ABSTRACT
Whengatheringdataforstatisticalanalysis,researchersmakenumerousassumptions.Infact,mostareprobablymadeatthesubconsciouslevel.Here'san example:Manyyearsago,ProfessorGeorgeMahlhadaresearchassistantadministersometestsinapilotstudythatwouldextendhisworkintoanewarea.The resultswereexcitinglyconsistentwiththeprofessor'shypothesis.Thestatistics collectedbytheassistantwouldhavebeenpublishedexceptforthefactthatwhen heaskedtwoadditionalassistantstorepeatthetests,theresultsdidnotagreewith theoriginalassistant's.Notethattheprofessorwascarefulandconductedreplications(repetitionsofthestudytoseeifthesameresultsareobtained).
