In 1985, the average American reported having three close friends. By 2021, that number had dropped to fewer than two — and a growing share of adults reported having none at all. This is not a story about introversion or individual preference. It is a structural collapse measured across decades, replicated in multiple national surveys, and showing no signs of reversal.

The data comes from the General Social Survey, one of the most rigorous long-term social research instruments in the United States. When sociologists McPherson, Smith-Lovin, and Brashears published their 2006 analysis of GSS data, they found that the modal number of confidants — people with whom Americans discuss important matters — had dropped from three to zero. Not one. Zero. The most common answer had become "nobody."

That finding was controversial, partly methodologically contested, and widely misreported. But the directional trend was not seriously disputed. Follow-up surveys from the American Enterprise Institute, the Survey Center on American Life, and Pew Research all confirmed the same trajectory: Americans are maintaining fewer close friendships, spending less time with friends, and reporting higher rates of loneliness than at any point in the modern survey record.

Several structural forces drove this decline in parallel. Suburbanization and car-dependent design eliminated the incidental contact that neighborhoods once produced. Rising work hours — particularly in the professional class — consumed the time that friendship requires. Delayed marriage and extended geographic mobility disrupted the anchored social contexts that had historically held friend groups together. Digital communication created a simulacrum of connection that absorbed social energy without generating the density of mutual disclosure that close friendship requires.

The collapse is not evenly distributed. It tracks class, geography, and gender. Rural and working-class Americans lost friendships faster than urban professionals. Men lost more than women. The young, counterintuitively, are now lonelier than the old — reversing a pattern that held for most of the 20th century.

What makes this a collective phenomenon rather than a personal one is its synchrony. When millions of people independently arrive at the same outcome — fewer close friends, more surface-level contact, growing reliance on digital substitutes — the explanation is not individual psychology. It is systemic. The built environment changed. The labor market changed. The social institutions that once scaffolded friendship formation — churches, civic associations, stable neighborhoods, shared workplaces — weakened or dissolved. Friendship, which requires time, proximity, and repeated unstructured contact, was left to survive on whatever individuals could carve out from a structure that no longer supported it.

This document traces that collapse: what the data shows, what drove it, and what the distributional pattern reveals about whose friendships the system was willing to sustain.