/O 



407 



% ■ 



LUNG CANCERr'xThe •Utistical basis for this claim arises, 



^ 



in large part, from two report<^ observations: first, an increase in ciga- 

 '>ette consumption and an increase'tf) lung cancer through the years; and, 



secbnd, a greater Incidence of lung cane^r among smokers than among non- 



- C '^ 



»moker<^vi ->- 



■ /^ 

 ^ <^^ 



L^us make one point clear: we are not {rying to discredit the use 

 of statistical ana^sis in medical research. When p/^i^r\y used, the technicae 

 can be mostlMlsful t^ndicating certain relationships anA-\;»^so<^tiBf trends, 



^•^ %. o' *- . 



and thus can play an impoMant role in the development of hyp<^eses. Like 

 other research techniques, h^l^ver, the use of statistics has its limitattoits. 

 One of these is, as the 1984 Surge^oGeneral's Report pointed out, the fact .; 

 that "Statistical methods cannot cstabRs^ proof of a causal relationship in an 



association. " ^<>„ 



^-/ 



There are numerous examples In the History of science where it 



*. >> 

 was erroneously believed that one factor caused a dfit^ase because of a statis- 



tical association between the two. For example, because'of a statistical asso- 



■'/■ ^ 

 elation, it was once thought that malaria was caused by ''nigh<'air, " because 



people who contracted malaria had been exposed to the night air. ^fiswever,- 



9 

 C 



further research proved that it was a microbe carried by mosquitoes which w 



o 



caused the malaria. In certain areas people believed pellagra was caused *" 



by eating corn, because fhose people who contracted pellagra had eaten a 



