Syndicate content

Benchmarking Expectations: Pre-Election Polling and Accountability

Taeku Lee's picture

The on-going controversy around the presidential election result in Iran raises an important curiosity.  It is clear at the present moment that the official results have defied expectations and dashed hopes for many.  From the standpoint of political accountability, there are at least two important questions that arise.  First, where do these expectations and hopes come from?  Perceptions that the election was "stolen" must be based in some sense of a range of plausible outcomes, and the declared 63% to 34% split clearly fell out of this range for Moussavi supporters and comfortably within this range for Ahmadinejad supporters.  The problem of conflicting pre-election expectations is an old one, rooted in what social scientists often call "homophily."  Where we stand is often determined by where we sit, and we tend to sit in deeply embedded and entrenched social information networks amongst others who are very much like us in body, mind, and spirit.  Those in the Ahmadinejad camp most likely set their expectations in the company of other Ahmadinejad supporters and those in the Moussavi camp most likely set their expectations in the company of fellows who championed Moussavi's cause.

The second key question, then, is whether some mechanism or institution, prior to the election, could have managed such conflicting expectations.  Could better – more precise, more neutral, less partisan – information have averted, or at least restrained either the degree of electoral fraud or the violent response to the perception of fraud, or both?  Information networks and epistemological grooves are awfully difficult to extricate oneself from.  But if expectations can be fairly and accurately benchmarked and the "will of the people" ascertained, the two sides of a fight will at least have some basis for knowing if they are on the side of the angels or if their zeal is weighted down by a cloud of dissonance.

This sort of neutral, information-based mediation is a role that political opinion polls were meant to play, from their inception.  Pioneers of pre-election polling like George Gallup and Archibald Crossley fully believed that the science of statistical sampling could, in James Bryce's words, achieve "government by public opinion" if "the will of the majority of citizens were to become ascertainable at all times, and without the need of its passing through a body of representatives, possibly even without the need of voting machinery at all."  In more mature electoral democracies like the United States, there is broad consensus about the proper place of "horserace" polls.  Candidates may complain and voters may turn a cynical and skeptical eye, but polls are almost always quite accurate about where voters stand and everyone expects them to be so.

In the 2009 Iranian elections, there were a number of pre-election polls.  For the most part, however, they were disregarded either because they were directly sponsored by one of the candidates or widely viewed as unfairly biased to favor a particular candidate.  Perhaps one exception in this was in fact a pre-election poll commissioned by two Washington, DC based think tanks (Terror Free Tomorrow and the New America Foundation) and conducted by KA Europe.  Public opinion polls conducted by Terror Free Tomorrow (TFT) aim explicitly, "by finding people's own priorities," to "empower public opinion against authoritarian dictatorships and terrorist minorities."

The results of this poll – in the field from May 11th to 20th, 2009 – are revealing.  On the presidential election itself, the publicly downloadable report is a bit opaque.  The report indicates that 34% of Iranians polled would vote for Ahmadinejad, 14% for Moussavi, 2% for Mehdi Karroubi, and 1% for Mohsen Rezai.  Of the remainder, 8% explicitly opted for "none" of the four candidates, 15% refused to answer that question, and 27% did not know for whom they would vote.  The nub, here is that roughly 1 in 2 Iranians appear to have expressly avoided choosing one of the candidates running for president. 

The prevalence of non-response evokes a whole host of limitations with polls that have emerged over the years.  Seemingly small and arbitrary differences in statistical sampling design, response rates, interview context, question wording, question ordering, and other confounding factors can result in large differences in a poll's findings about where "public opinion" stands on a given issue.  Moreover, there are often wide cultural differences – across nation-states as well as within them – in the perceived legitimacy of polling and in the degree to which citizens feel free to voice their political opinions.  This latter concern is especially likely to have a bearing for a poll funded by non-Iranian entities bearing names like "Terror Free Tomorrow," "the New American Foundation," and "KA Europe." 

Did roughly half of the Iranians in this particular poll report avoid choosing one of the four candidates running for president because they genuinely did not know who to vote for, because they did not like the choices before them, because they were afraid to disclose their true opinions, or because they did not want to reveal their views to this particular polling firm?  We have no way of knowing.  What can be said is that cultural anxieties about information technologies like polling will not change over night.  Given this reality, a minimum condition for polls is transparency over its methods and questions.  Here the TFT poll report does an excellent job of meeting such standards of transparency. 

Furthermore, because one in two Iranians in the TFT poll chose not to favor a candidate officially running for president, we have no way of knowing whether the officially sanctioned 63% to 34% result is within the realm of possibility or clear evidence of fraud.  If the 34% for Ahmadinejad to 14% for Moussavi split reported in the TFT poll simply carried over to these "non-responders," the election-day outcome would be roughly a 70% to 30% split (close to the Interior Ministry's official results).  If, on the other hand, non-responders were more likely to be reform-minded voices of dissent, the election-day projection could be entirely different.  To lend some credence to this scenario, 60% of those who "didn't know" how they would vote supported political reform and change in principle, a position that is likely to have pushed them to favor Moussavi.  More generally, report also shows that 77% specifically supported choosing Iran's Supreme Leader (currently the Ayatollah Ali Khamenei) through direct elections, a further challenge to the alliance between Ahmadinejad and Khamenei.

Where does this leave us?  My main point in this lengthy entry is simple enough: pre-election polls, properly conducted and imbued with legitimacy, can serve an important accountability function.  They can benchmark public expectations and, in doing so, perhaps they can limit the extent to which votes are miscounted and perhaps they can defuse public outcry among the electorally vanquished.  As the present case of the TFT poll and the presidential elections in Iran demonstrate regrettably clearly, however, this aspiration for pre-election polls is easier conceived than it is achieved.

Photo Credit: Flickr User Hamed Saber


Add new comment