¹ See, for example, Miller, Warren E. and Stokes, Donald, “Constituency Influence in Congress” this Review, 57 (03, 1963) 45–56Google Scholar; and Miller, , “Voting and Foreign Policy” in Rosenau, James (ed.), Domestic Sources of Foreign Policy (forthcoming)Google Scholar.

² See Campbell, Angus, et al., The American Voter, (New York: Wiley, 1960), p. 527Google Scholar. For summaries of poll data on public support of World War II and the Korean War, see: Campbell, J. T. and Cain, L. S., “Public Opinion and the Outbreak of War,” Journal of Conflict Resolution, 9 (1965), 318–328CrossRefGoogle Scholar.

³ This survey was conducted in late February and early March, 1966, by the National Opinion Research Center (NORC) with a questionnaire designed and pre-tested at Stanford. The respondents were 1495 adults. To obtain the sample NORC drew probability samples of blocks within each of its seventy-five sampling points throughout the country and sent its interviewers to select men and women within those blocks (or comparable areal units) to correspond to the nation in age, sex, and employment status.

⁴ New York Times, March 10,1966, p. 10; see also Sulzberger, C. L., New York Times, March 7, 1966, Section IVGoogle Scholar; and Newaweek, March 7, 1966. These reports have references both to the “Harris Survey” (Washington Post, Feb. 28, 1966) and “private” polls.

⁵ See, among others, Converse, Philip E., “The Nature of Belief Systems in Mass Publics,” in Apter, David E. (ed.), Ideologies and Discontent (New York: The Free Press, 1964)Google Scholar; McClosky, Herbert, Hoffmann, Paul J., and O'Hara, Rosemary, “Issue Conflict and Consensus Among Party Leaders and Followers,” this Review, 44 (06, 1960), 406–427Google Scholar; Key, V. O., Public Opinion and American Democracy, (New York; Knopf, 1961)Google Scholar; Stouffer, S., Communism, Conformity and Civil Liberties, (Garden City: Doubleday, 1954)Google Scholar; and Almond, Gabriel, The American People and Foreign Policy (New York: Harcourt, Brace, 1950)Google Scholar.

⁶ For example, the thrust of the February Harris Survey (op.cit.) was that the increasing disaffection with the “job President Johnson is doing in handling the war in Vietnam,” was from those who favored increasing our military effort. This was interpreted in the White House (New York Times, March 10, 1966)Google Scholar as indicating that a majority of those opposing the President were “Hawks.” But an analysis of the Harris report does not support this interpretation. If we examine the Harris data to ascertain whether those who had an opinion on escalation (8% were “not sure”) approve or disapprove of the President's handling of the war we find the following:

Compared with the January data, the February poll results show a trend in the direction of the interpretation but not majority support. The trend, moreover, was apparently temporary; by late April, according to Harris, (Washington Post, 05 9, 1966)Google Scholar, support for the President had dropped to 47% “good-excellent,” 53% “fairpoor” and anti-escalation sentiment had increased:

Since the May press release does not report the cross-tabulation of support and escalation sentiment, it is not certain that Johnson opposers were anti-escalation but it is a reasonable surmise.

⁷ Congressional Record, March 17, 1966, p. 5863 and April 28, 1966, p. A2320. For editorial comment on the poll, see, for instance, New York Times, March 17, 1966; St. Louis Post Dispatch, March 16, 1966; and San Francisco Chronicle, March 16, 1966. See The Washington Post, April 7, 1966, p. A25 for the original attack on our survey and ibid. April 16, 1966, Letters to the Editor Column, for our reply. For an account of press reaction to our initial report, see: Polsby, Nelson W., “Doves, Hawks, and the Press,” Transaction (04, 1967)Google Scholar.

⁸ For a full report on all questions and all marginal results, consult Public Opinion and the War in Vietnam (March 15, 1966, Institute of Political Studies, Stanford University, Stanford, California).

This report has been published, in Spanish: “La opinión publica en Estados Unidos de América sobre la guerra de Vietnam,” Revista Mexicana de Sociología, 28 (Abril-Junio, 1966) 357–76.

⁹ For instance, 86% of our respondents knew that Congress had not declared war, 68% knew we were bombing targets in the North at the time of the survey, 47% could name the capital of South Vietnam and 41% could name the capital of North Vietnam. It is hard to be certain whether to consider these figures “high” or “low.” It is our impression that they represent fairly high levels of information on an issue of foreign policy.

¹⁰ The Gallup Polls, covering the period in which we were in the field, show results on the same question somewhat at variance with our results. AIPO reported lower levels of “approval” than we found although the level of “disapproval” was about the same:

“In general, do you approve or disapprove of the way President Johnson is handling the situation in Vietnam?”

a. AIPO data are from press releases carried in the San Francisco Chronicle, March 9, and April 6, 1966.

¹¹ Since these results might be affected by who it was that the respondent perceived the Vietcong to be—their identification as North Vietnamesedominated or indigenous South Vietnamese rebels being a matter of debate—we cross-tabulated these results against identification of the Vietcong. Attitudes toward negotiations with them is not dependent upon the way in which they are indentified or mis-identified.

¹² See footnote 4 above.

¹³ Guttman scaling procedures were used to test whether sets of items which seemed intuitively to measure the same underlying attitude were in fact related in a unidimensional pattern. An eight-item ‘escalation’ scale and an eight-item ‘de-escalation’ or peace-initiative scale met the criteria for satisfactory Guttman scales in two separate random subsamples of the total. Coefficients of reproducibility were in excess of .90. Coefficients of scalability were in excess of .60 (see Menzel, Herbert, “A new coefficient for soalogram analysis,” Public Opinion Quarterly, 17, 1953, 268–280)CrossRefGoogle Scholar. The distribution of scale scores was significantly different from chance by chi square test at the p<.001 levels.

The escalation scale included questions asking approval or disapproval of the level of military involvement at the time of the survey. The respondents were asked whether they would be willing to continue the fight in Vietnam if it meant: bombing of military targets in North Vietnam (77% approved), 200,000 troops in Vietnam (61% approved); questions asking about increases in manpower involvement: 500,000 troops (45%), all-out mobilization (40%); and escalation questions involving China and Russia: fighting the Chinese in Vietnam (56%), fighting the Chinese in China (32%), fighting an atomic war with China (29%), and fighting an atomic war with Russia (22%).

The de-escalation scale included items ranging from approval of withdrawal from Vietnam even if it meant a Communist takeover in Laos and Thailand as well as Vietnam (13% approve) to approval of negotiations with the Vietcong (88% approve). Items in between included: immediate withdrawal (15%), letting Vietcong eventually gain control (28%), gradual withdrawal while letting Vietnamese work out their own problems (39%), letting the Vietcong participate in a coalition government (52%), permitting free elections (54%), and agreeing to a UN-supervised truce (69%). That both sets of items—those on escalation and those on de-escalation—can be ordered on Guttman scales and that these orderings parallel the ordering of items that a politically sophisticated observer would arrive at as being more or less extreme escalation or de-escalation measures (atomic war with Russia being a more extreme move than atomic war with China, for instance) lends support to the argument that attitudes on the war form a meaningful pattern.

We asked respondents several questions concerning their willingness to pay possible domestic economic and welfare costs “if we are to continue the fighting.” Approval for “increasing taxes at home” was 31% and disapproval 66%. “Putting ting government controls over wages and prices” got 41% approval and 53% disapproval. “Reducing aid to education” got 19% approval and 79% disapproval. “Spending less money for the War on Poverty” recived approval from 46% and disapproval from 51%. “Reducing the Medicare program” received 28% approval and 65% disapproval. For correlational analysis we combined the three welfare costs items into a single index and the two economic items into another index. (Guttman scaling analysis indicated that the two kinds of items were not sufficiently correlated to justify a combined scale).

¹⁴ See our original report (cited in footnote 8) for a full report of these data. Consistent “hawks” took the following positions on four items: they favored increasing our commitment to 500,000 men and favored bombing North Vietnam cities; they opposed a coalition government that included the Vietcong and opposed elections if there were a chance the Vietcong might win. Consistent “doves” took the opposite four positions. We found 6% of our sample to be consistent “hawks” and 14% to be consistent “doves.”

¹⁵ See Lipset, S. M., “The President, the Polls, and Vietnam,” Transaction (09-Oct. 1966), p. 20Google Scholar.

¹⁶ See footnote 13 for a description of these scales.

¹⁷ The pattern of consistency can also be observed if we look at the mean escalation scores in relation to willingness to see taxes raised and aid to education cut “if needed to continue the war.” Those who approve of reducing aid to education and raising taxes have the highest mean escalation score, while those who oppose raising taxes and reducing education aid have the lowest.

¹⁸ See Warren Miller, “Voting and Foreign Policy,” in Rosenau, op. cit.

¹⁹ The eight information items in the questionnaire failed to fit the Guttman scale pattern, and were combined into an arbitrary index ranging from 0 to 8, with the higher score indicating more correct answers. Questions included identification of the capitals of North and South Vietnam, knowledge of number of U.S. troops in Vietnam and number of U.S. casualties at the time of the survey, whether Congress had declared war, whether North Vietnam was then being bombed, and identification of the Vietcong. For the last item, both North Vietnamese or South Vietnamese communists were accepted as “correct” answers.

²⁰ For an exploration of three models of the relation of information and opinion, see: Gamson, W. A. and Modigliani, A., “Knowledge and Foreign Policy Opinions: Some Models for Consideration,” Public Opinion Quarterly, 30 (1966), 187–199Google Scholar.

²¹ To keep these conclusions in perspective it might be useful to point out that, even though the correlations are significantly different from zero, the approximately –.20 correlation between information and support for de-escalation policies means that less than 5 percent of the variation in policy scores can be accounted for by differences in information level.

²² See Converse, Philip E., Clausen, Aage R. and Miller, Warren E., “Electoral Myth and Reality: The 1964 Election,” this Review, 59 (06, 1965), 321–323Google Scholar.

²³ It must be pointed out that we are not talking of letter writers specifically on the Vietnamese war. We found only 17% of our respondents had ever written a letter on any subject, and any attempt to isolate a specific group that had written on the Vietnamese war would have given us too few cases to analyze. This limits the meaning of the following analysis, but not completely. The Converse, Clausen and Miller article uses a similar general measure of letter writing. But this illustrates the limitation of the cross-section survey in obtaining enough cases of significant policy groups.

²⁴ Op. cit., pp. 332–336.

²⁵ They differ, however, from the non-writers in terms of consistency roughly as do the informed from the less-well-informed. They are somewhat more likely to be consistent in that the negative correlations between escalation and deescalation scales are larger, (—.52 for the letter writers and —.34 for the non-writers), the correlation between escalation scores and willingness to pay welfare costs is higher, (.39 for the letter writers and .21 for the non-writers.) But, as with the information items, the distinction is not clear. They are no more likely than the non-writers to be willing to pay the economic costs of escalation. The correlation between escalation scores and willingness to pay economic costs is .37 for both writers and non-writers.

²⁶ Actually, a closer look at the findings in the Converse, Clausen and Miller article suggests that their findings are similar to ours in one important respect. Though they find a sharp difference between the mass public and the letter writers on most issues, they find much greater similarity on the one foreign policy issue that is reported, “negotiation with the communists.” The discrepancies on this issue are so slight in comparison with the discrepancies discovered on the domestic issue (op. cit., figures 2a-e, p. 334), that the analysis of opinion on this issue via letters could have served as a reasonable estimate of the shape of the distribution of public opinion. On the four domestic issues, in contrast, expression through letters was a highly inaccurate guide to public opinion. See below for a further discussion.

²⁷ See Miller, op. cit., and John Corry's news story on Vietnam as an issue in the 1966 Congressional election (New York Times, Oct. 20, 1966, p. 1).

²⁸ See Greenstein, Fred, “Sex-Related Political Differences in Childhood,” Journal of Politics, 23 (05, 1961), 353–371CrossRefGoogle Scholar; Lane, Robert E., Political Life, (Glencoe: The Free Press, 1959), 209–216Google Scholar; and Duverger, Maurice, The Political Role of Women (Paris, 1955)Google Scholar.

²⁹ Gamson and Modigliani (op. cit., p. 189) hypothesize and demonstrate that “the greater the attachment to the mainstream, the greater the degree of conformity of one's foreign policy opinions to official policy.”

³⁰ cf. Ekman, P., Tufte, E., Archibald, K. and Brody, R., “Coping with Cuba; Divergent Policy Preferences of State Political Leaders,” Journal of Conflict Resolution, 10 (1966), 180–197CrossRefGoogle Scholar.