stonepatientsreportedapainscoreoftenatleastonce
during the procedure.
Our natural assumptionis that the final ratingswould
represent something like the sum of the
moment-by-moment ones. We believe that having a
longerdurationofpainisworsethanashorterduration
andthathavingagreateraveragelevelofpainisworse
thanhavingaloweraveragelevel.Butthiswasn’twhat
the patients reported atall. Their final ratings largely
ignoredthe duration ofpain. Instead,theratings were
bestpredictedbywhatKahnemantermedthe“Peak-End
rule”: an average of the pain experienced at just two
moments—thesingleworstmomentoftheprocedureand
the very end. The gastroenterologists conducting the
proceduresratedthelevelofpaintheyhadinflictedvery
similarlytotheirpatients,accordingtothelevelofpainat
themomentofgreatestintensityandthelevelattheend,
not according to the total amount.
People seemed to have two different selves—an
experiencingselfwhoendureseverymomentequallyand
arememberingself whogives almostalltheweightof
judgmentafterwardtotwosinglepointsintime,theworst
momentandthelastone.Therememberingselfseemsto
sticktothePeak-Endruleevenwhentheendingisan
anomaly.Justafewminuteswithoutpainattheendof
their medical procedure dramatically reduced patients’
overallpainratingsevenafterthey’dexperiencedmore
thanhalfanhourofhighlevelofpain.“Thatwasn’tso
terrible,”they’dreportedafterward.Abadendingskewed
the pain scores upward just as dramatically.