Skip to contents

Segregation placements under the Mandela Rules

Source: R/mndla.R
mndla.Rd

The United Nations Standard Minimum Rules for the Treatment of Prisoners (the Mandela Rules) define solitary confinement as 22 hours or more a day without meaningful human contact (Rule 44), and prolonged solitary confinement, which Rule 43 prohibits, as more than 15 consecutive days of it. mndla() classifies each placement and gives the share that meets the rules, overall and by group (a year, a region), with three denominators that answer different questions (Sprott and Doob 2021):

Usage

mndla(
  days,
  id = NULL,
  by = NULL,
  w = NULL,
  thr = 15,
  hours = NULL,
  miss = NULL,
  alert = NULL,
  amin = 2,
  elig = NULL,
  excl = NULL,
  level = 0.95
)

Arguments

days

The length of each stay in days: numbers from 0.

id

NULL, or the person of each stay: labels.

by

NULL, or the group of each stay: labels.

w

NULL, or the count each row stands for: numbers from 0.

thr

The days beyond which a stay is prolonged: one number from 0.

hours, miss

NULL, or the average hours out of cell (from 0) and the percentage of days without four hours out (0 to 100) of each stay.

alert

NULL, or alerts on each stay: one column or several (a matrix, data frame or list), each TRUE/FALSE, 0/1 or "yes"/"no".

amin

The alerts at which a stay qualifies whatever its length: a whole number from 1.

elig, excl

NULL, or for each stay TRUE when it may qualify (elig) or when contact meeting the rules was recorded (excl), in the forms alert takes.

level

The level of the exact intervals: one number between 0 and 1.

Value

An object of class mndla, a list: tab, a data frame with one row per group (then "all") and denominator (group, den, n, k those that qualify, share, lo and hi the Clopper-Pearson interval for whole counts, and with hours sol and tor, the stays in each class); cls, the class of each stay (0, 1 or 2); alerts, the count of alerts on each stay (with alert); and rule, "duration" or "rule44".

Details

  • stays: the share of placements;

  • people: the share of people with at least one qualifying placement;

  • cumul: the share of people whose days in segregation add up to more than thr.

With only the length of each stay, a stay qualifies when it is longer than thr days, and with alert, also when it carries at least amin alerts (mental health, suicide risk). With hours (the average hours out of cell a day) and miss (the percentage of days without the four hours out of cell the law requires), the rule of Sprott and Doob: a stay with at most 2 hours out and every day missed is solitary confinement (class 1) up to thr days and torture (class 2) beyond. elig restricts the qualifying stays (to stays with an alert, say, as a proxy for the absence of contact) and excl removes those where meaningful contact was recorded. Rows of an aggregate table enter with their counts as w; the people denominators then need ids.

References

United Nations (2015). United Nations Standard Minimum Rules for the Treatment of Prisoners (the Nelson Mandela Rules). General Assembly resolution 70/175. https://docs.un.org/A/RES/70/175

Sprott, J. B. and Doob, A. N. (2021). Solitary Confinement, Torture, and Canada's Structured Intervention Units. University of Toronto, Centre for Criminology and Sociolegal Studies.

Examples

set.seed(6)
d <- data.frame(id = sample(1:80, 150, TRUE), yr = sample(2021:2022, 150,
                TRUE), days = rgeom(150, 0.08) + 1)
mndla(d$days, d$id, d$yr)$tab
#>   group    den   n  k     share        lo        hi
#> 1  2021  stays  78 26 0.3333333 0.2305852 0.4491667
#> 2  2021 people  49 22 0.4489796 0.3066538 0.5976946
#> 3  2021  cumul  49 25 0.5102041 0.3633783 0.6557526
#> 4  2022  stays  72 20 0.2777778 0.1786429 0.3959058
#> 5  2022 people  47 18 0.3829787 0.2450682 0.5362006
#> 6  2022  cumul  47 24 0.5106383 0.3606447 0.6592430
#> 7   all  stays 150 46 0.3066667 0.2340520 0.3871003
#> 8   all people  68 38 0.5588235 0.4331781 0.6791799
#> 9   all  cumul  68 43 0.6323529 0.5066791 0.7461021
# the rule of Rules 43 and 44 with hours out of cell
mndla(c(5, 20, 30), hours = c(1, 1.5, 3), miss = c(100, 100, 40))$tab
#>   group   den n k     share         lo        hi sol tor
#> 1   all stays 3 2 0.6666667 0.09429932 0.9915962   1   1