import pandas as pd
from pandas.tseries import offsets
import numpy as np
import datetime as dt
import yaml
import matplotlib.pyplot as plt
from matplotlib.pyplot import cm
import matplotlib.ticker as mtick
from matplotlib.ticker import (MultipleLocator, FormatStrFormatter,
                               AutoMinorLocator, NullLocator)
import matplotlib.dates as mdates
import seaborn as sns

Use of Solitary Confinement at the Northwest Detention Center: Data Appendix

3. ICE SRMS Data for NWDC, 2nd Installment Aug. 2020

UW Center for Human Rights, 30 November, 2020

Back to Data Appendix Index

Data analyzed:

1. ICE Segregation Review Management System dataset of reported solitary confinement placements at NWDC released to UWCHR via FOIA on August 12, 2020 (second installment, "SRMS 2").
   • For a parallel analysis of a previous installment of ICE SRMS records for NWDC released to UWCHR in February 2019 ("SRMS 1"), see Data Appendix 3.1. NWDC SRMS Records Installment 1

csv_opts = {'sep': '|',
            'quotechar': '"',
            'compression': 'gzip',
            'encoding': 'utf-8'}

# Import and test asserts for UWCHR segregation log dataset
srms2 = pd.read_csv('input/srms-2.csv.gz',
                 parse_dates=['placement_date', 'release_date'],
                 **csv_opts)

assert len(srms2[srms2['release_date'] < srms2['placement_date']]) == 0
assert len(srms2) != len(set(srms2['tracking_number'])) # Tracking number redacted
assert sum(srms2['placement_date'].isnull()) == 0

Dataset Structure and Cleaning Process

In a private repository, the document released by ICE (2019-ICLI-00003_-_SRMS_Report_from_CM_-_NWDC_Redacted.xlsx) has been edited to remove fields containing un-redacted sensitive and personally-identifiable information; converted to CSV format and compressed. The resulting CSV has been minimally cleaned in a private repository, dropping 6 duplicated records and generating unique identifier field hashid; cleaning code available upon request. The resulting file, srms-2.csv.gz, is described and analyzed in the continuing sections. Code for this process is available for review upon request.

Dataset Description and Analysis

The dataset released to UWCHR includes 461 records of solitary confinement placements at the Northwest Detention Center. Each record relates to a specific solitary confinement "placement". Names and A-numbers of detained people are redacted from the dataset, making it difficult to determine whether any records relate to detainees experiencing multiple solitary confinement placements; however, comment fields excluded from the public version of this dataset make clear that some represent repeat placements.

The earliest placement date (start of solitary stay) was September 3, 2013 and the latest placement date was March 16, 2020. The earliest release date (end of solitary stay) was October 3, 2013 and the latest release date was March 20, 2020.

The maximum number of solitary confinement placements in a single month was 17 in December 2013. The overall average number of solitary placements per month was 5.84.

Figure 1: Total monthly solitary confinement placements at NWDC

423 people placed in solitary were male, 38 were female. People placed in solitary confinement at NWDC were from 64 countries. The top 5 countries of citizenship are listed in Table 1 below.

Table 1: Top five countries of citizenship of detainees placed in solitary confinement at NWDC

Length of Segregation Placements

8 records do not have a release_date specified. All of these records are segregation placements beginning during the latter portion of the dataset, which suggests that these refer to individuals who remained in segregation at the time of production of this dataset.^[1] For purposes of analysis of placement length, we exclude these records, leaving 453 completed solitary placements.

The maximum completed segregation placement length was 691 days. The average segregation placement length was 59.41 days and the median segregation placement length was 31 days. The most common placement length was 29 days.

The median and mode values are close to ICE's maximum length of 30 days for most disciplinary segregation placements.^[2] 49.23% of placements were for 30 days or less, and 50.77% of placements were for more than 30 days.

71, or 15.67% of placements were for 15 days or less, and 382, or 84.33% of placements were for more than 15 days. Under the United Nations Standard Minimum Rules for the Treatment of Prisoners, segregation for more than 15 consecutive days is prohibited. Note that ICE regulations do not require reporting of all segregation placements, as described above.

The following histogram chart displays the distribution of length of stay for completed segregation placements:

Figure 2: Distribution of length of completed segregation placements (days_solitary) at NWDC

Placement Reason

Each solitary placement is categorized with one of 18 placement reasons (including "Other"). As mentioned above, the maximum solitary length for most "Disciplinary" placements is 30 days.

We can also aggregate these placement reasons into broader categories for comparison, as in the following figure:

Figure 3: Distribution of variable days_solitary by placement_reason_type

As seen in the following table describing the lenth of solitary placement for each detailed placement type, the most frequently used reasons include broad categories such as "Disciplinary", "Protective Custody: Other Detainee Safety" and "Facility Security Threat: Other".

Table 2: Description of variable days_solitary grouped by placement_reason

Application of different placement reason categories varies over time, as seen in the following figure, which shows trends in the proportion of placement reasons per fiscal year period. Note the proportional decrease in "Disciplinary" placements (which involve an administrative hearing process), and increase in "Protective Custody" and "Facility Security Threat" placements, which tend to be longer on average than disciplinary placements.

Figure 4: NWDC segregation placement reason categories as proportion of total placements

We can also illustrate placement reason and stay lengths over time. Note the frequent disciplinary placements during late FY2012 and FY2013 with stay lengths around 15-40 days, and the cluster of long "Facility Security Threat" placements during FY2016:

Figure 5: NWDC segregation placements by reason and stay length over time

Mental illness

While "Mental Illness" is occasionally employed as a placement_reason value, all records also include a mental_illness field indicating whether the detained person has been diagnosed with a mental illness or serious mental illness. 131, or 28.42% of records indicate a "Mental Illness" diagnosis; 27, or 5.86% of records indicate a diagnosis of "Serious Mental Illness". Remaining records indicate no mental illness diagnosis.

Figure 6: Count of placements by mental_illness Category

As shown in the preceding section, 6 records indicate a placement_reason of "Mental Illness". Of these records, 3 indicate a mental_illness value of "None".

As illustrated in the following table and chart, stay lengths for individuals with a "Mental Illness" diagnosis tended to be shorter than those without. However, stay lengths for individuals with a "Serious Mental Illness" diagnosis tended to be longer:

Table 3: Description of variable days_solitary grouped by mental_illness

Figure 7: Distribution of variable days_solitary by mental_illness

"Facility-Initiated" versus "Detainee Request" placements at NWDC

Each placement is coded as either "Facility-Initiated" (325 placements) or "Detainee Request" (136 placements). As seen in the below table, the majority of reported detainee requested placements were for "Protective Custody", with a small minority classified as "Other."

Average stay lengths for "Detainee Request" placements were considerably longer than for "Facility-Initiated" placements (100.29 versus 42.24) days; however, we note that stay lengths vary greatly, as seen in the following figure.

Figure 8: Distribution of variable days_solitary grouped by detainee_request

```

Table 4: Description of variable days_solitary grouped by detainee_request and placement_reason_type

Next section: Data Appendix 4. National ICE SRMS Data

Back to Data Appendix Index

Notes

<sup>[1]</sup> No records are missing placement_date values, and no record has a release_date preceding its placement_date.

<sup>[2]</sup> See Section 2.12 "Special Management Units" in ICE’s detention standards manual: https://www.ice.gov/doclib/detention-standards/2011/2-12.pdf