Ever since General Petraeus testified before Congress several weeks ago Democratic strategists have faced a complex challenge in interpreting the subsequent opinion polls on Iraq. On the one hand polls asking if the number of troops should be “decreased” or the troops “withdrawn” or if “deadlines” should be imposed continue to receive majority support. Yet at the same time, solid majorities of the electorate also offer their “support” or “approval” of the plan put forth by General Petraeus.
Now for many routine political purposes – campaign stump speeches, e-mail fundraising letters, rousing sermons to grass-roots supporters – ambiguous data like this are easy to handle – one just dismisses the incompatible poll data as a “distortion” and not the “real” public opinion.
Here’s how it’s done. Every politically aware person starts with a very firm gut feeling that he or she knows what the average person “really” thinks. We say to ourselves “If I could just talk to that average guy for 10 minutes across a kitchen table, I know he’d agree with me that the troops should come home. As a result, it’s obvious that adding the name of a General like Petraeus into a survey question just artificially biases the outcome. Poll questions without any well-known names included in them are a better measure of what the public “really” thinks about withdrawal”.
This kind of logic seems entirely reasonable until we happen to overhear the other side doing exactly the same thing. They say “I know if I could talk to that average guy for 10 minutes, he’d agree with me that he really doesn’t want America to lose this war. Vapid, hypothetical questions about when people wish the troops could come home are meaningless. They don’t indicate the consequences or describe the price of pulling out. The only “real” measure of public opinion on withdrawal from Iraq is the number of people who support or oppose the Petraeus plan.”
In practice what actually happens when poll data is ambiguous like this is that both sides just cherry-pick the polls and use only the data that supports their particular view – a process which ends up being more than slightly tedious. Politicians and commentators relentlessly throw opinion data at each other like World War I doughboys lobbing hand grenades across the trenches. Fortunately for the Dems for the last year or so the polls have generally contained as much or more data that that suggested support for prompt withdrawal as against it, making these rhetorical “poll wars” more or less a draw.
But for the serious formulation of Democratic political strategy, on the other hand, cherry-picking the data is not an acceptable solution. To make plans for the coming elections the Dems have to try to understand what the contradictory polling data actually indicates about the state of American public opinion and what the real balance between anti-war and pro-administration opinion really is.
Fortunately, last week Gallup released one of the methodologically clever studies they periodically pop up with – a study that goes a long way toward providing some useful answers.

First, among a variety of other questions Gallup asked their respondents if they favored a plan that would:
“…reduce U.S. troop levels in Iraq from roughly 170,000 to about 130,000 by next summer without making commitments for additional troop reductions until that time”
66% agreed.
Then a little later they asked the same respondents if they would favor a plan that:
“…would withdraw most U.S. troops from Iraq within 9 months time, with any remaining U.S. troops serving in Iraq only filling support roles.”
63% agreed.
What was going on? It turned out that almost half of the respondents who said yes to plan A also said yes to plan B. What this indicated was that, as Gallup noted: “any plan that includes withdrawal has a good chance of gaining at least initial support from Americans… either of the plans tested in this research has majority support from the American public.”
From a common sense point of view, this actually makes sense. Most Americans no longer believe that “victory” in Iraq is in any meaningful sense possible and, on an emotional level want the troops somehow out of harms way. But they do not know how it should be done and are willing to energetically endorse essentially any plan that is presented to them as feasible, regardless of the details or authors.
Gallup then took the additional step of forcing the respondents to choose between the two proposals above but without explicitly indicating their political origin. 53% chose the “most troops out in 9 months” proposal – essentially a Democratic initiative that failed in the Senate – while 44% preferred the very minimal, non-binding approach that is the Bush Administration’s preferred solution.
On the one hand, advocates of prompt withdrawal from Iraq will be pleased to see that – purely on the merits of the two proposals – more Americans prefer a prompt departure rather than the administration approach. But at the same time, it must also be recognized that in the real world of the political campaign next year the choices will not be presented in this “purely on the merits” sort of way.
With Bush’s personal popularity reduced to Nixonian levels and Cheney’s popularity hovering somewhere around the level of OJ’s and Michael Jackson’s, it is quite certain that the administration will relentlessly describe its Iraq strategy as “The Petraeus Plan” every single time it is discussed. With support for that plan running in the mid-50% range and Petraeus’ personal popularity up somewhere in the mid-60’s, the 53-44 split between the two plans will quickly narrow when Petraeus’ name is attached to the second proposal – probably bringing support for the two plans to within the margin of error. Thus for serious planning of political strategy the prudent assumption will have to be that current public opinion on prompt withdrawal from Iraq is basically divided about 50-50.
Some anti-war advocates will resist accepting a downward revision of this sort in the estimate of national support for prompt withdrawal from Iraq and will prefer to continue to cherry-pick the polls to prove that substantial majorities actually support their view. For conventional day-to-day partisan politics this is both entirely inevitable and not necessarily inappropriate.
But for the serious discussion of Democratic political strategy taking this approach would represent a serious mistake. There is nothing actually progressive, principled or resolute in deliberately overestimating the numbers of one’s own supporters or underestimating the size of the opposition.
Businessmen who underestimate the size and number of their competitors end up bankrupt; military officers who underestimate the size and number of their opponents end up captured or dead. Democratic strategists, when they talk among themselves about strategy, should seek only the maximum possible accuracy in their estimations. It is on this basis, and this basis only, that campaigns can be effectively planned and elections successfully won.
James Vega is a strategic marketing consultant whose clients have included major American nonprofit organizations and leading high-tech firms.