ABSTRACT
Ofcourse,anyincreaseincocaineusebyadolescentsiscauseforconcern. However,weshouldkeepinmindthatthisparticularincreaseisquitesmalland mightbeaccountedforinlargemeasurebyinaccuraciesinpolling.Allpollsare presumedtobeoffbyatleastasmallamountbecauseoferrorssuchastheselectionofonlyarandomsampleofhouseholdstocontact,thecurrentstateofmindof therespondents,failureofsomerespondentstoanswersomequestions,andsoon. Ifthefirstpollinourexamplewasoffasmallamountandunderestimatedcocaine usewhilethesecondpollthreeyearslaterwasoffasmallamountandoverestimateditsuse,muchoftheobservedincreasecouldbeattributedtotheerrorsin-
herentinconductingpolls.Clwouldbedesirablebefore jumpingtoimportantconclusio
Thegovernmentreportedgeincreaseinadolescents tryingcigarettesoverthesamethree-yearperiod.Thebeginningpercentagewas 18.4%andtheendingonewas20.2%.(Inotherwords,20.2isabout10%higher than18.4.)
Table1showsthestatisticsthatwe'veconsideredsofar.