The Minnesota state legislature in 2001 considered installing cameras
at intersections with traffic lights which would photographs cars that
run through red lights. Legislators are concerned that the number of
accidents caused by cars running red lights is becoming a serious
public health concern.

Another city recently conducted a pilot study to test whether or not
both installing cameras and increasing the frequency of police patrols
cut down the number of accidents taking place at city
intersections. The study involved 40 intersections, at each of which
the number of accidents was recorded weekly for five consecutive
weeks. 20 of those intersections had cameras installed while 20 did
not. In addition, 20 of the 40 had extra police patrolling and 20 did
not. (There was no randomization; it is not known how it was decided
which intersections got which of the "treatments.") Characteristics of
each intersection were also recorded: (1) average number of cars/hour
going through the intersection during rush hour, and (2) whether or
not the intersection had designated turn-only lanes.

The outcome is the number of accidents for each
intersection. Questions of interest are: whether patrolling decreased
the number of accidents, whether cameras decreased the number of
accidents, and whether both decreased the number of accidents even
more.