Published online by Cambridge University Press: 27 January 2009
Are Happy Days for government really here again? Recent polls suggest that the long slide in public confidence in America's political institutions and authorities has finally ended. Ronald Reagan, who came to Washington to bury government rather than praise it, ironically has presided over a restoration of trust in the competence of national leadership. We begin this article by charting the contours of the unanticipated improvement in the public's image of government, assessing the magnitude of the increase in confidence, identifying the social groups whose outlook has changed and specifying the institutions that have gained in popular esteem. Our main purpose, however, is to provide an explanation for the resurgence of trust in government that addresses persistent controversies about the theoretical and empirical status of this concept.
1 For evidence on the decline of public confidence in government, see Miller, A. H., ‘Political Issues and Trust in Government: 1964–70’, American Political Science Review, LXVIII (1974), 951–72CrossRefGoogle Scholar; Citrin, J., ‘The Changing American Electorate’, in Meltsner, A., ed., Politics and the Oval Office (San Francisco: Institute of Contemporary Studies Press, 1981)Google Scholar, and Lipset, S. M. and Schneider, W., The Confidence Gap (New York: Macmillan, 1983)Google Scholar. On the recent shift in public outlook, see Miller, A. H., ‘Is Confidence Rebounding?’ Public Opinion, VI (08–09 1983), 16–20Google Scholar; and Lipset, S. M., ‘Feeling Better: Measuring the Nation's Confidence’, Public Opinion, VI (04–05 1983), 6–9, 56–8.Google Scholar
2 See Miller, , ‘Political Issues and Trust in Government’Google Scholar; Crozier, M., Huntington, S. P. and Watanuki, J., The Crisis of Democracy (New York: New York University Press, 1979)Google Scholar; and Caddell, P., ‘Trapped in a Downward Spiral’, Public Opinion, II (10–11 1979), 2–7, 52–5, 58–60.Google Scholar
4 This issue was joined in the exchange between Miller and Citrin in the 1974 American Political Science Review and is carefully reviewed in Abramson, P. R., Political Altitudes in America (San Francisco: W. H, Freeman, 1983), pp. 195–205.Google Scholar
7 Lipset, S. M. and Schneider, W., ‘Confidence in Confidence Measures’, Public Opinion, VI (08–09 1983), 42–4Google Scholar. They attribute this different result to a heavier ideological ‘load’ in the Michigan measures of trust. Our own view is that the conflicting findings about the timing and extent of the upturn in trust in government are as much a function of differences in the particular object on which pollsters' questions focused. Thus, between 1981 and 1984, confidence levels always appeared higher when respondents were asked about the ‘White House’ instead of the ‘federal executive branch’, a term that encompasses the entire national bureaucracy and not just the president and his leading associates. In fact, the proportion of the public expressing a ‘great deal’ of confidence in the White House increased from 17 to 20 per cent between November 1980 and November 1982, a small change to be sure, but one that is consistent in direction with the shift recorded by the Michigan questions over the identical period.
8 These data thus confirm earlier research demonstrating that, while the mass public may discriminate among political institutions, images of any particular institution or ‘government in general’ tend to be diffuse and undifferentiated. See Citrin, J., ‘Political Alienation as a Social Indicator: Attitudes and Action’, Social Indicators Research, V, pp. 391–4.Google Scholar
11 Because of the erratic composition of the NES interview schedules and our desire to compare the 1980, 1982 and 1984 data, we are forced in this article to employ a two-item index combining answers to the ‘trust government’ and ‘big interests’ items as our summary measure of political confidence. Miller and Citrin used a measure composed of five items. Although we could have included the ‘waste tax money’ question and created a marginally more reliable measure, this item was omitted because previous analyses have indicated that it has relatively weaker interrelationships with the other political trust items and relatively stronger ideological and racially oriented connotations. Moreover, the two-item measure has been used in cross-national research that included an American sample. For this research see Barnes, S. H. and Kaase, M. et al. , Political Action: Mass Participation in Five Western Democracies (Beverly Hills, Calif.: Sage, 1979), especially Chap. 14Google Scholar. However, the use of the longer three-item index changes none of the substantive findings of this study.
14 In order to conserve space, we do not report the supporting statistics here. They are presented in the extended version of this paper delivered at the 1985 World Congress of the International Political Science Association. A copy of these data can be obtained by writing to the authors.
15 For a comprehensive review of these issues, see Kiewiet, D. R., Macroeconomics and Micropolitics (Chicago: University of Chicago Press, 1983).Google Scholar
17 For a review of opinions concerning national problems and their management during these years, see Goodman, J. L. Jr., Public Opinion During the Reagan Administration (Washington D.C.: The Urban Institute Press, 1983).Google Scholar
18 Keeter, S., ‘Public Opinion in 1984’ in Pomper, G., ed., The Elections of 1984 (Chatham, N. J.: Chatham House, 1985), pp. 91–5.Google Scholar
19 He makes this argument in the 1974 article, ‘Political Issues and Trust in Government’, and repeats it with reference to the recent trend in ‘Is Confidence Rebounding?’, p. 19.Google Scholar
20 A useful summary of these trends is Ladd, E. and Lipset, S. M., ‘Public Opinion and Public Policy’, in Duignan, P. and Rabushka, A., eds. The United States in the 1980s (Palo Alto, Calif.: The Hoover Institution, 1981), pp. 49–84.Google Scholar
22 Abramson, P. M., Aldrich, J. and Rohde, D., Change and Continuity in the 1984 Elections (Washington D.C.: Congressional Quarterly Press, 1986), Chap. 6.Google Scholar
23 Brody, R. and Page, B., ‘The Assessment of Policy Voting’, American Political Science Review, LXVI (1972), 450–8CrossRefGoogle Scholar, and Shanks, J. M. and Miller, W. E., ‘Policy Direction and Performance Evaluation: Complementary Explanations of the Reagan Elections’, unpublished paper presented at the 1985 American Political Science Association meetings.Google Scholar
24 Since respondents at the ideological extremes may be more politically cynical than their more moderate counterparts, arguably out of frustration that the incumbent administration has not gone far enough in redirecting the course of government, we have collapsed the seven-point issue continua into three groups (1–3 as liberal, 4 as moderate, 5–7 as conservative) in the computation of these correlations.
25 Kinder, D., ‘Presidential Character’, unpublished paper delivered at the 19th Carnegie Symposium on cognition at Carnegie Mellon University (1985), p. 4.Google Scholar
26 See the summary in Page, B., Choices and Echoes in Presidential Politics: Rational Man and Electoral Democracy (Chicago: University of Chicago Press, 1978), Chap. 8.Google Scholar
27 The independent variables were scored with missing data responses placed at the midpoint in order to increase the number of available cases. However, a separate analysis in which missing data cases were excluded did not change the nature of the results.
28 We have considered and rejected several methodologically oriented objections to the conclusion that these temporal changes in the estimated coefficients reflect genuine shifts in the underlying parameters. One such argument is that the observed changes simply reflect ‘errors in variables’. But there is no reason to believe that the reliabilities of identically measured independent variables differed from one year to the next. Sampling error is another technical reason for the changes in the coefficients over time. But the size of the observed differences and the monotonicity of the trends between 1980 and 1984 make this explanation implausible.
29 For a discussion of how to estimate ‘contributive’ importance, see Miller, W. E. and Shanks, J. M., ‘Policy Directions and Presidential Leadership: Alternative Interpretations of the 1980 Presidential Election’, British Journal of Political Science, XII (1982), 299–356CrossRefGoogle Scholar, and Achen, C., Interpreting and Using Regression (Beverly Hills, Calif.: Sage, 1982)CrossRefGoogle Scholar. A complication arises because both sampling error and change in the causal structure may cause this parameter estimate to differ from one year to the next. We use the average value of the two unstandardized coefficients as the multiplier on the grounds that in the absence of additional information about the behaviour of the slopes during the unobserved intervals, the most reasonable procedure is to assume monotonic change where change occurred and to use the mean value where estimates differ due to sampling fluctuations.
30 Our estimation technique may understate the contribution to trust made by the perceived improvement in the nation's economic health. If the decline in coefficients between 1982 and 1984 occurred monotonically during that period, then the earlier the improvement in the economy, the more likely it is that the ‘effect’ on trust at the moment that feelings about the economy changed was greater than the average 1982–84 slope. Data from the 1983 NES Pilot Study conducted in July of that year suggests that perceptions about the nation's economy improved drastically during the first six months of 1983, achieving the same mean value later registered in 1984. An explanation for the diminished effect of National Economic Health in 1984, therefore, is that once a problem becomes less pressing, attitudes about it are less relevant to one's generalized image of government.
31 We are handicapped here, as in the measurement of policy preferences, by the absence of equivalent questions in the three NES Surveys.
33 Wildavsky, A., ‘The Three Cultures: Explaining Anomalies in the American Welfare State’, The Public Interest, No. 69 (1982), 45–58.Google Scholar
34 This controversial point is made by Lichter, S. R. and Rothman, S., ‘Media and Business Elites’, Public Opinion, IV (10–11 1981), 42–6, 59–60Google Scholar and Weaver, P., ‘Newspaper News and Television News’, in Cater, D. and Adler, R., eds, Television as a Social Force (New York: Praeger, 1975), pp. 91–4Google Scholar, among many others. For an opposing point of view, see Hallin, D., ‘The Media, the War in Vietnam and Political Support: A Critique of the Thesis of an Oppositional Media’, Journal of Politics, XLVI (1984), 2–25CrossRefGoogle Scholar
35 Corrected for reliability, the correlation between Presidential Job Approval and Trust in Government was a healthy 0·46 in 1984. The reliability of the Trust in Government Index (0·517) was calculated from the 1972 National Election Study pre-/post-election panel. The reliability of Presidential Job Approval (0·799) was estimated from the pre-/post-waves of the 1984 NES. The formula used to correct the observed correlation (0·296) may be found in Achen, C. H., ‘Mass Political Attitudes and the Survey Response’, American Political Science Review, LXIX (1975), 1218–31, p. 1224CrossRefGoogle Scholar