In the relationship between public opinion and government decision-making, the trajectory of influence goes from top to bottom. Policymakers try to mold public opinion to suit their needs, not mold policies to suit the public. On many controversial foreign policy issues, there is often a gap between informed public opinion and government policies. For example, according to polls, the majority of the public did not support the Reagan administration policy of arming the contras in Central America or the policy of “constructive disengagement” in South Africa, which amounted to support for the apartheid regime. In the Middle East, the majority has tended to favor a more neutral and therefore less pro-Israel policy than that of successive American administrations. 
Polls are designed to provide reliable, scientific information about public attitudes. As poll results are reported in the press, however, they also interact with opinion, giving officials the opportunity to generate changes in the views receiving public expression. As Benjamin Ginsberg so convincingly shows, in his analysis of how polls during the Vietnam era were manipulated by government officials, polls contribute to the domestication of opinion by helping to transform it from “a politically potent, often disruptive force, into a more docile, plebiscitary phenomenon.” 
US press coverage of Third World issues usually follows certain premises. First, only matters that relate in some way to US interests are reported. Issues in the world outside of Europe and North America are covered on a crisis basis. Second, the media are likely to take their cue from the State Department. Polls sponsored by major media organizations and survey organizations are usually designed to register the degree of public support for US government decisions or rationales. Not surprisingly, poll results tend to reflect media emphasis on various issues.  The extensive polling of public attitudes during the Gulf crisis amply illustrates this point. When the public cited reasons for US involvement, their answers tended to reflect official rationales.  In the Gulf war, the Bush administration was deeply worried about the effects of the “Vietnam syndrome,” and the emergence of a strong anti-war movement. The task of domesticating opinion and of controlling the flow of information became key aspects of the administration war policy. There was considerable public support for the initial deployment of troops to Saudi Arabia, but the arguments made by the administration in the period of escalating military buildup did not convince the majority of respondents. A November 20, 1990 CBS News/New York Times poll revealed that a majority of Americans were “increasingly worried that President Bush…has been too quick to commit American military forces to the Persian Gulf and has not been working hard enough on a diplomatic solution.” Approval of the way President Bush was handling the crisis dropped to about 50 percent (down from 59 percent a month earlier) and only 21 percent thought the US should initiate military action against Iraq.
A telling instance of how polls can contribute to the construction of policy rationales — though not policy itself — emerged as the administration sought to build its case for war. Only one rationale, which had not even been mentioned by any administration official, appeared convincing to the public: “to stop Saddam Hussein from developing nuclear weapons.” In addition, the majority thought that Congress “should have to vote a declaration of war before American troops go into combat.” This poll led to a revised emphasis in public pronouncements. Officials began to cite Iraq’s nuclear threat on a regular basis, and featured it when asking Congress for a resolution approving military action.
Probably more public opinion polls were conducted during the Gulf crisis than during any other conflict in contemporary history. The wording of questions revealed the level and the nature of the concerns. A CBS News/New York Times survey of February 12-13, 1991 asked the following range of questions: Was the war to defeat Iraq worth the potential loss of several thousand American lives; could Iraqi troops be forced out of Kuwait by bombing from the air or was a ground war necessary; should the US stop fighting once Iraqi troops were out of Kuwait or should they fight until Saddam Hussein was removed from power; were respondents satisfied with the amount of information about the effect of US bombing on Iraq; should the public have been allowed to see pictures of the ceremonies for the US war dead at Dover Air Force Base; and were respondents “edgy or short tempered or having trouble sleeping” because of the war?
Media accounts of poll results are often at variance with the actual data. Reporting on an August 16-19, 1991 CBS News/New York Times poll, for example, Michael Oreskes wrote that “the Persian Gulf crisis has propelled public approval of President Bush back to its highest level, with support for his stand against Iraq’s invasion of Kuwait overwhelming misgivings about his handling of domestic problems.”  The poll results, however, show that nearly half the respondents did not think that the Iraqi invasion of Kuwait constituted a real threat to the supply of oil to the US. Nearly 51 percent thought the conflict could be resolved diplomatically, without recourse to war. Of the 75 percent who approved Bush’s handling of Iraq’s invasion, half offered lukewarm support. A majority (53 percent) said they were unwilling to pay higher taxes to help defray the cost.
Another instance of misinterpretation was analyzed by Howard Schuman of the University of Michigan Survey Research Center. He painstakingly scrutinized the New York Times’ interpretation of a major poll of Israeli public opinion sponsored by the newspaper itself and released on the eve of Israeli Prime Minister Yitzhak Shamir’s 1989 visit to Washington. As a result of the Palestinian intifada, Shamir was under US pressure to come forth with a proposal for talks with the Palestinians. The Times’ lead interpretation is that Shamir is constrained by his intransigent public opinion. The actual data generated by the poll suggests that, in reality, Israeli public opinion is remarkably fluid, and willing to consider compromise solutions based on trading land for peace, on dealing directly with the PLO, and on recognizing the principle of Palestinian statehood, provided Israeli security concerns are taken into account. 
Schuman concludes that reports of public opinion results must consider the following issues: the exact wording of questions; the order in which questions are presented; the timing of the study; the nature and the size of the sample (the margin of error changes with the size of the sample). Cross-tabulations and yields by gender, by political affiliation, by levels of education and income usually reveal statistically significant differences among respondents. It is also important to examine whether the questionnaire is structured in such a way as to test for intensity of feeling about various issues: How strongly do people favor or oppose certain options? Finally, it is crucial to check for long-term trends, for evolution in attitudes, assuming that similar samples have been drawn over a period of time.
Asking the Right Question
Historically, polling questions about Middle East issues have been framed either by people who know very little about the region, or by people whose main concern is supportive American attitudes toward Israel. For many decades, the flow of information about the Middle East in the US was largely informed by pro-Israel concerns, widely seen as almost synonymous with American interests. This was true both in terms of media coverage and in the wording of opinion polls. A noticeable change began after 1978, culminating with Israel’s invasion of Lebanon and siege of Beirut in 1982. Public opinion polls then showed the greatest decline in pro-Israel sympathy since the high levels in the aftermath of the June 1967 war.  A number of subsequent events, including the Jonathan Pollard affair, Israel’s human rights abuses during the intifada, and Israel’s involvement in the Iran-contra scandal, contributed to a further erosion of American public support for Israel.
The polls finally began to reflect this underlying shift in public attitudes. The wording of questions also began to change: Opinion polls started to ask whether the American public sympathizes more with “Israel vs. the Palestinians,” in addition to the more standard combination of “Israel vs. the Arab nations.” The results showed more sympathy with the Palestinians vs. Israel than with the Arab nations vs. Israel. Dynamics in the Middle East and in US-Israeli relations had redefined the Arab-Israeli conflict to include the Palestinian dimension.
Opinion surveys presume that people have access to information which informs their attitudes. In reality, of course, the majority of the public has only the information that the media provides, especially about foreign policy issues. In the case of the Middle East, information is often selected and interpreted so that it reinforces existing stereotypes or creates new ones. A poll I sponsored in 1985, carried out by the Survey Research Center of the University of Michigan, asked respondents if they know enough about various national groups in the Middle East.
The majority said they lacked adequate information about the various countries. Asked whether they thought various Middle Eastern leaders were actively seeking a peace settlement or trying to block peace efforts, the majority also said they did not know.
Most questionnaires dealing with American attitudes on the Middle East limit responses to either/or. By contrast, a May 1985 study asked respondents: “In the Middle East conflict, do you think the US should favor Israel, favor the Arab countries, or should the US not favor one side over the other?” Only 20 percent said the US should favor Israel and 0.2 percent the Arab side. The great majority (70 percent) thought the US should favor neither side.
In October 1982, I commissioned a nationwide survey of American attitudes on the Middle East, conducted by the conservative Richard Wirthlin polling organization. I wanted to test what happens to respondents when they are given — through a filtering process — accurate, publicly known and reliable information about Middle East-related issues. The survey instrument had been pretested and refined to the extent that the sponsors as well as the pollster were satisfied that no questions showed a built-in bias in favor of one side or the other. We discovered that with each refinement, changes in attitude do occur, leading to the conclusion that better-informed persons are more likely to adopt a pro-peace, neutral position in the conflict between Israel and the Palestinians. This reflects a real and growing public awareness of a legitimate Palestinian grievance. Public sensitivity to the intricacies of the struggle may well be running ahead (not for the first time) of the familiar Washington reflexes.
Polls, which are quite costly, provide major private and public institutions and private corporations a powerful tool to shape, domesticate and manage public opinion.  Within the last two decades, substantial changes in American public opinion have occurred and the polls have begun to reflect them. Nevertheless, how the polls are read and how their results are interpreted are matters that often bear little resemblance to the reality embedded in the numbers.
 Poll results are usually available from polling agencies or from various centers such as Roper’s which store public opinion data. On Middle East policy and polls, see Barry Hughes, The Domestic Context of American Foreign Policy (San Francisco: Freeman and Co., 1978).
 Benjamin Ginsberg, The Captive Public (New York: Basic Books, 1988).
 Todd Gitlin, The Whole World Is Watching (Berkeley, CA: University of California Press, 1980); Herbert Gans, Deciding What’s News (New York: Pantheon, 1979); Gaye Tuchman, Making News (New York: The Free Press, 1978).
 In a Los Angeles Times poll of August 29, 1990, 50 percent of the public cited oil as the main reason, 28 percent suggested freeing of American hostages, 45 percent claimed the US was involved in order to deter aggression, 11 percent to restore the legitimate government of Kuwait and 13 percent to defend Saudi Arabia.
 New York Times, August 21, 1990.
 Howard Schuman, “The Polls: A Review,” Public Opinion Quarterly 54 (1990), pp. 409- 414.
 Fouad Moughrabi, American Public Opinion and the Palestine Question (Washington, DC: International Center for Research and Public Policy, 1986); and “Public Opinion and the Israeli-Palestinian Conflict,” American-Arab Affairs 30 (Fall 1989).
 A customized national poll costs in the vicinity of $100,000. Most polls are less expensive. One major pollster charges $1,400 per question for each sample of 1,000 respondents.