x Abu Dhabi, UAESunday 23 July 2017

Paint by numbers

The big idea The Arab opinion polls that drive headlines in the West are far from worthless, Marc Lynch writes, but we must understand their limits.

Enough about you, let's talk about me:
Enough about you, let's talk about me: "Post-September 11 surveys carried out by American organisations were overwhelmingly narcissistic: How do Arabs feel about America? About American policies? About American leaders? In short, how do Arabs feel about the issues Americans care about?"

The Arab opinion polls that drive headlines in the West are far from worthless, Marc Lynch writes, but we must understand their limits. According to a recently released survey of public opinion in six Arab countries conducted by Shibley Telhami of the University of Maryland, only 24 per cent of Arabs express negative views of Barack Obama, while 51 per cent are hopeful about his foreign policy. But 66 per cent say they have no confidence in the United States and only 18 per cent admit to positive views of it. Numbers like this drive headlines in American newspapers and produce anxious rumblings in the bowels of American diplomats. But do they matter?

Since the September 11 attacks, American concern with the state of Arab public opinion has produced a boom in survey research: a wide range of private and governmental agencies now survey Arab views on almost every imaginable topic. As polling has exploded, it has also become intensely controversial: those who refuse to believe that American foreign policy is the main cause of anti-American sentiment routinely dismiss the validity and significance of polls showing precisely that.

In the past few years, when American policy was devoted almost exclusively to "combating violent extremism", it was possible to argue that such broad survey results had little use for policymakers. The salafi- jihadists of al Qa'eda and their passive supporters were never a mass movement: their numbers were microscopic and unlikely to be represented in opinion surveys - and they did not, at any rate, tend to respond to queries from pollsters. Views of America had little to do with support for al Qa'eda, and vice versa: campaigns to diminish support for terror groups did not necessarily increase support for American policies.

But today that insight no longer holds: al Qa'eda has become even more marginal, but among the Arab mainstream broad support has formed for a mass-based "Resistance", which is now the principal focus of Arab political discourse and argument. The views that matter now are those of the mainstream rather than a narrow slice of extremists - and the issues are political, not religious. So understanding public opinion is critical, but without considering the context for each poll - and the methodological challenges facing pollsters - some of these numbers may be worse than useless. In private, survey research professionals offer scathing critiques of the methodological problems - some of which are endemic and structural, while others are rooted in the specific practices of particular pollsters. The realities of authoritarian states with pervasive intelligence apparatuses, where self-censorship is a well- ingrained survival strategy, make it less likely that individuals will offer honest opinions on sensitive topics. The absence of reliable census data in many countries, often for political reasons, poses challenges for pollsters trying to create reliable frames for random sampling.

The false patina of science, and the obvious appeal of numbers for journalists and policy analysts, has led to uses and abuses far beyond what the data can bear. But polls are still better than the alternatives (talking to taxi drivers, for instance), and the best surveys today have been designed with these methodological concerns in mind - based on intensive internal discussions about possible biases and omissions.

These internal debates, however, are rarely matched in rigour by the public consumption of the findings. While the last American presidential election turned political junkies into expert poll-readers, able to discuss, if not understand, how different survey methods tilt findings, there is very little comparable analysis for Arab public opinion polls. One reason for this difference is that Arab public opinion surveys are still sparse, unevenly distributed and rudimentary compared to the sophisticated apparatus surrounding American political campaigns. In addition, polling in America aims to measure opinion in advance of elections - where attitudes are ultimately converted into outcomes. Nothing similar exists in most of the authoritarian Arab countries, where public opinion does not easily translate into changes in the political arena. This helps to explain the regionwide alienation from and apathy toward national institutions found in the surveys themselves - one recent survey of Jordanians found that only four per cent had a positive view of their Parliament.

Where there are significant elections, on the other hand - in places like Palestine, Kuwait and Lebanon - there has been much more survey research, and it shows that opinion shifts rapidly in response to the actions and statements of political leaders. A steady stream of Palestinian polls, for instance, has shown that public views about Hamas and Fatah move in response to their behaviour and performance in power.

Public opinion can only be understood in the context of its political system: it matters in non-democratic countries, but in very different ways than in a democracy. Without such a political context, analysis becomes almost impossible. The results of one survey of confidence in the Jordanian government led the liberal newspaper columnist Jamil al Nimri to throw up his hands in frustration: "these surveys continue to give results that it is difficult to do anything with… the numbers rise and fall without any comprehensible justification… this is not the fault of the surveyors, who follow professional standards and methods, but because of the absence of a real political life."

It is easy to conclude that in such authoritarian societies mass opinion matters little - though elite opinion may matter a bit more and the opinion of the leaders most of all. But that is too narrow a view. The rise of pan-Arab satellite networks and newspapers, the growing penetration of the internet, and the increasing turbulence of contentious politics makes public opinion more relevant. Even more crucially, almost every actor in the region - even if they say the opposite - pays close attention to public opinion and acts as if it matters.

For instance, Iraqi Prime Minister Nouri al Maliki may agonise privately about the impending American withdrawal from Iraq, but as a smart politician he responds to the fact - according to the latest ABC News survey - that 86 per cent of Iraqis want the US to leave either on the current schedule or sooner. Arab leaders may not really care about the fate of the Palestinians, but they have learnt through experience that it would be political suicide to say so in public. And some leaders may see eye to eye with Israel on key issues such as Hamas or Iran, but are rarely stupid enough to admit it - again, because they believe that public opinion does in some way matter.

Beyond the immediate political process, sophisticated public opinion surveys are a necessity for academic research on values, attitudes and worldviews. Survey research after September 11, for example, eventually demonstrated the true extremism of al Qa'eda's doctrines, countering common claims that such views were widespread in the Arab world. Gallup's extensive research on Muslim opinion has produced a deeply textured picture of a diverse and relatively moderate Islamic world. The Program on International Policy Attitudes found with a well-designed survey that less than 10 per cent of Egyptians approve of attacks on civilians in the United States, but that nearly 90 per cent agree with attacks on American forces in Iraq.

Such a finding confirms the extent to which al-Qaeda's tactics fall outside the Muslim mainstream - but it also shows that the public supports what it perceives to be legitimate resistance to foreign occupation in Palestine or Iraq. A plethora of opinion surveys have demonstrated beyond question that the vast majority of Muslims like the idea of democracy and see no conflict between democratic reforms and Islamic values, no matter how many people try to claim otherwise by citing verses from the Quran.

There has been a dramatic increase in recent years in the amount of polls conducted by Arab governments and NGOs - helping to correct the emphasis of the first wave of post-September 11 surveys carried out by American organisations, which were overwhelmingly narcissistic: How do Arabs feel about America? About American policies? About American leaders? In short, how do Arabs feel about the issues Americans care about? These polls produced results, but without any sense of how much the issues really mattered to the people being surveyed. For instance, survey research has consistently found that economic and quality of life issues - rather than American policies or politics - are foremost among Arab concerns.

Opinion research that explores deeper cultural matters and local political issues will be far more useful than news-making surveys about anti-Americanism. Mark Tessler, who heads the Arab Barometer project - and is the leading American academic working on Arab public opinion - argues that efforts should be directed toward "explanation rather than descriptions" in order to assemble a complex picture about attitudes and their causes rather than bullet-point numbers.

"FIFTY PER CENT ARABS HOPEFUL ABOUT OBAMA" is fine for headline writers, but it is far less important than acquiring a sense of why they are hopeful, or about what would vindicate, or dash, their hopes.
Marc Lynch is associate professor at the Elliott School of International Affairs at George Washington University. He writes a blog on Arab politics and media for Foreign Policy.