17.2 Definitions changes

Comparing this census with the 2001 Census the following changes in definitions have occurred.

Ethnicity:

2001: Ethnicity was defined using “ethnic trumping” i.e. in cases where an inmate identified multiple ethnicities, the assigned ethnicity reflected the following precedence: Māori, Pacific, European, Asian, Other (e.g., an inmate who identified Māori and European ethnicity would be classified as Māori).

2003: Ethnicity was recorded as the primary ethnicity indicated by the inmate.

Note: the “Combined Ethnicity” data will have been relatively unaffected by this change.

Convictions vs. sentences vs. custodial episode:

The manner in which previous convictions and sentences is counted has changed. To illustrate the complexity in this area, an individual offender may attend Court for sentencing on, say 100 individual charges. He may be convicted on all charges, but some of the convictions will be amalgamated and result in a single sentence. However, there may be multiple individual sentences, all of imprisonment, and some of differing lengths. Further, some sentences may be made concurrent with others, while some may be cumulative upon others. Given that in reality the offender can serve only one custodial period at a time, an “aggregate” sentence is determined from all of the combined sentences, with a single start date, parole eligibility date, and statutory release date.

Recent censuses have counted aggregate sentences in determining number of previous sentences for each inmate. However, given the changing definitions of the aggregation processes of sentences, this current census has adopted the custom of counting the individual sentences passed down by the Courts. A new time series has been started on this basis.

The data used in this report also does not include sentences handed down by the Courts but which the Department of Corrections is not required to manage. For example, convictions leading to a sentence such as a fine, or discharge after conviction (i.e., the kinds of sentences applied to minor offences) are not included in this analysis.

Summary:

2001: Counts of “Aggregate Sentences” (a Department of Corrections management construct) were incorrectly labelled “Convictions”.

2003: Counts of “Sentences” are now presented, the term convictions is no longer used. Sentences are as specified by the courts, and do not include minor sentences such as conviction and discharge, fines, etc.

Education levels and qualifications:

2001: Two separate questions asked, one for “highest school qualification achieved”, and a second for “other qualification”, the latter question for multiple selections.

2003: Single question asked for highest educational qualification achieved; the inmate chooses a single item from a list that included school, vocational and other post-secondary options, thus collapsing the two elements in 2001. An additional question on “educational level” was asked with the inmate able to choose a single item from schooling levels but excluding vocational training as a level.

Child dependents:

2001: “In the 4 weeks before you came into Corrections care, how many dependent children under 18 years old were living with you?”

2003: IOMS has input fields “Child Custodial Dependents” and “Child Financial Dependents”. These fields are completed at prison reception. Only the Child Custodial Dependents are summarised in the Children and Marriage chapter, however the combined effect of Custodial and Financial dependents is used in the trend data to try to align it with previous census interpretations.

Gang memberships:

2001: From their own knowledge on census day prison officers were asked to identify inmates as a gang association member and/or gang patch member.

2003: At the inmate’s receiving interview prison officers may add a gang membership record(s), selecting a gang and status for the record(s).

Risk Scores (RoC*RoI):

Note that risk scores are a calculated value and the method of calculation is subject to ongoing refinement and development. Between 2001 and 2003 an increased number of inmate records contained a calculated risk score. However, as a result of changes in both availability and calculation of risk scores, comparisons between 2001 and 2003 risk score profiles are problematic.

Total time in previous custodial sentences:

2001: This was described as “Time spent in prison, defined by days actually served”. However it would appear that the calculation used was in fact the days elapsed between sentence commencement date (SCD) and sentence expiry date (SED): i.e., the nominal sentence imposed.

2003: Calculated from SCD and SED, and described as “Custodial Period Imposed”. Note that sentence and parole laws mean that, in general time served is between one third and two thirds of the total Custodial Period Imposed.

Age at first custodial sentence for sentenced inmates:

2001: Data for only a subset of sentenced inmates were included in the analysis (i.e., sentenced inmates with one or more previous custodial sentence(s).

2003: Data was as described (i.e., includes all sentenced inmates, with age at first custodial sentence as for current sentence if no previous terms).