Social & Cultural Institutions

Religion: a long, asymmetric collapse

Confidence in organized religion is the second-steepest declining institution covered in Part II of this book — after the press, and before banks. From the early-1970s peak above 45%, it has fallen by two-thirds.

What makes the religion series unusual is that the decline has not run on partisan lines. It has run on attendance lines — and the attendance gap, in absolute terms, has narrowed.

Two readings of Figure 2 point in slightly different directions. On a relative basis, weekly church attenders have held confidence in the church somewhat better than the secular have: a 36% relative decline (44 → 28) versus a 62% relative decline (24 → 9). On an absolute basis, what is striking is that weekly attendance has not protected the institution from confidence loss. Believers themselves are 16 percentage points less confident in the church than they were in 1973. The 2002 Boston Globe Spotlight investigation, the cascade of clerical-abuse stories that followed in subsequent years, the ongoing decline of mainline Protestantism, and the slow politicization of evangelical Christianity all show up somewhere on this series — but the GSS does not tell us which. It tells us only that the result, summed across causes, is a roughly two-thirds decline in confidence over half a century, regardless of the respondent's relationship to the institution.

The party gap on religion is small but recently widening on a particular dimension.

The Democratic line on church confidence is at 11% in 2024 — the lowest reading from any major partisan group on any institution in this chapter. The Republican line is at 21%, also a historical low for that party. What makes the partisan gap relatively more visible now is just that the topline floor is so close to zero: a 10-point gap matters more when the average reading is 15% than it did when the average reading was 35%.

Generationally, the church story is a slow handoff. In 1973 the 65+ cohort was the most church-confident at 50%; in 2024 the 65+ cohort is the most church-confident at 21%. The gap between age cohorts has narrowed, but the order is the same. Religion is one of the few institutions where younger Americans have not "diverged" from older ones — they have moved together, in the same direction, at roughly the same rate.

Military: the Gulf War peak, the Trump-era reversal

The military series has a completely different shape, and it carries a finding that, more than any other in this chapter, would have been hard to predict five years ago.

The military's long arc is well-known to historians of public opinion. After Vietnam, in the late 1970s, fewer than one in three Americans expressed high confidence in the military. The Reagan defense buildup, the Iran hostage crisis, and the cumulative effect of professionalization (the post-draft volunteer force) brought the number back up through the 1980s. The 1991 Gulf War took it to 63%, the highest reading ever recorded for any institution covered in Part II. The post-9/11 War on Terror consolidated the level at 50%+ for nearly two decades.

What was not predicted, and what Figure 6 documents, is what the partisan structure of military trust did between 2018 and 2024.

This is a remarkable fact. For fifty years, "Republicans trust the military more than Democrats do" was one of the most reliable patterns in the entire institutional-confidence battery. As recently as 2018, three-quarters of Republicans expressed high confidence in the military. By 2024 the figure had collapsed by 30 points. The Conservative line by ideology shows a similar 29-point fall.

What happened? The data themselves cannot answer that, but the period contains a number of plausible drivers: the late-Trump and Biden-era civil-military controversies (the ouster of Mark Milley, the politicization of military leadership decisions); the August 2021 Afghanistan withdrawal; the Pentagon's vaccine mandate; recruitment shortfalls and the public conversation around them. Several of these are events that conservatives have particular reasons to be unhappy with. The data make clear only that something in this period unmade the institutional priors of the modern Republican coalition with respect to its own armed forces.

The age structure of the recent military decline is also worth noting briefly: the 2018 → 2024 fall is sharpest among older cohorts, who had been the highest-confidence group on military matters throughout the series. The 50–64 cohort fell from 70% to 46% (24 points), and the 65+ cohort fell from 63% to 47% (15 points). Younger cohorts (18–34) have actually been more confidence-stable. Whatever happened in this period landed disproportionately on the cohorts whose adulthood was structured by the post-Vietnam recovery and the Reagan-era buildup — not on the cohorts who came of age during Iraq and Afghanistan.

Two institutions, one larger pattern

The pairing of religion and the military in this section was not arbitrary. Both are institutions whose authority comes from belief rather than performance. Both have, over the period of this book, become objects of reduced confidence on the American Right specifically. Religion's loss is older and steadier. The military's loss is sharper and more recent. But on both, the largest absolute losses in the most recent six years are inside the Republican and Conservative coalitions — the very coalitions for which "trust the church" and "trust the troops" had been near-axiomatic.

This is the third domain in this chapter where the recent partisan story is not "Republicans diverging from Democrats" but "the Republican coalition's relationship to a longstanding institutional commitment is changing." Major companies (Ch. 2.5) showed it on business. The GSS's institutional battery, taken as a whole, suggests that something has been happening to the structure of conservative institutional trust that is broader than any single political event. Whether the cause is Trump-era civic disruption, generational replacement, the cumulative effect of culture-war controversies, or all three, the data alone cannot say. They can say, very clearly, that the fact of the change is real.

The next chapter switches register entirely. Part III turns from what Americans say about institutions to what they say about each other — beginning with the most foundational measure of social trust the GSS asks about: whether people, in general, can be trusted.