Guns Figure

Phil Buckley phil.buckley at VKN.VARIAN.COM
Mon Sep 18 11:04:32 MDT 1995


As for stats that are reliable, see the post below...
 
 
 
 
 
The Conservative Edge (TM) - Thursday, September 14, 1995 - Vol.I No. 62
 =======================================================================
 
        Some of my relatives were more excited than I was over the fact
that my recently published book, "The Vision of the Anointed," was on the
Wall Street Journal's best-seller list. Partly that is because I have my
doubts about the reliability of any of the various best-seller lists -- and
doubts about statistics in general.
        Some books on the best seller list don't sell as many copies as
other books that are not on the best-seller list. That is because all these
lists leave out many of the places that sell books. The New York Times, for
example, leaves out book clubs and specialty book shops.
        The Book-of-the-Month Club may be selling your book like hot cakes
but, somehow, that doesn't count. Books sold by religious book stores also
don't count, even though some of these books sell big.
        This is not a problem peculiar to books. All sorts of statistics
that we accept uncritically are based on very shaky methods. The annual
college rankings in "U.S. News and World Report" magazine would be very
questionable, even if the statistics on which these rankings are based were
reliable. But now it has come out that colleges lie to the magazine about
such things as what percentage of their students graduate.
        In any field and in any country, the more some statistic is used as
an indicator, the less it will indicate.
        One of the fatal flaws of government-run economies is that the
people who control things have to rely on statistics, since they cannot be
everywhere to observe with their own eyes. Even when the statistics are
accurate, they are seldom up to date.
        Back in 1992, when the Clinton campaign was giving President Bush a
hard time over the state of the economy, the economy was in fact doing very
well. But neither Clinton nor Bush knew that and neither did the voters,
because the statistics showing what was happening at that moment did not
appear until long after the election.
        One of the problems that was never solved during the entire history
of the Soviet Union was how to measure the success or failure of a given
factory. When factories making nails were rated according to how many nails
they made in a given time, they began making lots of tiny nails, in order
to look successful.
        When the authorities realized what was happening and changed to
measuring the weight of the nails produced, the factories then started
making huge, heavy nails, even though the consumers might need smaller
nails and thumbtacks.
        It was the same story with chandeliers. At first, factories were
judged by how many chandeliers they turned out. Of course, they began
turning out lots of midget chandeliers. When the criterion was changed to
the weight of chandeliers, then huge, heavy chandeliers began to appear. By
the 1950s, Soviet leader Nikita Khrushchev was complaining that chandeliers
were so heavy that they were pulling down the ceilings.
        If there is anything worse than shaky statistics, it is shaky
statistics mixed with politics.
        Ever since it became a big thing politically to "solve" the
"dropout problem," all sorts of efforts have been made to have a high rate
of high school graduation. The most effective way, of course, is to reduce
the educational standards and make high school more of a fun place. As a
result, we have more and more people receiving diplomas that mean less and
less.
        The underlying problem in all these situations is that, when some
number is used as an indicator of success, everyone will try to make that
number as big as possible, whether or not that accomplishes the fundamental
purpose it is supposed to accomplish. Whether it is Soviet factory
managers, school principals, or people who try to manipulate best-seller
lists, they are likely to do whatever will produce the right numbers,
whether or not that produces the right result.
        If we all grew our own food, built our own houses, and made our own
clothes, none of this would matter. We would produce what we wanted and
that would be the end of it. But we live in a world where the vast majority
of the things we use are made by other people. The questions is: How do we
get those other people to do what we want, rather than maximize some
statistical indicator?
        That problem is solved very simply in a market economy, because
people buy what they want and leave what they don't want. Nails or
chandeliers that are too big or too small just don't sell. People who make
things that don't sell have to either change what they are doing or go out
of business.
        Those who flatter themselves that they are some kind of special
anointed who ought to be making other people's decisions for them fail to
understand how nearly impossible it is for them to have accurate
information by which to measure success or failure second hand.
        One of the scariest stories about President Clinton is that he
stays up nights poring over statistics, in order to determine what policies
to follow. It would be much more encouraging if he were spending that time
studying the definitions behind those statistics and realizing how
unreliable such numbers can be.
 
------------------------------------------------------------------------
Thomas Sowell is a senior fellow at the Hoover
Institute and is a nationally syndicated columnist.
COPYRIGHT 1995 CREATORS SYNDICATE, INC.
------------------------------------------------------------------------
 
 =======================================================================
The Conservative Edge (TM) is a publication of Broadcast Media, Inc.
Copyright(c) 1995. All rights reserved.
------------------------------------------------------------------------
To get the Edge, send e-mail to listproc at colossus.net with message
subscribe CONSERVATIVE-EDGE-L Your Name in the body of the message.
For other communication send e-mail to hughes at connix.com or call
Broadcast Media, Inc. at 203.786.5417.
 =======================================================================



More information about the Rushtalk mailing list