Data, Photo Radar and Traffic Safety

This morning as I was reading the Edmonton Journal I came across an article by David Staples criticizing Photo Radar.  I wasn’t really that surprised, these seem to come at about the same pace as trucks wedging themselves under the High Level Bridge.

What was surprising was that David was calling for abandoning photo radar all together as a traffic safety tool.  This was a shift from his earlier critiques where he would argue that the City was misusing Photo Radar for the purposes of revenue generation (Cash Cow) as opposed to solely as a safety tool, but would concede that photo radar has merit when used properly.  Indeed, I have had a few back and forth debates on this with David and he’s been reasonable (indeed we get along well and agree on a lot of topics) 

I have always liked photo radar. I have anecdotal experiences in Edmonton and Germany that have demonstrated its effectiveness.  Here in Edmonton, I used to commute on Whitemud Drive every day.  The average speeds were 100+kms/h.  After photo radar came in with strong enforcement, speeds dropped by about 20km/h.  It was clear to me at that time that photo radar affected behavior.  Similarly, I noticed when I went to Germany how nobody tailgated and speed limits were closely adhered to.  Photo radar was everywhere.  The roads were much much safer because of it.  

Those observations took place long before I was an academic or a planner though.  I now know that if there is clear evidence one way or another that I need to check my own experience and thought and accept the data (provided it is solid).  This is the basis of evidence based decision making.  Now, there is lots of published research on this supporting the use of photo radar for safety purposes, but I won't get into that now because David shared a local case study that has some interesting results.  

David noted that that Strathcona County removed photo radar in 2012 and this had no impact on per capita collisions when before and after data was examined in a new study by the University of Victoria’s School of PublicAdministration.  The study was carried out by a graduate student, Jason Gariepy.

I looked at the study and notwithstanding some criticisms I have about it, the data looks pretty clear and supports the statement regarding accidents per capita (the data was included at the back of the study).  I created my own chart below to show this.  Note, all data is taken from Jason Gariepy’s publication linked here and I have not checked the validity of this data.  I made no changes other than smoothing the population data for years between censuses.  Given that there are 4 years of data since the removal of photo radar, I included the prior 4 years of data from the period when photo radar was last active.  According to the report, this was a period where photo radar was heavily used.

Note, the author provided data for four types of collisions.  Property damage incidents, Minor injury collisions, Major Injury Collisions, and Fatal Collisions.  

As can be seen, the chart that supports both David’s and the report author’s conclusion that removing photo radar did not change accident rates significantly.

After reading through the report though, I came across the data tables at the end. As I looked at those tables and within two minutes, I noticed an apparent trend in accidents where individuals experienced major injuries and fatal accidents.  After some double checking it was a very clear trend.  In the four years prior to removing photo radar, there were 35 Major Injury or Fatal Accidents.  In the four years after removing Photo Radar, there were 69 Major Injury or Fatal Accidents.  This is a hugely significant outcome.  With a 30% increase in population, Sherwood Park saw a 96% increase in the number of Major Injury or Fatal accidents.  The shift is very clear as can be seen in the table below.
 

Whatever the underlying factors are causing this increase in major injury and fatality collisions, it is clear that the data strongly suggests that the removal of photo radar has greatly increased the rate of major injury and fatal accidents.  In other words, the roads are less safe than before photo radar was removed.  This is consistent with accidents occurring at higher rates of speed. Indeed, this data demonstrates a much stronger influence than I have seen in studies elsewhere.  To make really solid claims about this, I would probably dig a little deeper and perhaps request some data on speeds in key intersections or other data that would support my conclusions.  That didn’t happen in the initial study however when it's claims of 'no impact' were made.   

As David points out in his article, Jason Gariepy, who authored the study, is a former Strathcona County Councillor who voted to remove photo radar.  Further, Councillor Brian Botterill, who was the Councillor who introduced the motion to remove photo radar from Strathcona County, contributed funding for the research (something openly disclosed by Gariepy).  The question for the research was framed very specifically and the report did provide information on that.

It also provided great data to help proponents for photo radar.  While more work needs to be done, this data suggests that photo radar reduces serious injuries and saves lives.  

As always, the process of building knowledge is an ongoing process – standing on the shoulders of giants as they say.  I am glad that I could dig a little deeper into the data provided and help to shed some light on some other aspects of this.  I know we all support evidence based, data driven decision making, and as such, I look forward to David’s next article calling for the return of photo radar to Strathcona County.