiative is notoriously difficult,” the memo continues. “This is especially true in a state like Ohio where polling on ballot initiatives has been very unreliable.” […]
But the memo warns that the question wording in the two polls is so flawed as to be unreliable, because neither poll used the language voters will see on the Issue 2 ballot. “Keep in mind, neither of these polls tested the actual ballot language,” the memo says. “It’s a safe bet, if the actual ‘Issue 2’ language were polled that the margin would have been substantially narrower.” […]
“Bottom line: It’s nearly impossible to develop a reliable likely voter model for ballot initiatives,” the memo says, adding: “there are simply too many unknowns to believe these numbers are credible or even remotely possible.”
In a blog post, Progress Ohio adds two more points:
• SB5 = Issue 2? — Many people do not realize that SB5 will be repealed or upheld via Issue 2. After months of misleading ads from Building a Better Ohio and their Beltway allies, another subset of voters aren't certain which way they need to vote to get rid of SB5.
• Turn-out models — As was referenced in article, polling numbers were way off in most of the recent major ballot fights, such as the same sex marriage ban, 2009 casino measure and the 2005 election reform initiatives. Frankly, it is hard to know exactly who will turn out to the polls during for an issue campaign. This is true especially during an off-year election: the last round of polling on the casinos was 16 points off the final tally.
So should we be worried? I asked Public Policy Polling's Tom Jensen for his thoughts, since his firm's work was one of those being criticized:
My rationale for using a summary of SB5 instead of the actual ballot language is that this is what’s driving people to the polls, so I assume they’re going to know how they’re voting before they get there. If this was a Presidential election and the referendum was an afterthought, I’d use the exact language because people’s only exposure to the referendum might be reading it there on the ballot. But with a high profile one like this, I think it’s better to explain what the vote’s about than just to use the confusing language on the ballot.
I may be wrong in that thought process but that was my rationale. It’s definitely true that polling on Ohio ballot initiatives has been historically inaccurate.
I think that's a good point. Issue 2 is the key race on the ballot this November—the top of the ticket, as it were. This entire election is only taking place on account of this (and two other) ballot measures—no actual elective offices are up for grabs. Another point worth making is that even if you think PPP hasn't been asking the right question, they've been asking the same question all along, and that question showed improvement for the good guys in PPP's most recent poll. What's more, Quinnipiac (the other outfit which has polled Issue 2) has shown the exact same trendlines.
| PPP |
16-Oct |
14-Aug |
22-May |
13-Mar |
| Approve |
36 |
39 |
35 |
31 |
| Reject |
56 |
50 |
55 |
54 |
| Undecided |
8 |
11 |
10 |
15 |
|
|
|
|
|
| Quinnipiac |
25-Oct |
27-Sep |
20-Jul |
18-May |
| Kept |
32 |
38 |
32 |
36 |
| Repealed |
57 |
51 |
56 |
54 |
| Undecided |
11 |
11 |
12 |
10 |
Both firms showed strong support for repealing SB5, with a late summer/early fall tightening. Then, in polls released just a week apart, both saw improved numbers for repealing the bill. So even if the margins are too gaudy, they have at least both moved in the right direction. I certainly remain concerned, though, and I think Progress Ohio makes a number of valid points. With a week-and-a-half left before election day, we can't get complacent—but I'm sure no one in Ohio is.
