In a lot of industries, there’s a mechanism to review adverse events after the fact, in the hopes of avoiding similar events in the future.
In medicine, for instance, certain events are reviewed, such as unexpected post-op deaths, or patients sent home from emergency who die unexpectedly.
NASA conducts reviews when their rockets blow up, as with the Challenger disaster.
One could imagine something similar in the legal system, where the legal histories of repeat offenders were reviewed, to see if things were done at previous sentencings, bail decisions, parole decisions, etc. that might have been done differently.
I worked for a software startup a decade ago whose CEO demanded exit interviews and reviews whenever a major customer - or even an employee - left after a bad experience. I still remember the company-wide emails from him pushing everyone to learn from what happened and report how their protocols would change.
But we weren't unionized or protected by intrusive HR and privacy policies, so all that was possible.
Aren't court cases, with judge's and defendant's name, already public? If they are, it would be possible and quite straightforward to publish league tables for judges showing conviction/ acquittal and recidivism rates.
Obviously publishing such data in an easily read format would be instantly banned, so doing it offshore and with good security would be essential.
The main complication is that defendants' names aren't necessarily unique (and aren't even explicitly tagged) in the Canclii datasets. So, as far as I can tell, the only to definitively match recidivism events to a particular offender is by manually confirming identities. And that would require a lot of work.
How about age and region? It's not impossible that there would be two habitual repeat violent offenders named David Jones in North Bay, but it would seem unlikely. I don't think perfection is required for this kind of thing to be very useful. If the govt wants truly reliable data out there, then can publish it themselves.
Unfortunately, the data is basically a raw, unformatted copy of the court transcript. There's not a lot there that's useful for this kind of search - although I'm sure that provincial governments have complete databases available that are kept private.
Unions definitely work to prevent these types of conversations from happening!
Great column David! If only a crime reporter (or any reporter) at a mainstream media outlet would take this kind of data-driven, objective approach.
In a lot of industries, there’s a mechanism to review adverse events after the fact, in the hopes of avoiding similar events in the future.
In medicine, for instance, certain events are reviewed, such as unexpected post-op deaths, or patients sent home from emergency who die unexpectedly.
NASA conducts reviews when their rockets blow up, as with the Challenger disaster.
One could imagine something similar in the legal system, where the legal histories of repeat offenders were reviewed, to see if things were done at previous sentencings, bail decisions, parole decisions, etc. that might have been done differently.
I worked for a software startup a decade ago whose CEO demanded exit interviews and reviews whenever a major customer - or even an employee - left after a bad experience. I still remember the company-wide emails from him pushing everyone to learn from what happened and report how their protocols would change.
But we weren't unionized or protected by intrusive HR and privacy policies, so all that was possible.
Aren't court cases, with judge's and defendant's name, already public? If they are, it would be possible and quite straightforward to publish league tables for judges showing conviction/ acquittal and recidivism rates.
Obviously publishing such data in an easily read format would be instantly banned, so doing it offshore and with good security would be essential.
The main complication is that defendants' names aren't necessarily unique (and aren't even explicitly tagged) in the Canclii datasets. So, as far as I can tell, the only to definitively match recidivism events to a particular offender is by manually confirming identities. And that would require a lot of work.
How about age and region? It's not impossible that there would be two habitual repeat violent offenders named David Jones in North Bay, but it would seem unlikely. I don't think perfection is required for this kind of thing to be very useful. If the govt wants truly reliable data out there, then can publish it themselves.
Unfortunately, the data is basically a raw, unformatted copy of the court transcript. There's not a lot there that's useful for this kind of search - although I'm sure that provincial governments have complete databases available that are kept private.
Do you have a link?
Here's the link to the CanLii site: https://www.canlii.org/