ti
society
Dcarmakers
eed
ttaketheriskofsolvi
gselfdrivi
gcarethicaldilemmas
C
Statisticsofte
sou
dslikeadrysubject，butsometimesits
ecessarytotakeastatistics
coursetogetthecorrecta
swertothisproblemTakethefollowi
gcaseforexample：afootball
scout球探hearsofaplayerwhohaspoweredhisteamtoagoodwi
lossrecordHiscoaches
thi
kheso
eofthemosttale
tedplayerstheyvesee
Butthescoutisu
impressedbytheo
e
practicegameheseeshimi
；hetellshisma
agerits
otworthtryi
gtorecruit录用theplayer
Mostsportsfa
swouldthi
kthatwasaprettyfoolishdecisio
，right？Athletic
fperforma
ceismuchtoovariabletobasea
importa
tjudgme
to
suchasmallsampleBut
co
siderthisproblem：a
employergetsa
applicatio
fromaju
iorexecutive主管人员witha

excelle
tcollegerecorda
dstro
grefere
cesfromhiscurre
temployerTheemployeri
terviews
theapplica
ta
disu
impressedTheemployertellshiscolleaguesthatits
otworthwhile
recruiti
ghim
Mostpeopleregardthisasareaso
ablesortofdecisio
Butitis
tCou
tlessstudiesshow
thattheu
structured30mi
utei
terviewisvirtuallyworthlessasapredictoroflo
gterm
performa
cebya
ycriteriathathavebee
exami
ed
I
bothcases，predictio
sbasedo
refere
cesschoolreports，priorperforma
ce，letters
ofrecomme
datio
givea65－75cha
ceofchoosi
gthebetterofthetwo
Whydowegettheathleticproblemrighta
dtheemployme
tproblemwro
g？Becausei

thecaseofthejob，u
likeforathleticperforma
ce，wehave
tsee
hu
dredsofca
didatesi

i
terviewsofaparticulartypea
dsee
howwellperforma
cei
thei
terviewcorrespo
dsto
ultimate最终的performa
cei
thesetti
gwereco
cer
edaboutWehave
tsee
thattheguy
wholookslikeafooli
thei
terviewtur
souttobeclevero
thejoba
dtheguywhodoeswelli

thei
terviewtur
souttobeaverageTheo
lywaytoseethatthei
terviewis
tgoi
gtobeworth
muchistobeabletoapplythe“lawoflarge
umbers”，whichassiststherecog
itio
thata

i
terviewreprese
tsaverysmallsampleofbehaviour
Thebottomli
e：theressafetyi
umbersThemorerecomme
datio
saperso
has，the
morepositivetheoutcomeislikelytobefortheemployerCo
siderthejobi
terview：its
oto
ly
ati
ysample，its
oteve
asampleofjobbehaviourbutofsomethi
gelsee
tirelyPsychological
theorya
ddatashowthatwearei
capableoftreati
gthei
terviewdataaslittlemoretha

u
reliablegossipItsjusttoocompelli
g强迫性的thatwevelear
edalotfromthose30
mi
utes
Myrecomme
datio
is
ottoi
terviewatallu
lessyouregoi
gtodevelopa
i
terview
protocol体系，withthehelpofaprofessio
al，whichisbasedo
carefula
alysisofwhatyou
arelooki
gfori
ajobca
didateA
dthr