False positive and false negative tests

FALSE POSITIVE AND NEGATIVE TESTS 6

Falsepositive and false negative tests

Falsepositive and false negative tests

Question1aMonthwith thehighestsubscription:JanuaryMonthwiththelowestsubscription:FebruaryQuestion1b

Mostcompaniesororganizationsthat requiresubscriptionneedtheir clientsormembersto dosoon a yearlybasis.Since mostof thesecompaniesconsidera yearas runningfrom January to December, mostof theclientsare alwayssupposedto paytheir renewalsubscriptionsin January since italwaysmarksthebeginningof theNew Year.Question2 aFalse positive rate: (1-0.95) = 0.05Falsenegativerate:0.95Question2 bFalse positive rate: (1-0.975) = 0.025Falsenegativerate:0.975Question2 cFalse positive rate: (1-0.95) = 0.05Falsenegativerate:0.95Question2 d During hypothesistestingin statistics, an errorknownas a typeoneerrorfrequentlyarisesout when the correct null hypothesis is rejected when it a actually.Thisconditionisalsoreferredto as thefalsepositivestatein hypothesistesting.Thelevel of significanceis alwayschosenbasing on theseriousnessof makinga typeI error orin thiscasebasingon theproblemsassociatedwith thefalsepositive.Whencarryingout statisticaltests,a level of significanceis alwayssetto enablea numberoftestsin statisticsthatarecarriedout ona givensubjectunder study.In anytest,thestatisticalsignificanceof anygivenresultis alwaysattainedwhenthep-valueis lessthan thelevel of significance.

Thelevel of significance,which is also known as the p-value,is theprobabilityof observinga valuelessthan theset level of significance.It can alsomeanthenumericallikelihoodof makingan error referred to as atypeI error. Thislevel of significancealwayssetsthebasisforrejectionof thenullhypothesisstatedbeforeconductinganystatisticaltest.Asyoudecreasethelevel of significance,youreducethelikelihoodof creatinga typeI error. Asthepossibilityof makingthetypeI erroris lowered,theprobabilityof makinga typeII erroron theotherhandincrease. ThetypeII errorthat alsomeansthefalsenegativearises whenwefailto rejectthenullhypothesiswhenitis actuallya wrongstatement.Falsepositiveandfalsenegativetypicallyarisein a numberof binaryclassificationsmainlythoseto dowith medicaltesting.However,narrowlyspeaking,thetwo termsreferto statisticalhypothesistestingproceduresoutlinedin theNeyman-Pearsonframework.Asa matterof properscientificpracticeandstatistics, thelevel of significanceis supposedto be chosenbefore anydata collectionandanalysistakesplace.Mostmedicalstatisticaltestsalwayschosea verylowlevel of significance.Thisismainlydoneto decreasethecreatingof creatinga typeI error which is alwaysdetrimentalinthefieldof medicine.Thisimpliesthatthelevels like 95%, 99%, and 98% are frequentin themedicalfieldsas comparedto thesocial science field.

Question3 Testresultsare referred to as falsepositiveifthey areerroneous.The major reason for making such a conclusion mainly arises when theobtained results are different from what exists in real life orexpected results.Itmay representa conditionthat is presentbutdoesnot exist.Forexample,a falsepositiveresultmay be expressedin a givenparticulartestthat is designedto detectcancer.In such scenarios,thesubjectresultsofthe subject showthattheindividualhas cancerwhereasactuallythesubjectdoesnot havetheillness.Theparadoxof thefalsepositivein a numberof medicalandstatisticaltestsandresultsare verymuchmorelikelyto beobtainedthan theinstanceof therealpositiveresults.

Forinstance, whentheentirepopulationpossessesa lowincidenceof occurrenceof a givencondition,therateof theincidenceis alwayslowerthan thefalsepositiverate.Theoppositeof thefalsepositivetestis thefalsenegativeresult.Thelikelihoodof a givena positivetestresultis alwaysdeterminednot onlyby theaccuracyorpreciseness of thetestbutby thevariouscharacteristicsdepictedby thepopulationfrom which thesamplewasobtained. In situationswheretheincidenceandpercentageof theindividualspossessinga givenconditionis reportedto be lowerthan thestatedconditionunder study,theresultsare morelikelyto givea falsepositiveresulteventheresultsare not right.Itshould benotedthateventhestatisticalteststhat havea lowerlikelihoodof givingfalsepositiveresultsin a givenindividualcase.Theresultsmay giveresultsthat are moredishonestthan theoverall actualpositiveresults.Thisimpliesthan in anysocietypossessingveryfewpeopleinfectedby a diseasefewerpatientswill giveproportionately falsepositives comparedto theothers that reportotherwise.Whatthismeansis thatthere will be morepatientsreportingpositiveresulttestsforthegivendiseaseunder considerationincorrectly.

Thesepatientswill reportto havingthediseaseyetin actualsensetheydonot havethediseaseunder consideration.Thisstatisticalparadoxhas bothereda numberof researchers andscientistswhodealwith medicaltestsandotherresearchin relatedfields.Thisis synonymousto anotherparadoxthat could be referredto as thefalsenegativeparadox,a conditionin which a sickpatientwill givetestresultsthat revealthathe orsheisnot infectedwith thediseaseorconditionbeingtested.To minimize the incidence of the false positive and false negativeresults on anindividualthe tests could be donein a repetitivemanner.Therepetitionof thetestscan bedoneatfixedintervalsorcan be carriedout usingdifferent test reagentswhencarryingout thetest. In medicaltestingproceduresandmorein thebinaryclassificationscenarios,a falsepositivecould be an errorin thedata reportingsystem.Thiswouldarisewhena giventestresultwronglyindicatespresenceof a conditionthatdoesnot exist.Such a casewould beshownin thediagnosisof a diseasethatan individualis not necessarysufferingfrom.Afalsenegativeresult is a result inwhich a giventestresultimproperlyindicatesnopresenceof a givenconditionthatexistin reality.Falsepositiveandfalsenegativeare two majorkindsof errorsin a givenbinarytest.In medicine,theyare referredto as falsepositivediagnosisandfalsenegativediagnosisrespectively.Inaddition,in thestatisticalclassificationtheyare referredto as thefalsepositiveerrorandfalsenegativeerrorrespectively.Onwayto goover thescenariosof falsepositiveorfalsenegative,involveperforminga numberof testson a givenindividual.One wayto minimizetheincidenceof thetwo scenarioswill requiretheuseof differentpersonsto carryout thetestson thesubject.

Reference

Cutrer,W., &amp Glahn, S. (2002). Falsepositive.Colorado Springs: Water Brook Press.