USincarcerations {Ecdat}

US incarcerations 1925 onward

Description

Counts of prisoners under the jurisdiction of state and federal correctional authorities in the US. This does not include jail inmates.

Usage

data("USincarcerations")

Format

A data frame with 95 observations on the following 7 variables.

year

an integer vector giving the year c(1925:2019).

stateFedIncarcerees

Total number of incarcerees = maleTotal + femaleTotal.

stateFedIncarcerationRate

incarceration rate = stateFedIncarcerees per 100,000 population.

stateFedMales

Total number of male incarcerees.

stateFedMaleRate

male incarceration rate = maleTotal per 100,000 males in the US population.

stateFedFemales

Total number of female incarcerees.

stateFedFemaleRate

female incarceration rate = femaleTotal per 100,000 females in the US population.

Details

This dataset began as an effort to update File:U.S. incarceration rates 1925 onwards.png on Wikimedia Commons. Conveniently data on these variables was provided in a table for 1925 to 2014. And a description was given of how to update that table using files p*t03.csv and p*t05.csv from Prisoners In 2019.

An initial rationality check was to compute

checkTot <- stateFedIncarcerees - stateFedMales - stateFedFemales

This was 0 except for 1927 and 1973, when it was 637 and 684. The stateFedFemales for 1972:1974 was 6269, 6004, 7389. We replaced 6004 with 6688, which made the checkTot 0 for 1973.

Similar checks for 1927 yielded nothing as obvious. However, the stateFedIncarcerees increased 6.9 percent in 1926 over 1925, and 12.2 and 5.8 percent in the following two years. Subtracting 637 from 109983 for 1927 gave us 109346, which reduced the increase to 11.6 percent for 1927. It's no longer the maximum annual increase prior to 1975.

Next, these numbers were compared with those in p19t03.csv and p19t05.csv, which include numbers of incarcerees and rates per 100,000 population for 2009:2019. The numbers were identical for 2009:2011, but there were several differences for the more recent counts.

For USincarcerations, we used the numbers from p19t03.csv and p19t05.csv, because they seem likely to be more accurate.

However, these numbers include only people in state and federal prisons. It excludes jails.

Key Statistic: Total correctional population includes a plot of "Total adult correctional population 1980-2016", which does include jails. The data there are available as Total_correctional_population_counts_by_status.csv. Data on these variables covering 2008-2018 are available as cpus1718.csv from "Data tables" at Publication Correctional Populations In The United States, 2017-2018. The data in cpus1718.csv is mostly but not entirely identical to "Total adult correctional population 1980-2016" for 2008-2016, the period of overlap. We therefore used the older data up to 2007 and cpus1718.csv for 2008-2018.

Actual analysis of the jail data is left for another project.

Source

Data from 1925 to 2014 from File:U.S. incarceration rates 1925 onwards.png on Wikimedia Commons, accessed 2020-11-23.

The primary source for the more recent data are files p*t03.csv and p*t05.csv from Prisoners In 2019, accessed 2020-11-23.

Data on jails and community supervision dating back to 1980 are available in Key Statistic: Total correctional population with data on the most recent years available from Publication Correctional Populations In The United States, 2017-2018.

Some time in 2021 or later more recent data should become available. When that happens, it may be desired to update this table to include those numbers – and check for any revisions of earlier numbers.

References

United States incarceration rate.

Examples

data(USincarcerations)

matplot(USincarcerations[1],
  0.001*USincarcerations[c(3, 5, 7)], type='l', 
  xlab='', ylab='incarceration rate (%)')
abline(h=0.5, lty='dotted', col='gray')
lbl <- paste("US incarceration rate", 
  '(percent of the population)', sep='\n')
text(1955, 0.75, lbl)
text(2007, 0.86, 'male', col=2)
text(2007, 0.15, 'female', col=3)