For help on this article or on the parts of the dissertation on which it is based, the author wishes to thank the following individuals and institutions: James P. Sewell; William Foltz; Bruce Russett; Peter Busch; Robert Mandel; Joseph Micallef; The Carnegie Endowment for International Peace; ARPA (Advanced Research Projects Agency) Behavioral Sciences, Contract No. N0014-67-A-0097-007, administered by the Office of Naval Research (a grant administered by Bruce Russett); Nathaniel Beck; Robert Wilson; Andre Cohen, and anonymous referees for this journal. While the credit should be shared, the responsibility for any errors which remain is the author's alone.

¹ See especially Riggs, Robert E., “One Small Step for Functionalism; UN Participation and Congressional Attitude Change,” International Organization 31 (Summer 1977): 515–539CrossRefGoogle Scholar, and Wolf, Peter, “International Organization and Attitude Change: A Re-examination of the Functionalist Approach,” International Organization 27 (Summer 1973): 347–371CrossRefGoogle Scholar.

² Riggs, “One Small Step.”

³ Two other systematic studies are not discussed here since flaws make them of doubtful relevance to this discussion. They are: Volgy, Thomas J. and Quistgaard, Jon E., “Learning about the Value of Global Cooperation: Role-taking in the United Nations as a Predictor of World Mindedness,” Journal of Conflict Resolution 19 (06 1975): 349–376CrossRefGoogle Scholar, and the study reported by Galtung, Ingrid Eide in “Are International Civil Servants International? A Case Study of UN Experts in Latin America,” Proceedings of the International Peace Research Association Inaugural Conference (Assen: van Gorcum, 1966): 198–209Google Scholar, and in “The Status of the Technical Assistance Expert: A Study of UN Experts in Latin America,” Journal of Peace Research 3 (1966): 359–379CrossRefGoogle Scholar. For a discussion of the flaws, see Peck, Richard Lee, “United Nations Influence on Member-States Through the Socialization of Permanent Representatives” (Ph.D. dissertation, Yale University, 1976), pp. 73 and 81Google Scholar.

⁴ This is, moreover, a doubly interesting study because it is virtually the only published study which goes beyond delegate socialization to investigate possible influences on organizational behavior, a topic on which its findings are sobering, even though they do accord with functionalist theory in a sense.

⁵ Although a random sample was first drawn, it soon became evident that refusals to grant interviews were not evenly distributed, but were concentrated in the smaller missions of Africa, Asia, and Latin America. Since the uneven response rate was already causing a departure from randomness, I decided to attempt to preserve at least some representativeness by replacing those missions in which I could not obtain interviews with other missions randomly drawn from the same geographic areas. I also added the two superpowers to my sample on a presumption that a sample of UN countries which did not include at least one of the superpowers was representative only in a mechanical statistical sense. As the resulting sample is thus not random, the tests of significance presented in this paper cannot be taken as true tests of significance; but they do give some indication of the strength of the relationship relative to the size of the sample. Finally, several statistical tests presented in Peck, , “United Nations Influence,” pp. 230–237Google Scholar, suggest that the sample is as representative of the population as a true random sample might have been. The remaining tendencies toward unrepresentativeness are identified in the text.

⁶ Alger, Chadwick F. suggested this approach in a footnote to his “United Nations Participation as a Learning Experience,” Public Opinion Quarterly 27 (Fall 1963): 411–426Google Scholar.

⁷ See text near footnote 47.

⁸ A third group—essentially the LDC group without the Latin Americans—was also analyzed. Although there are several reasons for thinking that influences on attitudes might be different for these “new” countries, the analysis found them to be little different from the LDCs. Hence, results for this group are not presented. Although it would also have been logical to analyze the Soviets separately, they had too small a representation in the sample for a separate analysis to be possible.

⁹ Kendall's Tau-b was used as a measure of correlation, since it is non-parametric and requires fewer assumptions about sampling distributions than the more commonly used parametric measures. See Blalock, Herbert, Social Statistics (New York: McGraw-Hill Book Company, 1960), pp. 319–324Google Scholar, and Nie, Norman, Bent, Dale H., and Hull, C. Hadlai, Statistical Package for the Social Sciences (New York: McGraw-Hill Book Company, 1970), p. 154Google Scholar. The test of significance used was that used by SPSS.

¹⁰ The indices were constructed from responses to the questions by taking the means of the nonmissing observations (using standardized scores) on the questions included in the index. Since inter-item correlations on the indices suggested the indices were not very homogeneous, a variety of indices was used. The discussion takes account of the results on all indices used. If several indices, each highly imperfect itself, show similar results, that at least reduces the likelihood that the result is an artifact of a particular index.

¹¹ Although the principal approach was multiple linear regression, checks were also made for parabolic and hyperbolic fits. For details, see Peck, , “United Nations Influence,” pp. 107–109Google Scholar.

¹² Four factors generated from national attributes by Vincent, Jack E. were used as control variables. See his Factor Analysis in International Relations: Interpretation, Problem Areas, and an Application (University of Florida Press, 1971), pp. 44–46, 52Google Scholar. And see Peck, , “United Nations Influence,” p. 107Google Scholar, for further details.

The control variables help eliminate spurious results deriving from possible parallel influences of national attributes on both independent and dependent variables. They also give some assurance that our results reflect real socialization and not merely a pattern in which delegates learn over time more about their government's interests and the relevance of the UN to those interests (a pure “rational response” pattern rather than socialization). One might expect that government interests would be closely associated with national attributes. Controlling for those attributes has the effect of leaving to be explained by the delegate's tenure only that variation in the dependent variables which remains after the attribute variables have explained all of the variation that they can. While the association between national attributes and national interests is undoubtedly highly imperfect, it is hard to know how one might otherwise control systematically for the effects of national interests on the pattern of delegate attitudes.

¹³ The factor analysis used principal component analysis and varimax rotation. For the entire sample of respondents it included nearly all of the responses to questions discussed in this article. Separate factor analyses were run for the western countries and for the less developed countries, although with a more limited set of variables. A number of variables were eliminated from the matrices for these analyses to get around singularity problems which arose when the full matrix was used.

¹⁴ A similar effect of tenure at the EC was found by Pendergast, William R. in his “Roles and Attitudes of French and Italian Delegates to the European Community,” International Organization 30 (Autumn 1976): 669–677CrossRefGoogle Scholar.

¹⁵ Two indications of delegate tenure at the UN were used. One was the length of the delegate's full-time experience in his mission (”mission time”) and the other included as well sessions attended by the delegate even when he was not a member of the permanent mission (”UN time”).

¹⁶ A rough calculation finds fewer results for the non-western delegates on the correlations than might have been expected by pure chance. Moreover, many of the results found for these delegates were on questions which had very low response rates (questions 5, 6, 7, and 8) and appeared to be due as much to an earlier time of first arrival at the UN as to a longer period of exposure. This latter distinction is based on the following consideration:

In 36 of the 47 cases it was possible to distinguish between the number of years of presence of the permanent representative at the UN and the number of years since his first arrival at the UN. Generational influences—both those associated with chronological age and those associated with a possibly different kind of peer group in the early years of the UN—seem conceptually distinct from the influences of tenure in the UN milieu which are the focus here. Hence, generational influences were checked as well as tenure influences, and if it seemed possible to attribute observed results to generational influences, the results were discounted accordingly.

¹⁷ Kendall's tau is .43 with “mission time” and .28 with “UN time” with two-tailed significance levels of .02 and.14 respectively.

¹⁸ These results appeared in regressions on four of five indices for the full sample and were mirrored by results found among the western countries but absent altogether in analyses done for the LDCs. Several of the indices obtained their best fit in a hyperbolic equation rising asymptotically to the right. For more detail, and the estimated equations, see Peck, , “United Nations Influence,” pp. 107–110Google Scholar and the appendices cited there.

¹⁹ Clark, John F., Michael, K. O'Leary, , Wittkopf, Eugene R., “National Attributes Associated with Dimensions of Support for the United Nations,” International Organization 25 (Winter 1971): 1–25CrossRefGoogle Scholar.

²⁰ Stoessinger has argued convincingly that the “financial issue” in the United Nations is not really financial, but rather political—a demonstration that nation-states are unwilling to pay for UN activities which are against their interests. However, the question of whether the UN should have a financial independence from its member-states is clearly a question with strong supranationalist implications. It may in fact be the most “supranationalist” of the questions asked here, since a positive response strongly implies a willingness to see a weakened national control over significant resources of the UN. As Stoessinger argues, the issue at stake is not the continued existence of the UN, but different conceptions of its proper role. Even if the conceptions are politically inspired, they nevertheless have sharply different supranationalist implications. See Stoessinger, John G. et al. , Financing the United Nations System (Washington, D.C.: The Brookings Institution, 1964), and especially pages 29–33Google Scholar. See also Stoessinger, John G., The United Nations and the Superpowers: China, Russia, and America, Fourth edition (New York: Random House, 1977), pp. 121–145Google Scholar.

²¹ Few results of significance appeared in the correlation analyses for the LDCs, and no results of significance appeared in the regressions.

²² Kendall's taus were -.33 with mission time and -.45 with UN time, with significance levels of .09 and .02 respectively.

²³ Kendall's tau was .37 with UN time, with a significance level of .06. Some caution is needed in the inference that support for an active Secretariat is equivalent to support for the UN. Two respondents justified their preference for a moderately active Secretary General by observing that the current climate of opinion at the UN is such that an active role for the Secretary General might be counter-productive, and a western delegate suggested that “it may be useful to have a pendulum effect in the movement from passive to active and back again.”

²⁴ Kendall's tau was -.39 with UN time, with a significance level of .04.

²⁵ It is tempting to call this frustration-based expansionism the “in for a penny” syndrome (”In for a penny, in for a pound”) in contrast to the “bird in hand” syndrome in which delegates are pleased with past achievements of the organization and conservative in their reactions to future expansion. The “bird in hand” term is from Hansen, Roger D., “European Integration: Forward March, Parade Rest, or Dismissed?” International Organization 27 (Spring 1973): 239CrossRefGoogle Scholar. It might be worth noting that two other combinations of reactions to the past and preferences for the future are logically possible. One, combining dissatisfaction with the organization's past performance and restrictionist preferences for the future, appears close to the classical nationalist response to international organizations. The other, combining satisfaction with past performance and expansionist preferences for the future, appears close to the pattern which classical functionalists expected to see. Frustration-based expansionism, by contrast, would appear typical of organizations going through the “spillover” crisis so much analyzed by neo-functionalists.

²⁶ For one index, time in the mission appeared to be important in that the addition of this variable after the control variables were already in the equation raised the explained proportion of the variance by more than 10 percent. For the other index, the time in mission variable was statistically significant at .10 (two-tailed). Since the number of respondents in the western group is smaller than the number in the LDC group, relationships must be stronger to be statistically significant for this group. In an effort to compensate for this, I considered variables which contributed at least 10 percent to the explained variance (R²) in the dependent variables after the control variables were already in the equation to be “important” variables which might indicate the presence of a relationship. For detail on the content of the indices and for the estimated regression equations, see Peck, “United Nations Influence,” pp. 119–120 and the appendices cited there.

²⁷ Regressions using the factor score on this factor as the dependent variable, but in every other way exactly the same as the regressions reported above, suggested that the delegate's exposure to the UN had a positive influence on this factor which was statistically significant at .05.

²⁸ The factor discussed here is an unrotated factor.

²⁹ That is, in the linear equation tenure has a positive coefficient significant at .05; a separate estimation of a hyperbolic equation finds a curve rising asymptotically to the right to be significant at .01.

³⁰ See the text near Table 5.

³¹ These results appear clearly in correlations. The correlation between UN time and the response to question 25 gives a tau of -.30 significant at .03, and the correlation between UN time and the response to question 26 gives a tau of -.35 significant at .01. The correlation between UN time and the response to question 27 gives a tau of .44 significant at better than .005.

They are also reflected (although more weakly) in the regressions. None of the regression results is statistically significant, even at .10 two-tailed; but the number of General Assembly sessions attended by the delegate seems to have a negative effect on his level of expansionism which raises the R2 for the equation by more than 10 percent. See Peck, “United Nations Influence,” p. 255, for the estimated equations.

³² See note 20 for a further discussion.

³³ See note 31 for the results on question 27. On question 30 for the LDC's tau is .34 with UN time, significant at .01 two-tailed.

³⁴ Results on three indices support this conclusion. For details on the content of the indices and for the estimated regression equations, see Peck, , “United Nations Influence,” p. 131Google Scholar, and the appendices cited there.

³⁵ In regressions using this factor score as a dependent variable and using independent variables identical to those used in the regressions reported above, previous sessions of the General Assembly attended has a positive coefficient and is significant at .05.

Although the structure of these factors—frustration coupled with expansionism—is similar for the western countries and the LDCs, the fact that the LDCs are beneficiaries of the economic programs which are the focus of their expansionism in a way not true of the western countries for the expansion of the Secretariat suggests that the parallelism of structure may be partly irrelevant. I am grateful to an anonymous reviewer for this journal for pointing this out to me. See also the text following note 41 and following note 50.

³⁶ Nevertheless, very few respondents draw a link between their government's engagement in the UN and their career prospects, as the responses to question 38 indicate.

³⁷ An early study of relevance is Reissman, Leonard, “A Study of Role Conceptions in Bureaucracy,” Social Forces 27 (03 1949): 305–310CrossRefGoogle Scholar. For the most part, however, the literature really began with Gouldner, Alvin W., “Cosmopolitans and Locals: Toward an Analysis of Latent Social Roles—I,” Administrative Science Quarterly 2 (12 1957): 281–306CrossRefGoogle Scholar. More recent treatments are numerous. See, for example: Glaser, Barney G., “The Local-Cosmopolitan Scientist,” American Journal of Sociology 69 (1963): 249–260CrossRefGoogle Scholar; Goldberg, Louis C., Baker, Frank, and Albert, H. Rubenstein, “Local-Cosmopolitan: Unidimensional or Multidimensional?” American Journal of Sociology 70 (1965): 704–710CrossRefGoogle Scholar; Miller, George A. and Wager, L. Wesley, ”Adult Socialization, Organizational Structure, and Role Orientation,” Administrative Science Quarterly 16 (06 1971): 151–163CrossRefGoogle Scholar; Sheldon, Mary E., “Investments and Involvements as Mechanisms Producing Commitment to the Organization,” Administrative Science Quarterly 16 (06 1971): 143–150CrossRefGoogle Scholar; Bergerand, P. K. and Grimes, A. J., “Cosmopolitan-Local: A Factor Analysis of the Construct,” Administrative Science Quarterly 18 (06 1973): 223–235CrossRefGoogle Scholar; Flango, Victor E. and Brumbaugh, Robert B., “The Dimensionality of the Cosmopolitan-Local Construct,” Administrative Science Quarterly 19 (06 1974): 198–210CrossRefGoogle Scholar; Sorensen, James E. and Sorensen, Thomas L., “The Conflict of Professionals in Bureaucratic Organizations,” Administrative Science Quarterly 19 (03 1974): 98–106CrossRefGoogle Scholar. The analysis of dimensions underlying the cosmopolitan-local role orientations has become increasingly complex as the literature has matured. Gouldner originally suggested three dimensions: loyalty to the organization, commitment to specialized skills; and reference group orientation. See his “Cosmopolitans and Locals,” p. 290. Although the basic dimensions have persisted in more recent analyses, they have grown in complexity. Moreover, there is some ambivalence about the importance of the reference group orientation dimension in the literature. See in particular: Glaser, “The Local-Cosmopolitan Scientist”; Goldberg, Baker, and Rubenstein, “Local-Cosmopolitan”; Berger and Grimes, “Cosmopolitan-Local”; and Flango and Brumbaugh, “The Dimensionality.” The complexity of the dimensions seems in part a function of the particular group of respondents studied. The three dimensions used here represent a considerable disregard for this complexity. They were derived from an analysis of the questions used in Goldberg, Baker, and Rubenstein, “Local-Cosmopolitan,” Miller and Wager, “Adult Socialization,” and Sheldon, “Investments and Involvements.”

³⁸ Results on three of four indices consistently showed parabolic relations of reasonably high statistical significance between time spent in the mission and career orientation. Results on the fourth index were in the same direction, but were not statistically significant. The actual minima occurred at 4.6, at 4.4, and at 4.7 years on the three different indices. For details on the content of the indices and for estimated regression equations, see Peck, , “United Nations Influence,” pp. 138–140Google Scholar and the appendices cited there.

³⁹ These two separate factors appeared in each of the separate factor analyses run for the sample as a whole, for the group of western countries, and for the group of less developed countries.

⁴⁰ That is to say, orthogonal.

⁴¹ The estimated equation for a hyperbola rising asymptotically to the right was significant at .05. For details and for the estimated regression equations, see Peck, , “United Nations Influence,” pp. 145–146Google Scholar, and the appendices cited there. Since the factor involved here includes a negative emphasis on the question about tenure in the mission included in the indices of UN career orientation, this finding is roughly consonant with the falling half of the parabolic relationship found for these delegates between tenure and UN career orientation as measured by the indices.

⁴² There were some hints of links with tenure—suggesting that those longer at the UN cared less about their reputations with their colleagues or with the Secretariat. But, given the responses on the basic questions, the interpretation of these results which is most persuasive is simply that permanent representatives who have been longer at the UN are more self-confident and less concerned about their reputations in general than permanent representatives with less experience.

⁴³ For mission time, tau is .34 significant at .08; for UN time, tau is .36 significant at .07. See question 39 in the appendix.

⁴⁴ For UN time, tau is .21 significant, if one can call it that, at .45.

⁴⁵ The relationship was hyperbolic, that is to say. For details on the content of the indices and for the estimated equations, see Peck, , “United Nations Influence,” pp. 151–152Google Scholar, and the appendices cited there.

⁴⁶ For a more detailed development of this point, see ibid., pp. 152–155.

⁴⁷ The indication that this may reverse itself after the first four or five years at the UN, however, may be more a result of self-selection, given the earlier socialized attitudes of these delegates, and given the fact that by the fourth or fifth year opportunities for rotation are likely to have arisen which would not have been likely in the first year or two. It should be noted, of course, that if selfselection is operative in the later years of contact with the UN, the results observed may understate the true effects of socialization.

⁴⁸ Five respondents were not asked this question or did not respond to it for other reasons.

⁴⁹ ” Actually, the link between the delegate's tenure and his career orientation seems to be a parabolic link, with the minimum career orientation toward the UN found near the midpoint of the delegate's fifth year in his mission.

⁵⁰ For an attempt to examine links between permanent representative tenure and UN voting patterns, however, see Peck, , “United Nations Influence,” pp. 166–202Google Scholar.

⁵¹ See the text near Tables 1 and 2.

⁵² Riggs, , “One Small Step,” p. 537Google Scholar.

⁵³ One might even argue that the cross-sectional approach to the socialization question has one advantage over a longitudinal approach. Interviewing and re-interviewing the same individual at different times leaves open the possibility that any observed attitude change is not so much the result of socialization as the result of a rational response to a changing environment. (Even in a changing “rational response,” of course, one would expect to find socially influenced perceptions to be important.) The cross-sectional approach implies that the environment has not changed substantially over the time of the interviewing, since all interviews are conducted in a few months' time. Of course, it still remains the case that the permanent representative's environment differs from respondent to respondent and that patterns of association between tenure and attitudes may reflect a “learning” of rational responses rather than some more radical kind of socialization. As mentioned in note 22, the control variables used in the regression equations represent one attempt to reduce difficulties arising from this source. However, it still remains the case that for LDC delegates at least, the UN socialization which does take place is not much different from socialization to their governments' views—something one might want to call a “rational response” rather than socialization.

⁵⁴ And Siverson's study used documentary evidence, but found no attitude change. Of course, the Egyptian-Israeli case in 1956 would pose a very severe test of even the strongest socialization.

⁵⁵ Note, however, that Pendergast in “Roles and Attitudes,” used several dimensions with affective components.

⁵⁶ Indeed, note Karns’ interview evidence that many of his subjects had little exposure to the issues other than through the interparliamentary meetings (”The Effect of Interparliamentary Meetings,” p. 500) and evidence presented by Karns (p. 499) and Riggs, (”One Small Step,” p. 523Google Scholar) suggesting that very few Congressmen were on the UN delegation more than once. For those who were, Riggs analyzed only their first UN experience.

⁵⁷ For some hints toward this consideration, however, see: Volgy, and Quistgaard, , “Learning about the Value of Global Cooperation,” p. 350Google Scholar footnote 1; Sawyer, Jack and Guetzkow, Harold, ”Bargaining and Negotiation in International Relations,” in International Behavior; a Social-Psychological Analysis, ed. Kelman, Herbert C. (New York: Holt, Rinehart and Winston, 1965), pp. 507–508Google Scholar; Sewell, James Patrick, UNESCO and World Politics; Engaging in International Relations (Princeton, N.J.: Princeton University Press, 1975), p. 309Google Scholar; and Pendergast, “Roles and Attitudes,” p. 677Google Scholar.

⁵⁸ Riggs also makes some use of this kind of argument, emphasizing the greater strength of European political parties as a possible explanation for the lack of observed attitude change among European parliamentarians.