Scott Newark answers his critics
I see my analysis of the StatsCan 2009 Police reported crime statistics has ruffled a few feathers and caused some responses both critical and supportive. Frankly, that’s a good thing because this is a subject that could do with some light of day and informed debate.
The latest comment was authored by Professor Tony Doob and Criminal Defense lawyer Lawrence Greenspan. Their criticisms merit a response.
A little context is required at the outset. In my analysis (p.7), it was specifically noted that the purpose of the review was not to serve as a substitution for the work of Statistics Canada but rather to identify deficiencies in what is being by reported by them in the Police Reported Crime Statistics, to offer insights and examples of the deficiencies and to offer tangible suggestions to create more accurate and relevant information for Canadians. It does exactly that.
Professor Doob and Mr. Greenspan suggest I’m comparing data that can’t be compared. Actually, I’ve noted the difficulties in comparing data throughout the Report but more to the point, identified that this difficulty has been created by the frequently changing StatsCan methodology which makes relevant comparisons harder to do. They also inaccurately suggested that I criticized StatsCan for not presenting all data on all crimes which they suggest is irrelevant anyway because it’s supposedly all available somewhere on a StatsCan website. Nothing could be further from the truth. The point of my analysis was to suggest a modernization in the way data is collected, analysed and then reported to the public (who are paying the bills) so that the most relevant data is presented in the most accurate fashion possible.
Relevance is clearly in the eye of the beholder but what gets reported is a choice made by StatsCan that I strongly suggest needs to be revisited. The 2009 StatsCan Report, for example, reports the number of reported incidents of causing a disturbance but not first or second degree murder. I understand that may be more relevant for Professor Doob and Mr Greenspan but I’m willing to go out on a limb and suggest it isn’t for most Canadians.
My suggestion that certain most serious violent crimes be grouped together is also simply a recognition that a better measure of high end violent crime are the defined incidents rather than a body count. Once again, this needs to include an informed debate on what should be included but because StatsCan reports what it does should not be the end of the debate.
Mr Greenspan and Professor Doob also take exception with a number I used for youth violent crime (rate and number) from 2001. They might be interested in knowing that the source of my data was…a StatsCan Crime Report from 2002. I was actually looking for numbers to do a comparison but, again unfortunately, StatsCan has changed what they report by reduced numbers and reduced years of comparison which should be explained. It is possible that the difference in numbers is a reflection of youth crimes ‘diverted’ out of the courts rather than resulting in charges being included in their figures although, once again, that information is no longer reported by StatsCan which my report mentions.
The numerical comparison in youth crime homicides between the two reports, by the way shows 30 such youths ‘charged’ while 2009 shows 79 youths ‘reported’ for homicides.
Interestingly, since writing this Report, I’ve had several constructive exchanges with StatsCan officials. As a result of an inquiry about the reporting of murder as an offence, I was directed to other StatsCan reports one of which (Homicide 2009) confirm the dramatic increase in youths “accused” of homicide in 2009 which was an increase of 23 from the year previous and the second highest rate per 100K in thirty years.
That same report also confirms the disturbing fact that 152 of the total 610 reported adult homicides, or approximately 25%, were unsolved which is an increase in number and per centage from the averages between 1999 and 2008. Asking questions is a good thing it turns out even what it irritates some people.
In short, by all means let’s make sure we are making the most accurate comparisons possible and that what gets gathered, analysed and reported is as meaningful and reliable as possible. Does anyone really believe that the methodology used by Statistics Canada doesn’t affect what they report? Don’t take my word for it. Here’s what StatsCan says itself.
There are a number of ways of measuring the incidence of crime and each method will yield a different result. The characteristics of the counting process will affect the count which is obtained. Different data collection systems will produce different figures for the same series of events since the count of events is a reflection of the definitions which are used and the manner in which the data are gathered.(Canadian Centre for Justice Statistics: Policing Services Program- Uniform Crime Reporting (05/Dec.02)
Describing the Crime Severity Index as ‘objective’ is also a stretch for anyone who actually reads how StatsCan describes it which I invite everyone to do.
In the calculation of the CSI, each offence is assigned a weight, derived from sentences handed down by criminal courts. The more serious the average sentence, the higher the weight for that offence. As a result, more serious offences have a greater impact on changes in the Index. (p.9)
Got that? It’s really an undisclosed assessment of undisclosed offences based on sentences..which aren’t disclosed…and which are themselves inherently subjective…that are used for the ‘severity’ assessment. It also isn’t clear whether the ’sentence’ is what the judge announces for the media (which takes into account extra pre trial credit) or the real sentence which is what’s written on the warrant of committal. How about defined serious offences as the metric?
As for comparative data, lacking a social life, I am aware that Reports prior to 2008 contained crime volume and rate data for the preceding five years and, for some offences, data from ten years previously. They managed to fit it into the same sized reports as the 2009 version so maybe someone should explain why StatsCan chose to reduce the comparative data for more thorough reporting of crime volume and rate trends. Choices are made for a reason.
I also notice that neither critic disputes my recommendation that StatsCan should report on the criminal profile of the persons for defined offences. Does anyone really think it wouldn’t be helpful for systemic accountability or for operational and policy decision makers not to know how many of the most serious crimes were committed by persons on bail, probation, parole, subject to criminal deportation or with previous federal incarceration.
Meanwhile, the reported rationale for only reporting the ‘most serious’ offence in multiple offence incidents is that everyone else does the same thing. Are we capable of more accurate reporting? Shouldn’t we at least ask?
This is an admittedly complex subject but not everything is discernible only to learned statisticians or professors. From even the reduced comparative data (reduced in 2008) I noticed the following increases in the volume of crimes from 2008 to 2009 that are on the higher end of significance for most people.
*Homicide and attempt murder increased by 84 incidents
*Sex assaults against children increased by1185 incidents
*Using/pointing/discharging a firearm increased by 237 incidents
*Kidnapping/unlawful confinement increased by 76 incidents
*Child porn increased by 205 incidents
*Trafficking (not coke or marihuana) increased by 582 incidents
None of these facts are included in the StatsCan Report “highlights” (pp5-6) and I don’t recall hearing about this when their report was released. Do you?
Because StatsCan has also chosen to alter what crimes were included in certain reported categories of crime they now don’t report changes in robberies with firearms or aggravated assaults/assaults with weapons both of which used to be tracked. From the data I was able to find, the aggravated assault/assault with a weapon increased from 37,500 in 1999 to 49,600 in 2005, a rise of 32 percent, while the rate per 100,000 increased from 123 to 154.
By all means let’s consider population growth but volume of crime is not irrelevant especially considering age demographics and that justice system success is not simply having more of us out there for the criminals to choose from. Wasn’t the point of the gazillions we spend annually on our corrections system to get people to not commit crimes anymore? How’s that working? Oh…I forgot…we don’t ask.
Finally, anticipating the response from the criminology and academia crowd to any criticism, I actually had the Macdonald Laurier Institute include this quote (p28) at the release of the Report which has characterized my analysis of crime issues throughout my career “…instead of being “tough” on crime, it’s better to be honest about crime so as to be smart about crime.”
Vitriol aside, I’m glad to see the Report has generated a discussion on these crime statistics that includes the desirability of making them as accurate and relevant as they can be. I sincerely hope my Report contributes to that effort.
By Scott Newark
- The Crime Statistics Debate By Scott Newark I see my analysis of the StatsCan 2009 Police reported crime statistics...
- Scott Newark in the National Post: StatsCan’s reporting has improved, but Canadians still need better access to information about our safety August 2, 2011 – In today’s National Post, Scott Newark discusses StatsCan’s recently released report on 2010...
- Counting crimes By Brian Lee Crowley, National Post, February 28, 2011 Just how bad is crime in...
- MLI author Scott Newark in the Media: Crime-rate numbers don’t tell whole story Since the release of Scott Newark’s National Post article on August 2, 2011, “Let’s be...
- New MLI paper…Canadian Crime Stats Don’t Add Up February 09, 2011 – Ottawa The Macdonald-Laurier Institute (MLI) today released Why Canadian crime statistics...
Please note that the link above was live when posted. Older links may no longer be functional.