Lynchings

How many lynchings are documented, and where?

CrimeDB compiles one record of lynching victims from two research inventories. Tolnay and Beck's covers the ten states of the Deep South, 1882 to 1930. Seguin and Rigby's was built to fill in around it, covering the rest of the country from 1883 to 1936, each case confirmed in a period newspaper.

The record has holes: the Deep South after 1930, everywhere before 1882, and killings no newspaper reported.

Who was lynched, by race

Read the race labels with care. The old lists had two boxes, Black and white, and put everyone who wasn't Black under white: Mexican, Chinese, Native American, and Italian victims among them. Researchers who recoded the lists, Lisa Cook among them, found the white count too high for that reason, and the Chinese, Hispanic, and Native American counts too low. That is a point about labels, not about population.

The census had the same habit. Through 1920 it had no Mexican race and counted people of Mexican descent as White, so its White is whites and Mexicans together. A rate needs victims counted the same way as the people they are divided by, so CrimeDB shows that group as "Whites, Mexicans", named for everyone in it. In 1930 the census printed Mexican apart, and so do we.

The population point runs the other way. In the ten Deep South states, Black residents were lynched at 6 to 49 times the rate of Whites, Mexicans, decade by decade. The Black rate peaked in the 1890s, at 1.53 per 100,000 a year, and fell to 0.302 by the 1920s.

Outside those ten states the record is Seguin and Rigby's, through 1936. Two things limit it. The 1890 census printed two counts: the total population, and a general enumeration that leaves out Indians on reservations and in Indian Territory, and a few others. Lynchings happened in those places too, so the 1890s use the total population; the Indian Office's own 1890 count of Indians agrees with it more nearly. And the 1880 census did not count Indian Territory, so Oklahoma's 6 victims in the 1880s are in no rate.

A joined group can hide who it holds. In Texas in the 1910s, 29 of the 32 victims counted with Whites, Mexicans were Mexican.

Every state, by group and decade, where a rate can be made. A blank means the inventory doesn't cover that decade in that state, or the census had no such group that year.

What do modern records show?

Homicides and aggravated assaults police recorded as bias-motivated, and federal prosecutions under the Emmett Till Antilynching Act of 2022, from Justice Department announcements.