[ASC-media] Media Release: Figures Fool When Fools Figure, Says Statistician

Australasian Science science at control.com.au
Sun Mar 28 23:22:59 EST 2004


29 March 2004

For immediate release

Figures Fool When Fools Figure, Says Statistician

A leading statistician says Australian governments and public agencies are
risking bad policy decisions through poor statistical practice.

Writing in the conScience column of the April issue of Australasian Science
magazine, released today, Dr Nick Fisher says that members of the
Statistical Society of Australia are calling for a professional approach to
the use of statistics to avoid "an extraordinary catalogue of misadventures
-- some hilarious, some disastrous -- arising from failure to deal with the
statistical issues with due care".

Dr Fisher is Past President of the Society, Principal of ValueMetrics
Australia and Visiting Professor at the University of Sydney. He urges
"government departments and agencies responsible for major policy issues to
employ professionally accredited statisticians with oversight of the
collection, analysis and interpretation of data that underpin any major
policy decision". 

Reliance on managers, public servants, doctors, engineers, economists,
journalists, politicians and sociologists who may be well-qualified in their
own fields but are untrained or unqualified in statistics can lead to
decisions that "can be fatally flawed, and may adversely affect millions of
people".

"You are unlikely to trust your life to an untrained doctor, your office
building to an unqualified engineer or your teeth to an unaccredited
dentist. You probably wouldn't risk your money through an accountant without
professional recognition. Yet, every day we are made to rely on policies
affecting our health, prosperity and security that are founded on
unprofessional use of statistical methods."

Though acknowledging that "statistics as a discipline and statisticians as
practitioners have image problems", Dr Fisher deplores that "anyone can call
themselves a statistician, regardless of training and experience". A
professionally accredited statistician will have many years of experience,
plus at least one degree in statistics.

"Whereas statements by financial institutions about probity and risk have to
be authorised, legally, by properly qualified accountants and actuaries,
statistical conclusions derived from major environmental impact studies
require no similar authorisation. Yet, decisions of huge community import
are made as a result of the effect of airport noise on housing prices, or
the degree of residual contamination in a major industrial site."


Please cite AUSTRALASIAN SCIENCE MAGAZINE as the source of this story.

CONTACTS
Dr Nick Fisher on (02) 9922 1623 or 0407 017 016

Examples of statistical success and disaster stories are on
www.statsoc.org.au.

For the full text and permission to reproduce it (630 words) call the
Editor, Guy Nolch, on (03) 9500 0015.

A photo of Dr Fisher is available on request.

The full article can be downloaded as a PDF at www.control.com.au

-----------------------------
Guy Nolch
Editor, Australasian Science
Box 2155 Wattletree Rd PO
VIC 3145 Australia
Phone 61-3-9500 0015
Fax 61-3-9500 0255
Web www.control.com.au





More information about the ASC-media mailing list