This note addresses an issue that has received little attention from Canadian criminologists: changes over time in the quality of youth crime in Canada. There is a body of Canadian research on trends in the volume of youth crime, and of "violent" youth crime, but there is little published research apart from the brief annual reports by Statistics Canada (e.g. Brennan 2012) that analyse trends over time in the overall seriousness of youth crime.
The lack of scientific research has not deterred the public and commentators in the media from expressing definite opinions on the matter. While acknowledging the demonstrated decline in the amount of youth crime over the past two decades, these commentators have emphasized a supposed increase in its seriousness. For example, in 2001, the President of the Police Association of Ontario claimed that "the rate of serious crimes committed by youth continues to rise at an alarming rate" ("Police Association" 2001). Respondents to a poll in Halifax in 2004 overwhelmingly (79%) agreed that "teen crimes have become more serious over the last 10 years" (Boomer 2004). In 2008, the St. John's Telegram reported that "people [working] in the province's justice system say [that] ... the types of crimes youths are being charged and convicted of are of a more serious nature" (Morrissey 2008), and the National Post editorialized that "serious offences committed by young people have risen 12% over the past decade, and 30% since 1991" ("Fixing the Young Offenders Act" 2008). Recent initiatives by the Canadian government to "get tough" on young offenders have been justified partly by this supposed increase in the seriousness of youth crime, and claims to that effect have repeatedly been made by Canadian politicians (e.g. Llewellyn 2008; Moore 2006; Nicholson 2007; "Stop youth violence" 2006; Young 2011). Therefore, a report on the issue that draws on statistical evidence should be of value to criminologists, as well as the public and policy makers.
Canadian research on trends over time in the seriousness of youth crime began to appear in the 1990s as part of a broader effort to assess the impact of the Young Offenders Act. Carrington and Moyer (1994: 19-22) compared the average levels of seriousness of youth crime during 1985-90 and 1980-84 in seven provinces and territories. To operationalize "seriousness of youth crime," they used a sixfold ranked classification of Uniform Crime Reporting CUCR) Survey offence types, treating the first three categories as serious and the last three as less serious. They reported that, although the proportions of young persons apprehended for serious offences against the person increased by about one third, there were very few of them relative to all apprehended young offenders (about 3% to 4%) and their increase was more than offset by the decrease in the proportion of serious property offenders, who were much more numerous. Overall, their serious youth crime constituted a lower proportion of all youth crime during 1985-90 than during 1980-84.
Other research has assessed changes in violent youth crime rather than in the overall seriousness of youth crime. Corrado and Markwart (1992: 166-72) analysed police and youth court data for 1986 to 1988 and found a "sizeable increase" (19%) in the number of young persons charged with violent offences and a "modest" increase (9%) in such offences processed in youth court. Relying on more comprehensive analyses of data for 1986 to 1992, Corrado and Markwart (1994, 1995) concluded that youth violence had "more than doubled" from 1986 to 1992 but that its level remained low; they also pointed out that youth homicides had not increased during the period (Corrado and Markwart 1994: 350-55, 1995: 84ff). Responding to Corrado and Markwart's conclusions, Carrington (1995) compared 1986-92 with 1980-83, and reported an 83% increase in the per capita rate of young persons apprehended for offences against the person but argued that much of this was in "offences against the person that are generally nonviolent," especially level I (common) assault (Carrington 1995: 68-70). Moyer (1996) reported that the rate of young persons charged with both "serious" and "other" offences against the person increased from 1985 to 1993, showing no tendency to stabilize; but that there was no increase during that period in the most serious violent offences:
homicide, attempted murder, and level 2 and level 3 sexual assaults (Moyer 1996: 91-93). Similarly, Doob and Sprott (1998) found that the increase from 1991 to 1996 in the per capita rate of cases of assault in Youth Court was due entirely to "minor" (level 1) assaults and that the rates of more serious (levels 2 and 3) assaults were stable or declined slightly.
Measurement of the seriousness of crime
Stylianou (2003) and Ramchand, MacDonald, Haviland, and Morral (2009) provide detailed reviews of issues in the measurement of the seriousness of crime. One great impediment to measuring crime seriousness has been the difficulty of defining it. Opinions differ as to the absolute and relative seriousness of different types of crime. Methods for measuring the seriousness of different types of crime fall into three general approaches. The normative or perceptual approach attempts to measure what people think about the seriousness of crimes. Panels of raters that have been used include representative samples of the general public, available university students, and experts, such as police officers or judges involved in sentencing. Raters are provided with crime vignettes and asked to provide either quantitative estimates of the seriousness of the crime(s) in each vignette, relative to some benchmark (normative intensity), or rankings of the crimes in the vignettes relative to one another (normative ranking). The former method was pioneered by Sellin and Wolfgang (1964) and adapted for Canada by Akman and Normandeau (1967; Akman, Normandeau, Sellin, and Wolfgang 1968); the latter by Thurstone (Thurstone and Chave 1929). While the normative approach can provide useful estimates of the perceived seriousness of specific crimes, as described in detailed vignettes, it has limited applicability to the problem addressed in the present research; namely, to assess the overall seriousness of all known (youth) crime in Canada. Furthermore, the validity of ratings based on the perceptions of certain types of raters (e.g. university students) and expressed in the abstract as representative of public opinion is open to question. Finally, when used in longitudinal research, normative indices of seriousness need to be updated, at significant cost, because public opinion on the seriousness of specific offences is subject to change. As Francis, Soothill, and Dittrich (2001: 729) put it, the normative approach is "unwieldy, certainly unreal and volatile."
A radically different approach to measuring the seriousness of offences assumes that the seriousness of a crime is a function of the harm done, or "cost," to its victim(s). This "economic" approach estimates the cost to the victim of different types of crime, using data such as victim death and injury rates, weighted by the costs of death and types of injury as indicated by the amounts of monetary awards by juries in personal injury accident cases (e.g. Cohen 1988; Roman 2011). This approach, which is not yet well developed, has several limitations, including its one-dimensional definition of seriousness and its inapplicability to many types of crime, such as crimes without identifiable victims.
A third approach, which is adopted in the research reported here, relies on some authoritative indicator(s) of seriousness. A simple example is reliance on statutory classifications of offence categories, such as the distinctions in the Criminal Code between offences against the person, against property, and against neither (other offences), and between summary, dual-procedure (hybrid), and indictable offences (e.g. Carrington and Moyer 1994). Such indicators of seriousness are rather crude: They rely on two or a few ranked categories. A more precise statutory indicator is the maximum prison sentence specified for each category of offence (e.g., Kyvsgaard 2003). Using this as a measure of seriousness assumes proportionality in legislation--that statutory maxima are intended by the legislature to be (and actually are) proportional to the...