The Brutal Pervasiveness of Solitary Confinement

Editor’s Note: This article previously appeared in a different format as part of The Atlantic’s Notes section, retired in 2021.

Yesterday, I wrote about California’s landmark legal settlement to sharply curtail solitary confinement in the state’s prison system. The common estimate I used in that article suggested that between 70,000 and 80,000 prisoners are currently in solitary throughout the U.S. We just learned the actual total could be much larger.

A new report by the Liman Project at Yale Law School and the Association of State Correctional Administrators sought to reassess the scope of solitary. Researchers sent a survey with more than 130 questions to every state and federal prison. The foremost question was size: How many prisoners do you keep in some form of isolation?

Commentators have relied on estimates dating back ten years or more; the figures cited range from 25,000 to 80,000 prisoners. This Report is the first to update those figures; thirty-four jurisdictions, housing about 73% of the 1.5 million people incarcerated in U.S. prisons, provided numbers, totaling more than 66,000 prisoners in some form of restricted housing—whether termed “administrative segregation,” “disciplinary segregation,” or “protective custody.”

66,000 prisoners is below the range I cited yesterday, but that figure only includes data from 34 states that responded to the survey. By extrapolating those figures to the entire country, the report estimates that between 80,000 and 100,000 prisoners were in some form of physical segregation from other inmates in 2014—a substantial increase from previous estimates.

A troubling caveat in the report also notes that its survey did not include “people in local jails, juvenile facilities, or in military and immigration detention.” The true number of prisoners in some form of solitary may be significantly larger than today’s estimates.


Professionals who study mass incarceration often struggle to overcome gaps in criminal-justice statistics. My early foray into these numbers came through the death penalty, which is relatively well-quantified compared to the rest of the system. I grew curious about the problem earlier this year and first wrote about it in May. The challenges I ran into were philosophical ones: How do we know what we don’t know? How do these knowns and unknowns, to borrow a phrase, shape our understanding of the criminal-justice system—and its reform?

I’m still looking for answers. If you have any expertise on criminal-justice statistics or the gaps between them, feel free to reach out to me at