MANIPULATING STATISTICS
 FOR A PURPOSE
 SCRIPPS HOWARD NEWS
 SERVICE

 Date: Friday, March 12, 1999

 Section: Source: By LINDA SEEBACH

 Scripps Howard News Service

 Memo: COLUMN

 Advance for Sunday 3-14
 (Linda Seebach is an editorial writer for the Denver
 Rocky Mountain News in Colorado.)

 Edition: 

 When research shows that "aggressive driving" is less of
 a problem in New York and Boston than in the rest of
 America's big cities, there's something wrong with the
 research.

 But no one seemed to notice that when an advocacy
 group called the Surface Transportation Policy Project
 released a report last week purporting to show that
 availability of public transit causes people to drive less
 aggressively. A less tortured interpretation of the data
 would suggest that cities crowded enough to support
 mass transit also tend to have a low traffic fatality rate
 because nobody can drive faster than a crawl.

 But that wouldn't be news, would it?

 The project's own summary of its research says it
 focuses "on deadly aggressive driving; that is, fatal traffic
 accidents in which aggressive driving behaviors were a
 factor in the crash."

 It excludes crashes where drugs or alcohol were a factor,
 and includes speeding only if it is what they deem
 "excessive," that is, above 80 miles per hour.

 With that definition, the summary concludes, "we found
 that aggressive driving is a factor in about 56 percent of
 all fatal crashes."

 Sounds very impressive, and we are probably doomed to
 hear the 56 percent figure endlessly cited by people who
 have no idea what the report actually said. But two things
 about this conclusion are misleading.

 One is the definition of "aggressive driving," which
 includes just about every traffic violation you can think
 of. It's true that aggressive drivers commit these
 violations, including improper lane changes, passing on
 the wrong side or without sufficient distance, following
 too closely, driving erratically or recklessly, failing to
 obey traffic signals, changing speed suddenly, turning
 from the wrong lane and so forth.

 But it's false that everyone who commits such a violation
 is an aggressive driver. Some people just aren't very good
 drivers, and even good drivers are sometimes inattentive,
 or distracted by their children or their cell phones.

 The second misleading thing is that the report carefully
 fails to mention that the excluded alcohol- and
 drug-related cases amount to 41 percent of the fatal
 crashes. Together that's 97 percent, leaving only 3
 percent for weather, vehicle malfunctions and the like. It
 may be true that 95 percent, of all traffic fatalities not
 involving drugs or alcohol do involve a traffic violation,
 but it's not interesting. That 95 percent of them result
 from deliberate aggressive driving is untrue, and such a
 claim would be suspiciously implausible on its face.
 Claiming 56 out of 100, however, sounds as if it could be
 right.

 At least, reporters bought it.

 Next the project researchers took their dubious figures
 for "aggressive driving deaths" and calculated a death rate
 per 100,000 people. Among big cities, Riverside-San
 Bernardino, Calif., was highest at 13.4 and Boston lowest
 at 2.1.

 They then looked for statistical patterns and no doubt to
 their delight found that higher use of public transit was
 correlated to lower traffic death rates. This they attribute
 to the baleful effects of sprawl, which they define as the
 kind of environment in which people have to drive nearly
 everywhere they want to go.

 "People get frustrated when they have to spend so much
 time in the car, and that frustration can come out as
 aggressive driving," they say.

 Perhaps so, but wouldn't you then expect that as
 congestion gets worse, aggressive driving would also?
 Not according to the report, which says congested areas
 tend to have lower death rates, "partially explained by the
 lower travel speeds in congested areas, which would
 make aggressive driving crashes less deadly."

 A common-sense explanation, but an equally
 common-sense explanation for the correlation between
 high transit use and low traffic-death rates seems to have
 eluded them. Very crowded places like Manhattan have
 both; very uncrowded places have neither.

 The report is careful to acknowledge that "we cannot
 infer a causal relationship" based on the correlations, but
 in fact does so throughout. "New transit systems could
 help reduce aggressive driving," says the heading on one
 chart, noting that the new federal transportation law
 "authorizes many cities to make a new investment in rail
 lines." 

 In cities where transit use is relatively high, traffic deaths
 drop by far more than traffic decreases, raising an
 intriguing possibility the report barely mentions. Perhaps,
 they suggest, "travelers who find driving most frustrating
 may be choosing to stay off the road."

 Good. If the maniacs take the bus, the rest of us can
 drive in safety.