By Peter Muller
In a previous issue
www.all-creatures.org/cash/cc2005-w-can.html we
reported on the success of Strieter-Lites, a system of reflectors which,
if properly set up along a roadway, will reduce deer-car collisions.
Numerous studies and installation have reported the effectiveness of the
devices. The Georgia Department of Transportation (GDOT) funded a study
by the University of Georgia (UGA) to reexamine the efficacy of
Strieter-Lites. The study was reported in Volume 34, Number 4, of the
Wildlife Society Bulletin in an article entitled “Evaluation of Wildlife
Warning Reflectors for Altering White-Tailed Deer Behavior Along
Roadways,” by Gino J. D’Angelo et al. Gino J. D’Angelo, a graduate
student at the University of Georgia, concludes that reflectors were
ineffective in preventing deer-car collisions.
The reflectors that were putatively found to be ineffective were
Strieter-Lites, in spite of the fact that other studies had found them
to be 50% to 100% effective in preventing deer-car collisions.
(“A Study
of the Effectiveness of Strieter-Lite Wild Animal Highway Warning
Reflector Systems” by Robert H. Grenier at
http://www.strieter-lite.com/images/scientific_report.pdf
) The discord of the Georgia study with a wide body of evidence to the
contrary invites a closer look at the UGA study.
The UGA study protocol consisted of observing deer at night with an
infrared camera along a campus road at the UGA and evaluating their
behavior as cars approached. In their collection of data, the observer
chose one deer from a group near the road (no criteria for the choice of
the observed animal are provided) and recorded that deer’s motion as the
car approached. The subject’s motion was described as the relative
distance moved along two axes, one parallel and one perpendicular to the
roadway.
The study fallaciously assumed that only motion away from the roadway
along the perpendicular axis would constitute a successful response of
the subject to the Strieter-Lites. This assumption is the most obvious
and fundamental flaw in the study. The function of Strieter-Lites is to
prevent deer from bolting across a roadway when a car with headlights
approaches after dark.
Strieter-Lites make no claim to effect the motion
of deer along the roadway when they are not about to enter the roadway
in the presence of an oncoming car. The data show that the deer grazing
along the roadway continued to graze along the roadway and execute what
seemed an essentially random movement along the axes being observed.
Nothing in those data refuted any claim made for the efficacy of
Strieter-Lites. It is even more surprising that such a basic flaw in the
study escaped the peer-review of the Wildlife Society Bulletin.
In summary, the data showed a behavior of the subject deer that simply
fails to support, much less establish, the conclusion of the study in
the most basic way.
Other parts of the study are also weak. The site chosen is on the UGA
campus; the deer there are habituated to constant traffic and people.
Not surprisingly, they fail to take special note of traffic and people
in their vicinity. Most deer-car collisions occur in rural/suburban
settings where essentially deer-in-the-wild come only occasionally into
contact with motorists. It seems that a study measuring the effect of
those reflectors should be set up in such an environment.
Further, the lack of criteria for choosing a subject to follow upon the
approach of a vehicle seems very suspicious and invites the
“cherry-picking” of data points by the observer. Did the observer
consistently pick the deer furthest from the road so that movement even
further away from the road was less likely to occur? Did the observer
pick the most placid deer so that less movement would result due to the
temperamental disposition of the subject animal chosen? Did the observer
pick a deer that was moving with a group so that its feeling to maintain
cohesion with the herd outweighed any disposition on its part to move
independently? The lack of criteria for choosing a subject suggests a
basically flawed study.
The major failure of the study remains its failure to observe data that
have relevance to the effectiveness of Strieter-Lites. That the deer do
not move away from the roadway is no more relevant to the efficacy of Strieter-Lites than that the deer did not start jumping up and down, or
that they did not attempt to climb trees. Strieter-Lites make no claim
to cause or prevent any of these behaviors that did not occur. They are
designed to prevent deer from bolting across a roadway when a car with
headlights is approaching – and that they evidently do exceedingly well.
In the six-month study only a single car-deer collision occurred; the
“normal” number of collisions is between 12 and 24 collisions a year.
The proper, unbiased, scientific conclusion is again that Strieter-Lites
work exceedingly well.
Peter Muller is VP of C.A.S.H. and has written articles for the C.A.S.H.
Courier.
