This month we’re taking a look at criminal justice data in Canada. Justice is an area that has long been subject to statistical methods. Police statistics on crime were first published in Canada in 1921. These statistics were spotty and not generally comparable from year to year until the Uniform Crime Reporting System came into service in 1962. This long history and high level of public interest in crime and justice means that there is a wealth of information available for analysis. We took a look at the information available from Statistics Canada and developed some visualizations of data we found interesting. Before we get to that, let’s take a look at the most significant public ranking of Canadian municipalities related to Crime and Justice.
Macleans has a ranking for the most dangerous places in Canada. It is a good interactive listing tool that enables people to sort by type of crime, crime severity index which come from Statistics Canada, and a 5 year trend in crime. A clear link to their full methodology is available on their website. One of the interesting elements is that they rank as “most dangerous” which is somewhat contrary to how municipalities or other groups would often rank the “safest”. This could be to drive clicks to their website. The rankings break out a wide range of crimes, from violent and severe crimes, to fraud to firearms to impaired driving to various drug charges.
There are a number of interesting findings in this data.
First it appears that of the top 20 most “dangerous communities” all of them are West of Ontario based on the “all crime” ranking. The average Crime Severity Index score in Canada is 75.01. North Battleford Saskatchewan scored a CSI 366. On the other end, LaSalle Ontario which is a suburb of Windsor where we are all situated scored a 15.
There are a couple of outliers regarding youth crime rankings. Granby Quebec was ranked 14th in youth crime but 46 on the Crime Severity Index while Collingwood scores a 43 in Crime severity index but 18th most youth crime. Swift Current ranks 6th in youth crime while having the CSI of 83. Most of the other communities in the top 20 have a CSI over 100. For impared driving Whistler BC ranks number one so obviously people are having a lot of fun going to or from the ski hills.
The homicide capital of Canada is Thompson Manitoba which also has the second highest CSI score overall. I would point out that for some of the smaller or more rural areas, due to certain criminal offenses being rare or having small populations you get some skewed data. Thompson is the Cocaine production/trafficking with 92 incidence, Toronto had 504 but had a trafficking per population of 17.05 compared to Thompsons’ 650.
It has been widely reported that BIPOC populations are over represented in the criminal justice system, we can’t do this topic the justice it deserves in a segment on our show and it may be a future show that we comeback and discuss BIPOC data. Just this week Windsor Police Services for the first time released data on the use of force by racialized groups which was reported on our local CBC News. One snapshot that we found was a Statistics Canada research project completed in 2019 that compared crime, incarceration etc. rates between 2007/08 and 2017/18 for individuals that identified as indigenous.
Justice Department’s State of the Criminal Justice System Dashboard
The Justice Department manages a “State of the Criminal Justice System Dashboard”, which shows high-level outcomes of the criminal justice system in Canada and their performance indicators. The nine outcomes are Safe Communities, Fair and Accessible, Confidence in the System, Operation of the System, Resolution Mechanisms, Correctional Supervision, Victims and Survivors, Indigenous People, and Marginalized and Vulnerable People. You can also see more detailed information, including an explanation of what the indicator is, why it is important, its limitations, and its geographical coverage. Data can also be exported to a spreadsheet, which of course makes it much easier to work with.
Some of the indicators can be compared to make them more useful, so it’s a bit unfortunate that there isn’t a tool within the dashboard to easily compare indicators, but perhaps that’s something that will be added in the future! For example, in the Correctional Supervision outcome, there is an indicator called “Mental health services in federal corrections” which tells you the percentage of individuals under federal correctional supervision who identified as having a mental health need and did receive mental health services in response to this need. However, this doesn’t really tell you much on its own as there’s no hard number in this indicator of how many individuals did identify a need. Under the Marginalized and Vulnerable People outcome, there is an indicator called “Mental health needs in federal corrections” that does identify how many individuals indicated they had a mental health need. So by comparing the two numbers for 2016/2017, for example, then we can see that of the 1,106 individuals who identified a mental health need, 929 received mental health services, leaving 177 individuals who did not receive mental health services for their identified need.
This is a relatively transparent dashboard otherwise, with limitations of the data even identified for each indicator along with accompanying resources. You can see the Justice Department is trying to make this usable for the average person. There are, of course, always questions around methodology of how outcomes and indicators were selected. The dashboard preamble states these were “identified through extensive research and feedback from multi-phased consultations with criminal justice system partners, stakeholders, experts and other Canadians”, which doesn’t tell us much about the methodology used in this research. There is a “Data Development” tab that tells us a whopping thirty-eight additional indicators or areas have been identified as important for monitoring and reporting on performance, but they are not yet included. This does demonstrate a definite data gap in the dashboard. Nonetheless, it is still a useful tool and a good starting point for making this data user-friendly.
Statistics Canada Justice Data
Statistics Canada hosts a portal for crime and justice data, which tracks a few national key indicators and hosts 858 datasets on courts, crime, police services and victimization. A lot of this data is interesting and valuable but we found only a few data sets that met our criteria for this month’s podcast: data published over a long period of time that is available at a fairly low level of aggregation.
A lot of this data is published only at the national or provincial level, covers a relatively short period of time, or is presented to the public in the form of infographics and dashboards. Some data that would be really valuable to have are spread across data sets covering different time series. For example, spending by police departments is available in detail for current and recent years, but only rolled up to the national level.
One dataset that did meet our criteria for making an interesting visualization was Table 35-10-0077-01, “Police personnel and selected crime statistics, municipal police services.” It goes back to 2000 and provides a set of comparable statistics for police services across the country. We used Python and the petl library to process the data as published into a format more friendly for visualization, and then built a tool to generate charts comparing selected municipalities. You can find the Python source here and try out the visualization tool here. We’ve included a sample comparing selected Ontario municipalities in this post.
According to Statistics Canada, “The Crime Severity Index (CSI) measures changes in the level of severity of crime in Canada from year to year. In the index, all crimes are assigned a weight based on their seriousness. The level of seriousness is based on actual sentences handed down by the courts in all provinces and territories.” This is an improvement on the traditional “crime rate” calculation which expressed a simple count of all criminal incidents reported to and by police divided by the population. No matter how serious the offence, they are all weighted the same in a crime rate calculation. The Crime Severity Index is intended to provide a clearer view of the seriousness of crime in a given jurisdiction.
Police officers per 100,000 population is a measure used internationally and within countries to indicate the relative police presence in the population.
The “clearance rate” is calculated by dividing the number of crimes that result in a charge being laid (“cleared”) divided by the total number of crimes recorded. In Canada, the source of these statistics is the Uniform Crime Reporting Survey performed by Statistics Canada. This rate is weighted in a similar way to the Crime Severity Index: more serious offences are assigned a higher weight than less serious offences. There are some interesting criticisms of this measure as an indicator – that it can incentivize police to lay charges whether they have solved a crime or not, and that the behaviour of criminals influences clearance rates which obscures them as a measure of the effectiveness of the criminal justice system.
The Statistics Canada table reports a number of metrics related to officers and civilian staff, including the number of officers who are women. This allows us to chart the percentage of women officers in each police service.
If you’re interested in a study that delves into representation of women and indiginous people within police forces in Canada as a whole, Statistics Canada published a review of police resources in 2019. There are some interesting statistics and graphs in that paper. Here’s a few of our favorites:
That shows how we had a period of policy funding per capita increases in the decade between 1999 and 2009. Since 2009 we’ve levelled off on per capita police funding in real terms. How about police officer salaries? It looks like the RCMP tops the list for those but it’s civilian workforce is paid the least.
It’s interesting how the OPP civilian salaries are nearly on par with their officers and that is not the case in other forces. Finally, similarly to other areas of the workforce, police forces are ageing with the youngest force being independent First Nations forces and the oldest being the OPP.
Statistics Canada – Court Workload Data
That’s all about what crimes are occurring and how policing is conducted. What about after charges are laid? We looked into visualizing the Canada court workload indicators and built another tool to look at those indicators. Just like the graphs above, this uses Vega Lite to create the charts and Observable to expose the code and make it interactive. You can choose the geographic regions, the offence category you’re interested in, and the workload statistic. Here’s the initiated cases for homicides for all the provinces (If you’re wondering about Quebec, that province’s data was only incorporated in 2015).
Let’s look a little more closely at one offence – Fraud, for Canada’s provinces. The provinces with the most initiated cases are Alberta, Ontario, Quebec and BC. This makes sense given their populations, but you’ll notice the initiated cases do not perfectly reflect population – Alberta is over represented. It would be interesting to normalize this particular measure to the population of the province.
Alberta’s caseload (the number of open cases) has also been growing at around the same time that initiated cases have increased – that is, more cases are in the queue.
You can also take a look at the number of cases currently in progress delineated by their age. Even though Alberta’s caseload has been increasing in this area, the case age does not appear to differ greatly from the other provinces. Here’s the largest group (aged 6 to 12 months) for example showing all provinces between 20 and 26% of cases being this old.
A small number of cases are older than five years (thank goodness).
Even so, median, which aggregates over the entire set, shows a high case age for Alberta, but also for Quebec.
You can go through the offences and provinces in this way and get an understanding of what the caseloads and wait times are like for various offences. Happy digging!