With less than five months to go until Election Day 2010, political junkies everywhere are trying to find the perfect tea leaves with which to forecast the outcome.
One that received a little attention from the Washington Post this week revolved around one of those little bits of conventional wisdom that portends glad tidings for the GOP--the wide disparity in releasing of campaign, or internal, polling in this cycle:
House Republicans have unleashed a slew of internal polls in recent days, seeking to put the country but, more accurately, the media and the GOP donor base, on notice the playing field for the fall is rapidly expanding.
From Ohio's 13th district, which was on nobody's map until very recently, to Oregon's 1st district, which went for President Obama by 25 points in 2008, these internal polls are aimed at making the case that Republicans have a plethora of targets heading into November.
To be sure, anybody who follows the semi-nightly Polling and Political Wrap here at Daily Kos knows that there has been a huge gap emerging between the number of partisan poll released in this cycle by the GOP and those released in this cycle by Democrats. The ratio, at this fairly early stage, is somewhere in the neighborhood of 3-to-1 or 4-to-1.
The practical political value for the Republicans is obvious. Flooding the zone with positive internal polling reinforces the meme that Republicans are ascendant in this electoral cycle. Furthermore, as the Post article by Aaron Blake notes, it has the added benefit of convincing the donor base to dip into their pockets more fervently, since competitiveness tends to loosen bankbooks in political campaigns.
Of greater concern is trying to divine the reason why Democrats have largely gone silent on the internal polling front.
The most obvious answer is also the most unsettling one.
If the flood of GOP polling is not being answered by contrary data by the Democratic Party, it does not take a intellectual leap of faith to suggest that the reason for their silence is that their data is not...well...terribly contrary. Obviously, the Democrats have absolutely no incentive for confirming rosy Republican data with pro-GOP data of their own.
There is, for what it is worth, an alternate explanation. As Crisitunity at Swing State Project put it earlier this week: "do the Democrats just not have good news in those districts to counter with, or (as many have speculated) are they engaged in a bit of expectations gaming/rope-a-dope?"
That explanation, while clearly less plausible than the prior one, does have at least a thread or two of logic to it. Bad polling news can have one positive effect--it can knock loose any complacency and bring a sense of urgency for the Democrats (see an example of this strain of logic in a story about a recent poll showing the Oregon Governor's race tied up).
The problem, however, with that line of logic, is that there is no recent precedent. Looking at the past three electoral cycles, there has been a fairly clear connection between the partisan makeup of internal polling releases and how the parties performed respectively in November.
I arrive at these numbers by utilizing a (nerd alert!) database I compiled of over 2700 polls during the 2008 election cycle. I winnowed out presidential polling, and also excluded partisan pollsters whose primary releases are public in nature (insert Rasmussen jokes here). I wasn't nerdy enough did not generate such a database for the 2004 or 2006 cycles. For those cycles, the excellent polling database at DC's Political Report was utilized.
Internal Polling releases, by party, 2004-2008
2008: 164 Democratic Polls, 88 Republican Polls
2006: 209 Democratic Polls, 88 Republican Polls
2004: 122 Democratic Polls, 119 Republican Polls
As you can see, during the wave cycles of 2006 and 2008, the Democrats released significantly more partisan polling data than did the GOP (more than 2-to-1 in 2006). In 2004, which was a cycle narrowly carried by the GOP, the numbers were almost perfectly at parity.
Before Democrats should start adorning ourselves in funereal garb, however, there is a caveat. The one thing, anecdotally, which seems unique about this cycle is how early the internal polling releases on general election races have come. In the three cycles covered above, the bulk of the polling came in September and October. A lot of that partisan divide occurred in the later months. This stands to reason--as the push toward election day comes, those candidacies that are outperforming expectations are eager to push their polling out there to inspire both the media and donors alike.
Therefore, the predictive value of internal polling releases tends to be, as with all polling, what gets released late, rather than what was released early. In that sense, the polling divide is not necessarily a harbinger of doom. At least not yet.