COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	\\M0BASL30\M1BASI30\M2NGR40\.
C00007 ENDMK
C⊗;
\\M0BASL30;\M1BASI30;\M2NGR40;\.
\F2\CSTANFORD ARTIFICIAL INTELLIGENCE LABORATORY
\CDEPARTMENT OF COMPUTER SCIENCE
\CSTANFORD UNIVERSITY
\CSTANFORD, CALIFORNIA 94305
\F0
						

\C11 December 1974




To the Editor of the Stanford Daily:

\J	The article in Friday's \F1Daily\F0 on the Stanford Public Interest
Research Group (PIRG) quotes NorCal PIRG's research director as saying,
in connection with automobile miles-per-gallon figures, "...the absence of
such regulations requiring uniform quotation of standardized data provides
an opportunity for deception".  Indeed it does, and so does the absence
of regulations requiring the uniform quotation of
of standardized data by professors in class and politicians on TV.
The problem is that the presence of regulations requiring uniform quotation
of standardized data also provides and opportunity for deception - by
the authors of the standardized data.
In fact, the Environmental Protection Agency, the author of the standardized
data in this case, took the opportunity provided for deception, and the PIRG
seem to be following in their footsteps.

	The EPA data is based on a dynamometer test used to enforce the
emission regulation and doesn't involve driving the car at all.  It understates
the mileage of almost all cars, and, unfortunately, there are substantial
differences in how unfair it is to different cars.  Presumably, the EPA
doesn't prefer one manufacturer over another, and the differential treatment
is an accident.  However, tuning a car for best mileage on the EPA
dynamometer doesn't necessarily tune it for best road mileage, and the EPA
seems to be pigheadedly persisting in using the dynamometer.  It also has
to cover up its own share of the responsibility for the present low mileage
of American cars.  If PIRG attains its goal of "regulations requiring the
uniform quotation of standardized data", manufacturers will acquire an
incentive to manipulate this data by gaming the test, by lobbying for the
test being done in a way that favors their cars, or by kowtowing to or
even bribing the relevant officials.
I think the NorCal PIRG is aware of these defects in the EPA tests, but
doesn't want to confuse the public by mentioning it, since it is a minor
consideration in view of EPA's and PIRG's good intentions.  We may call this
phenomenon \F1the bureaucratic ethic\F0.

	In short, there is something to be said for preserving the First
Amendment - even for car dealers.\.



							John McCarthy
							Computer Science Department
							
DAILY.LE2[LET,JMC]:SU-AI