The effects of institutional design on offender behaviour

Given that CSC’s institutions were not constructed simultaneously, and that approaches to penitentiary design have changed over time, differences exist in the design of individual institutions and units. CSC’s Facilities Branch was interested in determining whether the design type of an institution or unit is associated with offenders’ behaviour.

We have undertaken a study to examine this question as it relates to medium security institutions. CSC categorizes these institutions into four designations, numbered T-2 through T-5, with greater security and supervision associated with those of higher number. The designations differ in terms of the living unit style (apartment style or cell block), the extent to which offenders’ movement can be viewed and controlled, and the location and type of control post.

Analyses include a total of 5,336 male offenders who were housed in the same institution (or unit therein for those including multiple unit types) from April 1 to June 1, 2011. So far, these offenders have been compared on a number of behavioural indicators according to the designation of the unit in which they resided.

As can be seen in the table, units with lower designations tended to have lower percentages of offenders found guilty of institutional charges or placed in involuntary segregation, and a greater proportion of transfers to a lower security institution (rather than to a similar or higher security institution).

Preliminary follow-up analyses demonstrated that, even when considering offender-level variables such as criminal history risk, unit designation continued to be associated with all of the behavioural outcomes except minor charges. The next step in analyses will be multi-level modeling to fully understand the magnitude of unit designation’s contribution.

Though results to date do not tell us the strength of the association between institutional designation and offender behaviour or establish that the association is due to a cause-and-effect relationship, they do indicate that a link exists. Further analyses are underway.

Prepared by: Renée Gobeil