The Foundations of Trust

It is no secret that public trust in politicians and the political system has declined markedly in recent decades. According to national surveys, many Americans no longer trust elected officials to keep their promises or to look out for interests of ordinary citizens. The proportion of those who say they have "hardly any" confidence in the people running the executive branch of the federal government rose from 18 percent in 1973 to 35 percent in 1994. Over the same time period, the number expressing low confidence in leaders of Congress increased from 15 percent to 40 percent.

But Americans are not only losing faith in politicians and other leaders. They are also less willing now than in the past to believe that people in general can be trusted. For example, the proportion agreeing that "most people can be trusted" fell from 55 percent in 1960 to 34 percent in 1994, while the proportion saying "you can't be too careful in dealing with people" rose from 40 percent in 1960 to 61 percent in 1994. Taken at face value, these surveys point to growing fears -- about crime, hatred, and selfishness -- and to rising cynicism about human nature. Viewed more closely, they suggest that Americans are losing the capacity to work with one another, to cooperate and to give each other the benefit of the doubt.

Of course, questions can be raised about these findings. It may be that the surveys are biased, eliciting responses that oversimplify people's own understandings of trust. It may be the case, too, that people have good reasons for exercising caution in their dealings with others. Still, if trust is declining in the United States, then one of the ways to renew civil society may be to understand better the conditions that facilitate trust and the ones that erode it. In what follows, I consider both the relevant survey data and the results of interviews in which people spoke in detail about their civic activities and beliefs, including some of their views about trust.

What Surveys Show

Though newspaper articles about trust tend to report aggregate measures, the evidence indicates that some groups in our society are more trusting than others. For example, Robert Putnam finds higher levels of trust among people with higher levels of education. Whites are more likely than blacks to give trusting responses, and people who hold memberships in voluntary associations are more trusting than non-members.

Eric Uslaner's analysis of the data suggests that the standard survey question about trust can almost be thought of as a measure of optimism: people who have more advantages in life and who are more confident about their futures have more reasons to be optimistic and thus are more likely to give trusting responses. By implication, trust may be declining because many Americans do not feel as confident about the future in terms of economic opportunities as they did a few decades ago.

Two other interpretations of trends in levels of trust also find support in the available data. First, some of the erosion in social trust is attributable to political scandals specifically, as well as to declining confidence in public leaders more generally. For example, National Election Survey questions about trust in people show abrupt declines during the late 1960s and early 1970s, when the Vietnam war and Watergate were raising public concern about national leaders. Other surveys conducted in those years show that large numbers actually believed their trust had been shaken by Watergate. The National Election data also show that some of the decline in generalized trust can be explained statistically by the decline in confidence in government officials.

The other interpretation of trends in generalized trust comes from looking more closely at its relationship with education. In 1973, the largest differences in trust were between respondents who had graduated from high school and those who had not. By 1994, this difference had shrunk, but that between college graduates and high school graduates had increased. High school graduates in 1994 were about as likely to be trusting as those without high school degrees had been in 1973, while college graduates in 1994 resembled high school graduates in 1973. In addition, the 1994 data show that college graduates were significantly less likely to trust than were respondents with graduate degrees, whereas in 1973 this difference was not significant.

These results confirm what other studies of college effects have long suggested: Higher education appears to be a measure of relative standing in the society, more so than a cognitive experience that shapes people's attitudes absolutely. As the proportion of Americans with college and postgraduate degrees has expanded, the large number who have high school degrees or less have become more disadvantaged, and their declining sense of their own opportunities appears to be reflected in a lower likelihood of giving optimistic responses to questions about trust. This is not to deny that even the most privileged in terms of education have also become somewhat less trusting. But it does appear that trust has fallen partly because people with the lowest levels of education have become more similar to one another in their views about trust, while there are greater differentials between these people and those with higher levels of education.

A second data set that permits further analysis of trust comes from a 1982 survey which asked adults in a representative national sample how they perceived themselves. One of the questions asked people to rate themselves on a seven-point scale anchored at one end by the word "dependable" and at the other by the word "undependable." This question is conceptually interesting because, as I shall suggest below, many people's willingness to trust others seems to be a reflection of whether or not they believe they can trust themselves.

The variables that are significantly associated with describing oneself as dependable are essentially the same as those that in other surveys correlate with saying that most people can be trusted. Having a college degree, being white, having an above-average family income, owning one's home, and scoring higher on a self-esteem scale are all associated with higher probabilities of regarding oneself as dependable (as is civic participation, as measured by volunteering to help the poor and writing letters to public officials). I take these findings as added evidence that trust is influenced by personal circumstances. If generalized trust has been declining, the reason may not be simply that people believe others are having more difficulty living up to their agreements, or even because their faith in politicians has been shaken. It has probably declined partly because people are less confident in their own capacities to hold up their share of the bargain. We shall have more opportunity to consider this possibility in looking at the qualitative data.

The final survey from which I have been able to draw observations on trust is my own 1992 national survey of 1,000 members of small support groups, from Bible study circles to twelve-step and self-help groups. This survey, unlike those that measure trust in unspecified "people," asked respondents about trust in a specific context, focusing on the trust that respondents place in fellow members.

The results show that the intimate and encompassing interaction characteristic of support groups -- in which people may eat together, sing, pray, and help others in their community -- is associated with higher levels of trust. Other factors that did not show significant relationships with trust were engaging in sports activities together (contrary to arguments about the beneficial effects of bowling leagues), discussing social or political issues, having elected officers, having an agenda or schedule, and having business meetings. In short, informal interaction was conducive to trust, but formal structure was unrelated to trust.

The study also suggests a reason why trust may be declining in the wider society. People in heterogeneous groups -- especially ones that included diverse political views and mixed religious views -- were less likely to rate their groups high on trust than were people in homogeneous groups. From this finding, it may be reasonable to suspect that trust among the public at large has suffered as a result of acrimony between political or religious factions; it may also have been influenced by the growing diversity of the population in ethnic identities, lifestyles, and values.

In sum, the various surveys show that opportunities to realize one's expectations are conducive to trust, whereas inequality and disadvantage are not. Perceptions of one's own reliability are an important consideration, and social interaction appears to enhance trust. The data also indicate that trust is always conditional: Nobody trusts everyone all the time. It is thus important to turn to qualitative data to see how people understand the conditions under which they can and cannot trust others. The goal is to identify the different kinds of mental frames that people bring to bear in deciding when and whether to trust.

Trust from Within

For some people, as the surveys suggest, trust in others depends on a prior belief in their own trustworthiness. As one man put it, "I feel if I can be trusted, I can trust other people. Throughout my lifetime I've found that to be true, that if you are up and above and honest with people, they will return that respect." His statement focuses more on himself than on any particular characteristic of others. It is also significant that he associates trustworthiness with being honest, rather than with being dependable -- making good on everything he promises. This allows him to see being trustworthy as entirely under his own control, rather than subject to contingencies that might prevent him from fulfilling other people's expectations.

Such seemingly minimal conditions can nevertheless become problematic. In other research, I have found that people are not at all sure (or in agreement) about what it means to be honest. For many, honesty is largely a subjective self-assessment that says, in effect, you are honest if you feel that you are honest. In the absence of clearer external standards, many people feel that they are operating largely within gray areas, especially in business or in their professional lives. As a result, they wonder whether or not they are truly honest.

Generally, respondents who emphasized certain characteristics of their own as the condition for trusting others relied on one of three frames. The first is illustrated by the man we have just been considering. It is aptly described as a frame that emphasizes self-knowledge -- believing, for example, that you are honest, civil, and capable of judging people. Others felt certain that they could keep their mouths shut when they needed to, or that they would not make a promise they could not keep.

A second frame emphasizes personal experience. For example, a 44-year-old newspaper editor attributed his trustfulness to his growing up in a neighborhood where people knew one another and helped one another. This man did not simply generalize from his childhood experience by assuming that virtually everyone he runs into in his present life can be trusted. Rather, he observed that his experiences had shaped him so that he is simply a trusting person.

The third frame explains trustfulness as a leap of faith on the part of the person doing the trusting. Unlike the first two, this frame implicitly assumes that trust may not be rational (at least not in terms of what one knows about how people behave), but argues that being trusting is still a good way to live. One woman who works as an artist told us that she trusts most people, even though she has been "burned" a few times and is not naive. She explained her reason for trusting in terms of a faith commitment that causes her to focus on an inner self: "What I'm seeing is what's going on inside that person, I behold the Christ in you, the good stuff that connects us all, that's in all of us. That's where I connect into people." But other respondents defended their leap of faith on pragmatic grounds; they said they were simply happier or found they could get along better if they decided to be trusting.

Objects of Trust

In contrast to those who made trust depend on something in themselves, most of the people we talked to focused on attributes and conditions of those who would or would not qualify as objects of trust. The most common of these conditions was also the most straightforward: people said they would trust others if those others had already proved to be reliable. In other words, trust depended on a rational assessment of someone's performance in the past and the likelihood that the future would hold similar performances by that person or by similar persons. This kind of trust, it should be noted, is highly contingent on the kind of people toward whom it is directed. Whereas a frame that emphasizes one's own trustworthiness suggests an inclination to trust most people under most circumstances, this one is more limited.

Another frame that permits people to say they can trust others emphasizes resources. Objects of trust who have sufficient resources to carry through on their promises and to behave reliably are said to be trustworthy; in contrast, people with limited resources should perhaps not be trusted. A 43-year-old woman who manages a small company illustrated this frame when she said that her employees can generally be trusted because her firm supplies them with the resources they need to fulfill their responsibilities. She generalized from this example, saying, "If you give people responsibility and try and give them as many tools as you possibly can, they won't disappoint you."

Significantly, this is a frame that may influence people's judgments of political leaders as well as workers or friends. One of our respondents, a draftsman in his fifties, had concluded that public officials were unreliable, and produced this explanation of their failure to perform up to the public's expectations: "I think they honestly believe that when they get in there they're going to be able to do all these things for you. My feeling is once they are elected and they're in the office, they learn the reality of those offices and they find out that things aren't as easily done as they think they are." Like many of the people we talked to, the draftsman denied that his qualified trust in politicians should be taken as incipient cynicism. He said he was merely being realistic in his expectations of what officials could and could not do.

The main alternatives to frames that emphasize reliable performance are frames that emphasize something about the self of the person being trusted. These frames come in several variants. One of the most common argues simply that you can trust people better if you know them well. It says nothing about whether or not these people have performed reliably in the past -- only that you know them well enough to assume that they are people of character, people who are somewhat like yourself in their values and their respect for common norms of decency.

Some respondents were more specific about what it is in other people that makes them seem trustworthy. One frame (which was hinted at by the woman who focuses on the "Christ" in other people) draws a distinction between an inner self that is presumed to be more trustworthy and an outer self that may be more calculating, devious, or unreliable. The people who can be trusted are those who have been willing to disclose something about this inner self.

This emphasis on self-disclosure means, perhaps obviously, that people are more likely to trust those with whom they have developed some intimacy. Yet this frame (like the one that emphasizes resources) sometimes figures in people's judgments of political leaders. For example, some of our respondents focus less on the official acts of politicians than on biographical details, body language, reports of affairs, and the use or absence of certain words. They are very much aware of their leaders' personal foibles, and base their opinions on characteristics that may have little to do with performance in office.

Alternatives to Trust

Though the people in our study often expressed concern about mistrust, many of them had found ways to negotiate their lives without being immobilized by it. They did so by relying on other mechanisms, such as laws, regulations, and markets, to minimize risk. For example, the head of a baby-sitting cooperative in New Jersey told us that although her organization had 200 members, only 8 families regularly used it. The reason, she thought, was that most of the members did not trust the other members well enough to feel comfortable leaving their children with them. Although she tried to build trust by having parties and other "get acquainted" meetings, most of the members said they didn't have time for such activities. Because they enjoyed upper-middle-class incomes, they used for-profit day care centers instead.

In a case like this, one could say that trust was being redirected from neighbors to for-profit day care centers. But this is not how the woman herself saw it. In her view, trust is "a personal thing." Like many of the people we interviewed, she makes it conditional on knowing something about the private life -- the inner self -- of another person. It seems better to her to say that people have confidence in day care centers. Trust is not at issue; confidence rests on knowing that these centers are licensed by the state, are governed by a number of laws and regulations, and must do their jobs well if they are to stay in business.

Political mistrust can elicit a similar response: Americans mistrust individual politicians but find other ways to get on with their lives. Some say it is healthy to be skeptical toward politicians, because critical questions get raised or because mistrust prevents politicians from assuming too much power. For many of the people we interviewed, mistrust was more of an irritation than a profound complaint because they felt the political system should not be doing very much anyway. Some believed that private volunteer organizations were better suited to solving social problems. Following the 1996 election, during which high levels of mistrust toward the leading candidates were registered, the stock market soared to record highs; analysts attributed the increases to investors who saw that a divided government incapable of inspiring confidence would do little to upset their expectations.

Implications for Civic Renewal

The foregoing nevertheless suggests that trust is important enough that people are prepared to talk about it, generally in more complex ways than have been captured in public opinion polls. For most people, trust is not simply a matter of making rational calculations about the possibility of benefiting by cooperating with someone else. Social scientists who reduce the study of trust to questions about rational choice, and who argue that it has nothing to do with moral discourse, miss this point. In interviews, people bring moral frameworks directly to bear on their thinking about trust, talking about how it is simply good to trust others, how trust depends on moral character, and how you have a moral obligation to fulfill the expectations of those who have placed trust in you. Even if trust is taken as nothing more than an expectation of how someone will behave, it is invariably conditioned by assumptions about one's own honesty, whether or not promises are morally binding, and how much the behavior of individuals is likely to be shaped by their moral commitments.

Our surveys suggest that deliberate repair work needs to be done to restore trust that has been damaged by broken promises, disrupted relationships, and fragmented self-images. Clearly, it is not enough to argue that trust will be restored simply by getting people to join bowling leagues and other civic organizations. Many of the respondents who are active in such organizations said that they had to work hard at building and maintaining trust. Rather than simply happening as members interacted, it required confrontations, staff meetings, bull sessions, phone calls, and mediation.

People also talked about constructing rules within civic organizations that served as heuristics for making decisions about trust. These rules were often quite explicit: Members did not have to rely on implicit norms about whether or not to trust someone who failed to show up for meetings or who took handouts but did not look for a job. Because these are gray areas, organizations created local understandings about how to regard such behavior. Consequently, participants did not have to trust a "generalized other," but could respond according to the rules they had learned in these organizations. The baby-sitting co-op provided detailed information to members about what to expect and what not to expect; garden club members developed a voucher system to avoid confusion about who owed favors to whom; a homeless shelter adopted a rule that in effect told volunteers to trust whatever clients told them the first time, but to ask more questions on subsequent visits. In these ways, the presence of civic organizations helped to define the conditions under which trust could be exercised.

The frames that privileged Americans use to justify their faith in people suggest a final lesson. Given that trust of others and trust of oneself depend so heavily on the resources and opportunities at one's disposal, the problem of mistrust can only be addressed adequately by including efforts to redress injustice and inequality. It is not enough to blame negative campaign rhetoric or the nihilism of popular culture. African Americans are not less trusting than white European Americans because they watch the wrong kind of television. They are less trusting because they have fewer economic resources to risk and are more in danger of being victims of violent crime. Those who generalize from their experiences in safe, affluent suburban families may not always recognize the importance of social circumstances in promoting a willingness to trust.

-- Robert Wuthnow

Robert Wuthnow, professor of social science at Princeton University, is the author, most recently, of Poor Richard's Principle: Recovering the American Dream through the Moral Dimension of Work, Business, and Money (Princeton University Press, 1996). Sources: General Social Surveys and other national surveys conducted by the National Opinion Research Center at the University of Chicago; Robert D. Putnam, "Bowling Alone: America's Declining Social Capital," Journal of Democracy (January 1995); Putnam, "Tuning In, Tuning Out: The Strange Disappearance of Social Capital in America" (The Ithiel de Sola Pool Lecture, American Political Science Association, 1995), P.S.: Political Science and Politics, vol. 27 (December 1995); Putnam, "The Strange Disappearance of Civic America," American Prospect (Winter 1996) [http://epn.org/prospect/24/24putn.html]); Eric M. Uslaner, "Faith, Hope, and Charity: Social Capital, Trust, and Collective Action," unpublished paper, Department of Government and Politics, University of Maryland (1996); Self-Esteem Survey, Gallup Organization (1982); Robert Wuthnow, Sharing the Journey: Support Groups and America's New Quest for Community (Free Press, 1994); Erving Goffman, Frame Analysis: An Essay on the Organization of Experience (Harper & Row, 1974); Robert Wuthnow, God and Mammon in America (Free Press, 1994); Adam Seligman, The Problem of Trust (Princeton University Press, 1997); Russell Hardin, "The Street-Level Epistemology of Trust," Politics and Society, vol. 21 (December 1993).

[ Table of Contents ]