ICE detention facility daily population codebook

Authors
Affiliations

Graeme Blair

University of California–Los Angeles

David Hausman

University of California–Berkeley School of Law

We produce a dataset at the ICE detention facility-day level to facilitate analysis of the number of noncitizens detained each day and overnight in total and by demographic group. The counts are calculated from the individual-level detentions dataset from our most recent data release from ICE. Each row represents a day in a particular detention facility.

Data structure

Each row in the dataset represents one detention facility on one day. For each facility-day, we report counts of distinct individuals in detention, in total and broken down by demographic subgroups (gender, age, and whether ICE recorded a past criminal conviction). We count individuals in two ways:

  • Any point in the day. The number of distinct individuals detained at that facility at any moment during the day.
  • At midnight. The number of distinct individuals still detained at the end of the day (after 11:59 p.m.). This nightly count is always less than or equal to the any-point-in-the-day count because some individuals are booked out before midnight.

Data processing

We describe the data construction steps here:

  • Create dataset for each day and each detention facility. We identify detention facilities by the “detention facility code” field in the detentions table and identify dates using the “book_in_date_time” and “detention_book_out_date_time” fields.
  • Count distinct individuals. We count each distinct individual detained (1) at any point in the day, or (2) detained at midnight (the midnight at the end of the day, i.e., after 11:59 p.m. of that date). Notice that the count of people detained at any point during the day is greater than or equal to the nightly count, since some individuals are booked out before midnight. For records produced by ICE without a unique identifier corresponding to a person, we count each row as if it is a distinct individual. In general this will be accurate for deduplicating data at the detention facility-day level, but a small number of individuals may be counted twice if they appear more than once in a given detention-facility-day. There are multiple reasonable ways to handle individuals with missing identifiers.
  • Count subpopulations. We code each individual according to their ICE-reported gender, age (under 18, based on birth year; note this is imprecise for those within a year of 18 at the time of the detention), and whether ICE recorded the individual having a past criminal conviction. We then count the number of distinct individuals in each group (e.g., women detainees) by facility and day/night.

Inspect code

Fields

We describe the fields in the facility daily population dataset below.