Attitudes about Government and Other Institutions

The Task Force interprets shifting attitudes about government trustworthiness and responsiveness to mean that “Americans are increasingly worried about disparities of participation, voice, and government responsiveness.” Americans are said to believe that government is in thrall to well-to-do “special interests” (2004, 5).

Changes in Americans' trust and confidence in government, and feelings that government is attentive and responsive to public opinion are a thicket, threatening to entangle the unwary. Alford (2001, 46) noted that “we have not yet adequately explained large scale over-time variation in levels of trust” (see also Citrin and Luks 2001).²

Most of the decline in public trust in government occurred between the mid-1960s and 1980 (Alford 2001; Citrin and Luks 2001), when the Task Force claims that income inequality was low (2004, 3).

Even more troublesome for the Task Force's interpretation of growing cynicism are findings that income had virtually no impact on trust in government or confidence in government leaders (Alford 2001, 36–38). The 2000 and 2004 ANES showed virtually no relationship between family income and a four-item Trust in Government Scale, and a very anemic association between income and a two-item indicator of external political efficacy.

Two other factors vex the Task Force's explanation for declining trust in government. First, Americans lost confidence in politicians and the leaders of “institutions of opposition”—such as the press, television, and organized labor—at the same time (Cook and Gronke 2005, 792). Second, Blendon and his associates' (1997) report of why Americans don't trust government casts even more pall on the Task Force's interpretation. According to Blendon et al., when asked why they distrusted the government in 1995, “the main reason Americans gave revolved around the way the federal government spends money” (1997, 210). Inefficiency, or “spending on the wrong things” exceeded “special interests too influential” as reasons why people distrusted the federal government (see Blendon et al. 1997, Figure 8.4, 210).

Political Participation and Equality

The heart of the Task Force's report (2004, 5–11) deals with several facets of political participation. There are several problems with this section, including the premise that activism matters, which may not be true (see Weissberg 2005).

As have others, Ginsberg (1986, 48–50) warned that, unlike spontaneous riots, voting and other forms of election-related activity tend to be top-down mechanisms for elite rule, and not a way for common folk to demand additional government services. As Ginsberg put it (1986, 48–49), “With the advent of the election, control over the agenda for the expression of political opinion passes from the citizen to the state” (see also Weissberg 2005).

The Task Force notes that the wealthy turn out on Election Day and perform more demanding political activities at higher rates than do the poor (2004, 5–8). The report also argues that since the well-to-do participate at higher rates than those from lower income groups, these “Disparities in participation ensure that ordinary Americans speak in a whisper while the most advantaged roar” (2004, 11). “Unequal voice[—created by class “disparities in participation”—] matters because the advantaged convey very different messages to government officials than do average citizens or those who are the least well off” (11). Apparently George Soros' policy attitudes are closer to Richard Mellon Scaiffe's than they are to Cindy Sheehan's.

Scholars have long known that the wealthy participate at higher rates than the middle- and working-classes. A more important question asks whether non-voting contributes to skewed public policies. A number of studies have not found large differences between the presidential preferences and policy attitudes of voters and nonvoters (Leighley 1995 summarizes this evidence).

The 2000 and 2004 ANES reveal no statistically significant differences between voters' and non-voters' attitudes about gun control, the death penalty, protecting homosexuals against job discrimination, government assistance to Blacks, and government spending for defense.³

All questions were identical.⁴ Voters' and non-voters' attitudes about abortion and the welfare state⁵ were significantly different, although none of the Kendall's tau-c values was higher than .16, and the rest were lower than .10. Non-voting might skew the “voices” heard by public officials concerning abortion and the welfare state, but not for other issues. Even in the cases of abortion and the welfare state, any policy skew would be slight.

Consider just one other type of participation: donating to campaigns. According to the Task Force, “Campaign contributors are the least representative group of citizens,” (2004, 7), because the wealthy are much more likely to donate to campaigns than are the poor and middle class. Although the relationship between family income and campaign contributing was statistically significant in 2000 and 2004, the relationships were weak. In 2000, the point-biserial correlation coefficient for the relationship between family income and donating to a party or candidate was r_pb = .20. Four years later, the same relationship was r_pb = .18.

There is no gainsaying the fact that wealthy persons contribute disproportionately to American campaigns, probably for the reason noted by Weissberg (2004–2005, 34). Someone making $20,000 a year might find better ways to spend her money than making a small contribution to a campaign, while multi-millionaires would not miss seven-figure donations.

The report also notes that, even though many groups representing the “previously unorganized” have been created lately, “Interest Groups for the Well-Off Dominate Washington” (2004, 8–9), which may be unassailable.⁶

But, is this new? In 1960, E. E. Schattschneider wrote that “the pressure system has an upper-class bias” (32). This was just before the U.S. sought greater equality for Blacks and Lyndon Baines Johnson launched “the Great Society.”

The Task Force contends that political parties, once the vehicles for mobilizing masses of ordinary citizens, have now become agents of what Schier (2000, 1), whose work is cited in the report, calls the politics of “activation,” or “identifying and activating the small segment of citizens most likely to ‘get their message’ and vote or lobby government.” What is worse, from the Task Force's point of view, the two major parties “intensify the skewed participation in U.S. politics by targeting many of the resources on recruiting those who are already the most privileged and involved” (2004, 10).

The Task Force's focus on parties' penchant for activating mainly their well-to-do adherents occasions three comments. First, as Schier noted (2000), the strategy of activation is a rational response by the parties to their weakened condition. At that, however, the link between family income and reports of being contacted by one of the parties is weak. In 2000, the point-biserial correlation coefficient for the relationship between family income and respondents' reports of being contacted by one of the parties during the campaign was r_pb = .12. Four years later, the same coefficient was r_pb = .20.

Finally, the parties may have become very dependent on wealthy contributors, but they remain far apart on issue after issue. Some wealthy persons are on the “left” side of the political spectrum, others are on the “right” side.

Income and Public Opinion

The Task Force evidently assumes that persons from society's lower social orders have fundamentally different, and conflicting, interests with those from the higher rungs of the economic ladder.⁷

The Task Force seems to assume that the type of “class conflict” that may have existed during the 1930s and 1940s still exists, or at least should.

The 2000 and 2004 ANES do not suggest that the poorest and wealthiest segments of society are at loggerheads on key policies: gay people's rights, the death penalty, abortion, government assistance to African Americans, defense spending, and the welfare state. (These are the same policies considered in the section above.) Family income is broken into quintiles.

It would be mistaken to assert that family income is orthogonal to the attitudes depicted in the table but the relationships are generally quite small. Indeed, as estimated by chi-square (χ²), several are not significant at ρ = .05.

The exceptions are for the relationships between family income and attitudes about abortion and toward the welfare state. Table 1 displays the relationships shown by χ² to be statistically significant. For abortion, the table displays each income quintile's attitudes about the issue. For the Welfare State Scale, the table depicts each income quintile's mean score and standard deviation. Even when the χ² values indicate statistical significance, the values of Kendall's tau-c are typically quite modest. Even the association between income and attitudes about the welfare state was not strong. In 2000, the value of Kendall's tau-c was .16; four years later it was .21.

It should also be noted that the correlations between income and partisanship and ideology are generally weak. In 2000, for example, the value of Kendall's tau-c for the relationship between family income and ideology was .05. The same statistic for the association between family income and party identification was .11. Four years later, the value of tau-c for the relationship between family income and ideology was .07, while the tau-c for the relationship between family income and partisanship was .12. Controlling for race reduced the strength of these relationships.

Finally, the 2000 and 2004 ANES do not show strong correlations between income and presidential voting. Looking at just George W. Bush vs. Al Gore in 2000 and Bush vs. John Kerry in 2004, the relationships were statistically significant, but the value of Kendall's tau-b was .11 in 2000 and .16 four years later. Moreover, controlling for race weakened the relationships.