PRINCETON, NJ -- One of the more interesting phenomena we pollsters run across is the local positivity bias. Basically, here's what we observe: When survey respondents are asked to rate something at the national level, they almost universally rate it more negatively than they do when asked to rate the same thing at their local level.
In other words, Americans tend to say that things are bad "out there" across the country, but that things are much better where they live.
Two of the most prominent examples of this local positivity bias relate to Congress and education. When asked whether "most members of Congress" should be re-elected, Americans are always more likely to say "no" than they are when asked whether their "representative to Congress" should be re-elected. Americans think that members of Congress in general need to be turned out of office, but that their local representative is okay. In similar fashion, the "nation's schools" always receive a much lower grade than respondents' own local schools do. Indeed, when we narrow a sample down to parents of kids who are actually in the local schools, the educational system receives its highest marks.
The same positivity bias occurs when Americans are asked to rate how satisfied they are personally, as opposed to how satisfied they are with the way things are going "in the country today"; when they are asked to rate health care (the nation's health care system is rated negatively, while at the same time individuals' own health care gets good marks); and when they rate crime.
Logically, of course, this shouldn't happen. The overall state of the nation's educational and health care systems should be nothing more than the sum of all of the individual educational and health care systems. If everyone is happy with his or her local schools, for example, then the overall state of the educational system in this country should be positive. If everyone likes his or her own member of Congress, then why do these same people rate Congress as a whole so poorly?
One explanation for the local positivity bias revolves around a classic social psychological theory: cognitive dissonance. Humans strive to reduce the unpleasant state of having inconsistent cognitions or thoughts. Rating one's schools, safety, and health care situation negatively can be inconsistent or dissonant with the human desire to think that one has made smart choices and that one's kids are getting a good education. Few parents want to think they are sending their child to a sub-par school. We tend not to want to focus on the consequences of believing that we have sub-par health care. Hence, to reduce cognitive dissonance, respondents may have a tendency to tell interviewers that indeed everything is in relatively good shape at their local level.
Additionally, humans engage in quite a bit of impression management and presentation of self in their day-to-day interactions. A survey interview is a dyadic interaction, and individual respondents may feel significant pressure to appear to have made wise choices in terms of where they live, the health care available, and so forth to the interviewer on the other end of the line.
Indeed, there is another social psychological theory, called self-perception theory, which says that in many ways we figure out how we feel about things by observing our own behavior, verbal or otherwise. We figure out that we are embarrassed, for example, when we observe our bodies blushing. We know we are frightened when we observe our knees knocking and adrenaline surging into our bloodstream. And as a classic example has it, we know that we like black bread because we observe ourselves eating it often. The extension of this theory is that, in response to an interviewer's questions, we have some pressure to say we like our current situation because doing so allows us to conclude -- when we observe our own verbal behavior -- that indeed our situation is not all that bad.
There is another way of looking at the local positivity bias phenomenon. Most Americans, except perhaps for itinerant salespeople or well-informed pollsters, have little firsthand knowledge of the true state of education, health care and crime across the nation. Thus, when asked by an interviewer to give an opinion, the respondent falls back on what he or she has heard. In many instances, that involves news coverage from the media. The news media, as they have from time immemorial, cover the exceptional and the out-of-the-ordinary, and that, more often than not, tends to be bad, negative news.
Additionally, national news coverage amplifies and underscores local problems. For example, the news media naturally report on the school districts that have problems, even if there are only 10 of them out of the thousands of such districts across the country. The average observer could be forgiven if he or she makes a leap of logic and assumes that these problems must be symptomatic of a large number of problems across the country.
There are also great numbers of special-interest groups, charities and other organizations that exist to focus attention on particular problems in order to raise money and consciousness to fix these problems. These groups thus have a vested interest in calling the nation's attention to problems. There are fewer groups that exist in order to tell people how well things are going in a particular sector. The net result may be to encourage Americans to assume that massive problems exist across the nation, and hence to respond in the negative when a pollster calls and asks for assessment of conditions nationwide.
There is certainly evidence that the population responds to national news coverage of problems. In 1993, the percentage of Americans saying that health care was the number-one problem in the country zoomed up -- no doubt in direct reaction to the fact that Bill and Hillary Clinton emphasized the critical condition of the nation's health care system as part of their ill-fated national health care plan. This was presumably not because the actual state of health care changed across the country in a short period of time, but because respondents were reacting to what they saw and heard in the media. Indeed, within the span of a few months, the percentage of Americans who said that health care was one of the country's most important problems dropped down to where it had been before the Clinton initiative.
Does all of this mean there is little value in asking Americans to rate the state of affairs "in the country today" or "in the nation today"?
Certainly, one can argue that asking respondents about the state of affairs in their local area produces answers that may be more accurate in the precise sense of the word. Respondents have actual experiential data and daily experiences on which to base their evaluations of what goes on around them, and thus these local measures may well be more likely to reflect reality as opposed to "perceptions of reality."
But this certainly doesn't mean that national evaluations are worthless. As social psychologist W.I. Thomas once said, "If men define situations as real, they are real in their consequences." Even if distorted, perceptions of the country as a whole can be important indicators of the overall mood of Americans, and as such, can have a specific impact on elections and in turn, national legislation.
In other words, if the public is convinced that there is a health care crisis in this country, then it may not matter as much whether there are indeed health care problems in reality. People's perceptions make them more receptive to legislation to change health care, and ultimately could affect election results.
Thus, at Gallup, we are committed to the idea of continuing to measure all aspects of public opinion on key issues -- Americans' perspectives on the state of affairs at the national level, as well as their perceptions of what is happening in their local area and in their local community. The key, in our opinion, is the careful analysis and treatment of what it all means. Blended together, the assessment of what Americans feel about social issues at both levels gives us a better and more detailed portrait than would result from either of these two levels in isolation.