Americans Feel Fine About Their Own Lives - But Trust in Institutions Has Collapsed Over 50 Years

There is a persistent puzzle in American public opinion research. On measures of personal welfare - economic satisfaction, health, happiness - survey responses have been relatively stable over decades. People report that their own lives are roughly as they always were. But ask about the country, its institutions, or whether the government listens to people like them, and the data tell a different story entirely.

James N. Druckman and colleagues analyzed long-term data from two of the most rigorous and sustained survey infrastructure projects in American social science: the General Social Survey (GSS) and the American National Election Studies (ANES), both supported by the National Science Foundation. Their analysis, published in PNAS, examined trends across seven dimensions from the 1970s through 2024: economic satisfaction, health, happiness, satisfaction with democracy, affective polarization, political efficacy, and confidence in institutions.

Personal measures vs. national measures

The divergence between individual and national assessments is one of the analysis's most striking features. Personal economic satisfaction, self-reported health, and happiness have all remained broadly stable over the decades covered, though happiness showed some decline after 2020. These are not indicators of a population in acute distress.

The national measures diverge sharply. Satisfaction with democracy dropped significantly between 2008 and 2012 and has not recovered. Affective polarization - measured as the gap between how favorably people view their own political party versus the opposing party - has increased by roughly 30 percentage points since 2000. This means people have not simply become more positive about their own side; they have become substantially more negative about the other side. The gap between in-group warmth and out-group hostility is much wider than it was a generation ago.

Political efficacy - the sense that ordinary citizens have influence over what government does and that public officials care about people like them - has declined substantially since the 1990s. The combination of lower efficacy and higher out-party hostility constitutes a portrait of democratic alienation that personal well-being numbers do not capture.

Institutional confidence has fallen broadly

Confidence in specific institutions has declined across a wide range of sectors. Congress, education, medicine, organized religion, the press, and science have all seen notable drops in the proportion of Americans expressing high confidence. The declines are not uniform - some sectors have fallen further than others - but the breadth of the pattern is significant. This is not skepticism about one particular institution but a general erosion of institutional credibility.

The partisan dimension of institutional confidence has widened considerably. Fifty years ago, Democrats and Republicans differed primarily in their confidence in labor and in business - a partisan gap that reflected traditional economic coalitions. Today, the partisan gap extends across almost every major institution. Democrats report higher confidence in education, science, the press, and medicine. Republicans report higher confidence in the military, organized religion, and the Supreme Court.

The sorting of institutional trust along partisan lines is a different phenomenon from overall decline. It means that confidence in institutions has become less about the institutions themselves and more about which team they are perceived to be on. This creates self-reinforcing dynamics: as institutions become coded as partisan, partisans on the other side withdraw confidence, which further confirms the partisan coding.

What long-term data enables

The value of the GSS and ANES data lies precisely in their longevity and consistency. Both surveys have used comparable methodologies over decades, enabling the kind of trend analysis that shorter or more recent studies cannot provide. The researchers emphasize that the sustained investment in this infrastructure - face-to-face interviewing, consistent sampling, standardized questions - is itself worth highlighting at a moment when such investments face funding pressure.

Cross-sectional data has limitations that longitudinal survey data helps address. A poll taken in any given year captures a snapshot that may be driven by temporary events. The patterns visible in the GSS and ANES span administrations, economic cycles, and news events, making them more robust indicators of structural trends than any single study could produce.

The authors argue that these trends paint a picture of a nation under civic stress - not primarily in terms of how individuals experience their own lives, but in how they relate to the collective institutions and political processes that govern public life. Whether that stress is reversible, and through what mechanisms, the survey data alone cannot answer.