Ban the Cams note:  A MO Dot commission recentley "embraced" photo enforcement at a hearing that was less a hearing than marketing plan staged by the scamera crowd.  READ BELOW:

http://www.thenewspaper.com/news/33/3376.asp

Quote:  Those policies were developed by "stakeholders" identified by MoDOT as traffic camera companies and local government officials. No representatives from the motoring public or state lawmakers named as participants. The new restrictions developed by this group apply only to traffic cameras on state roads. The new rules are so closely aligned with existing practice that they will not put any of the existing eighty-eight cameras out of business.

http://washmo.com/blogs/guy-midkiff/961

MoDot's 45% Change - Statistically Insignificant     

Local Blogs - Guy Midkiff 
Friday, 14 January 2011 

In my last blog post: MoDot's Immaculate Connection http://washmo.com/blogs/guy-midkiff/960, I drew attention to MoDot's decision to lift the moratorium on red light cameras. My major concern is, "why would MoDot release a study that will have millions of dollars in impact on Missouri drivers - without a shred of evidence."

After several requests, MoDot released their numbers today. What is clear is that they have opened up the proverbial Pandora's Box. The following is my (and several others) initial assessment of their determination. Check back soon, I promise to raise even more questions in the near future:

First, there is no information on how this crash data was gathered:

- What was the distance from intersection crash inclusion zone?

- Were any crashes below $X excluded?

- Were any other crashes excluded from these tallies?

- What specific intersections were excluded?

- What makes a crash a "severe" crash?

- How long is the before data period?

- What method of Bayes analysis was used?

- Were there any changes in traffic volume over the study period?

- Why were no control intersections used in this analysis to validate the data?

- Why is there no breakdown of rear-end crashes?

Without these definitions this data is useless.

Second, their analysis of severe crashes uses numbers that are so small(0 and 1) that they are not suitable for this kind of Bayesian analysis. Given that severe crashes are so rare at these intersections, you cannot make an effective statistical study of this data.

Third, the math doesn't seem to add up on the severe crashes changes. There is no explanation of how they came up with the change numbers they show here.

To me, the real story here is that more than 85% of these RLC intersections had ZERO serious right angle crashes BEFORE the cameras were installed, which is something they conveniently omit in their summary, and something that should make it clear that the 45% change they tout was statistically insignificant.

I would also like to see some control intersection data which could show if these RLC intersections are doing better or worse than their non-RLC counterparts, after all, as Einstein said "everything is relative", and nationally crashes and fatalities have been declining steadily for the last 5 years.

(Another view of the "study" Florida Comments on MoDot "Study" http://washmo.com/blogs/guy-midkiff/963)

Finally, I have not yet been able to find out how much input the red light camera companies had in this study or if a dedicated statistician was used in the analysis compilation. In the power point presentation listed in my previous post, it is clear MoDot at least solicited input from the RLC companies. And since RLC companies stand to make - or lose - vast fortunes, my hunch is RLC companies are up to their necks in this study. I think RLC companies should have to publicly disclose any and all contact they have with public and government officials. I have sent several letters to RLC companies trying to finding out how much money they have spent courting public officials. They have yet to answer a single request. 

Guy Midkiff

Subscribe to this comment's feed

Title
Comment	smaller | bigger
Name
Email
Write the displayed characters
Subscribe via email (registered users only)
	Add Comment Preview