Correctional Service Canada
Symbol of the Government of Canada

Common menu bar links

The Incidence of Family Violence Perpetrated by Federal Offenders: A File Review Study

Warning This Web page has been archived on the Web.

Method

Sample

The Family Violence Incidence Study is based on file reviews of a random sample of 935 offenders who were admitted to federal institutions between June and November in 1992. During this period a total of 2,806 offenders were admitted to federal institutions. Hence, the sample constitutes 33.3% of the intended population. The random selection method was stratified in a manner that allowed for better regional representation of the population of newly admitted offenders. Figure 1 shows the breakdown of the sample by region. The population breakdowns for admissions during the period of study are also included to demonstrate the effect of stratification on the regional sub-samples. Our results are weighted to correct for the disproportionate regional sub-sample sizes resulting from stratification.

Figure 1

Regional Sample and Population Proportions (%)

full graphic of Regional Sample and Population Proportions (%)

All offenders comprising the current sample are male with an average age of 31.8 (S.D. = 10.01) years. At the time of admission, 44.6% were involved in a marital union (including legal marriages and common law relationships) and approximately 18% had three or more previous marital unions. According to the files more than half (59.1%) of the offenders had children or step children. Most of the offenders had not completed high school (71.1%), although almost half (46.3%) had completed some high school. Thirteen percent had completed some post-secondary studies including trade or technical school. Only 1.5% had completed university degrees and 3.6% had completed some university training.

The mean number of previous convictions on record for this sample of offenders was 19.3 (S.D. = 21.9) and 40.5% had previous federal admissions. Thirty-nine percent had previously violated the conditions of community supervision. For the current admission 76.9% percent of the offenders had been admitted on a new warrant of committal, 16.5% had been revoked after violation of community supervision conditions, 0.1% were admitted on exchange of service agreements with the provinces, and 6.3% had been admitted on other warrants. The average sentence length for determinant sentences was 3.74 (S.D. = 2.66) years and 41.3% had been sentenced to terms of less than 3 years. Only 3.9% of the sample had received life sentences. The major admitting offence for the current admission included 20.7% of offenders convicted of property offences, 39.4% for violent offences (non-sexual offences including murder, manslaughter, assault, confinement, arson, robbery), 15.4 for sexual offences, 11.9% for drug offences, and 12.7% for other offences. The major admitting offence refers to the offence that received the longest sentence for the index admission.

While every attempt was made to meet the criteria of random sampling of files, some exceptions were necessary in order to ensure that the project proceeded in a timely and cost-efficient manner. Institutions where random selection of less than seven files occurred were not included in the sample. In addition, the unavailability of some files led to the random replacement of additional files. The lack of availability of files was normally due to an offender transfer to another institution or release. In four institutions, replacement lists were exhausted. Files from other institutions were substituted in order to meet the regional sampling quota. It is unlikely that these departures from the random sampling design have resulted in major misrepresentations of the regional and national admission populations. Table 1 compares demographic characteristics and offence history indicators for the current admission sample with the population of admissions from which the sample was drawn. Overall, the figures demonstrate that the sample is representative of the population on the various indicators that were examined. The only exception involves admission type, where the number of offenders who had been admitted on a revocation appears to be slightly under-represented in the sample.

Table 1

Comparison Of Sample And Population Characteristics

Comparison Of Sample And Population Characteristics

File Review

File reviews were completed by five research assistants who completed a 5-day training program prior to the data collection. Files were reviewed at the institutions where offenders had been admitted. The file reviews were based primarily on information contained in case management and psychology files, although occasionally sentence administration files were also consulted.

The file review protocol contained information on personal and demographic variables, family background, intimate relationships, children/dependants, history of violence, psychological/emotional issues, participation in treatment programs and substance abuse history. The file review instrument is included in Appendix A. To increase the reliability of the information collected, a data collection supervisor conducted interrater reviews of files completed during the first 4 weeks of the project. At that time, file reviewers were provided with feedback and written clarifications for improving on the reliability of the information. Periodic information bulletins were released to file reviewers as required throughout the project in order to increase the quality and consistency of the file review process.

In many cases the file review protocol demanded information which required subjective judgement on the part of the research assistants. Often inconsistencies were detected within files concerning the occurrence or non-occurrence of events (e.g., childhood victimization by a parent). The quality of the data is obviously limited by the necessity that file reviewers were sometimes required to make judgements on the occurrence of events when detailed information was scant.

The subject of the file review exercise involved material that was not necessarily gathered and recorded in a systematic fashion by CSC staff who maintain offender files. Two important factors condition the quality and completeness of information about family violence contained in offender files. Firstly, file information reflects information available to CSC employees at the time it was recorded. Secondly, the recorded information reflects what CSC staff members regarded as important documentation. There is a possibility that social history information for offenders who were introduced in the federal system several years ago (e.g., recidivists) may not contain the same degree of detail as files for more recent first time offenders.