New Semester
Started
Get
50% OFF
Study Help!
--h --m --s
Claim Now
Question Answers
Textbooks
Find textbooks, questions and answers
Oops, something went wrong!
Change your search query and then try again
S
Books
FREE
Study Help
Expert Questions
Accounting
General Management
Mathematics
Finance
Organizational Behaviour
Law
Physics
Operating System
Management Leadership
Sociology
Programming
Marketing
Database
Computer Network
Economics
Textbooks Solutions
Accounting
Managerial Accounting
Management Leadership
Cost Accounting
Statistics
Business Law
Corporate Finance
Finance
Economics
Auditing
Tutors
Online Tutors
Find a Tutor
Hire a Tutor
Become a Tutor
AI Tutor
AI Study Planner
NEW
Sell Books
Search
Search
Sign In
Register
study help
business
statistics informed decisions using data
Foundations Of Statistics For Data Scientists With R And Python 1st Edition Alan Agresti - Solutions
Refertotheanalysesoftheendometrialcancerdatain Section 7.3.2.(a) Explainthedifferencebetweentheinterpretationsoftheclassicalone-sided P-valueof 0.0011 andtheBayesianposterior P(β1 ≤ 0) = 0.0002 that weobtainusing σ =
Suppose y = 0 at x = 10, 20, 30, 40 and y = 1 at x = 60, 70, 80, 90.(a) WhatisthetrueMLestimate ˆβ for logisticregression?Reporttheestimateandstandard error providedbyyoursoftware.Whyisthereportedstandarderrorsolarge?(b) UseaBayesianmethodortheFirthpenalizedlikelihoodmethodtofitthemodel.Compare
Fortheclinicaltrialsdatamentionedin Section 7.2.6 of 10successesin10trialsforthenew drug and5successesin10trialsforplacebo,fitthelogisticregressionmodel.(a) Whatdoessoftwarereportfor ˆβ1? Whatistheactualvalue?(b) GivetheresultoftheWaldtestof H0: β1 = 0? Comparewiththeresultofthelikelihood-ratio
Comparisonof oddsratio with ratioofproportions:(a) Accordingto ourworldindata.org, in2017theproportionofthepopulationthatdied due tosuicidesdonebyfirearmswas0.00006291(i.e.,about6.29per100,000)inthe United Statesand0.00000168(i.e.,about0.17per100,000)intheUnitedKingdom.Use these tofindan oddsratio
“Standyourground”(SYG)lawsempowerindividualstouseanyforcenecessarytodefend themselvesagainstanyonetheybelievetobeanimminentthreat.Forcourttrialsinwhichthe defendantclaimedSYG,logisticregressionwasusedtoanalyzetheprobabilityofconviction.26 The
FortheItaliansurveyintroducedin Section 7.2.5, resultsforthe27,775subjectshavingage over65areinthe Employment2 data fileatthebook’swebsite.Sampleproportionsemployed at thevariouscombinationsofvaluesofexplanatoryvariablessuggestthatthegendereffect
Fittheinteractionmodelreferredtoin Section 7.2.5 for Italianemploymentinthe Employment data file,usinglogitandidentitylinkfunctions.Ineachcase,explainhowtointerpretthe estimated gendereffect.
Forfirst-degreemurderconvictionsinEastBatonRougeParish,Louisiana,between1990and 2008, thedeathpenaltywasgivenin3of25casesinwhichaWhitekilledaWhite,in0of3 cases inwhichaWhitekilledaBlack,in9of30casesinwhichaBlackkilledaWhite,and in 11of132casesinwhichaBlackkilledaBlack.25 Construct adatafile,letting
ForindictmentsincaseswithmultiplemurdersinFloridaduringa12-yearperiod,thedeath penaltywasgivenin53of467casesinwhichaWhitekilledaWhite,in0of16casesinwhich a WhitekilledaBlack,in11of48casesinwhichaBlackkilledaWhite,andin4of143cases in whichaBlackkilledaBlack.24(a)
The SoreThroat data fileatthebook’swebsitecontainsdatafromfromastudy23 about Y =whether apatienthavingsurgeryhadasorethroatonwaking(1 = yes,0 = no) asafunction of D = duration ofthesurgery(inminutes)and T = typeofdeviceusedtosecuretheairway(1 = trachealtube,0 = laryngeal maskairway).(a)
The Soybeans data fileatthebook’swebsiteshowsresultsofanexperiment22 to comparethe proportionsofsoybeanseedsthathavehealthyradicleofatleast2 mm after beingwrapped with wetgerminatingpaperundertreatmentAof25degreesCelsiusfor24hoursandunder
The Afterlife data fileatthebook’swebsiteshowsdatafromthe2018GeneralSocialSurvey on beliefintheafterlife(postlife = 1, yes;postlife = 2, no),religion(1 = Protestant,2 = Catholic, 3 = Jewish, othercategoriesexcluded),andgender(1 = male, 2 = female). Analyzethesedata
Usethefollowingtoydatatoillustrateeffectsofgroupedversusungroupedbinarydataonthe estimates andthedevianceforlogisticregressionmodeling:x Numberoftrials Numberofsuccesses 0 4 1 1 4 2 2 4 4 Denote by M0 the null modelandby M1 the modellogit(πi) = β0 + β1xi.(a)
AstudyofsheepmentionedinExercise1.27analyzedwhetherthesheepsurvivedforayear from theoriginalobservationtime(1 = yes,0 = no) asafunctionoftheirweight(kg) atthe original observation.(a) Doesthesurvivalofthesheepseemtodependontheirweight?Ifso,howdoestheweight of
Analogouslytothepreviousexercise,randomlysample30 X observationsfromauniformin the interval(-4,4)andconditionalon X = x, 30normalobservationswith E(Y ) = 3.5x3 −20x2 + 0.5x + 20 and σ = 30. FitpolynomialnormalGLMsoflowerandhigherorderthan that
Refertothesimulationin Section 7.1.6. Randomlysample6observationsfromthesameuni-form/normal structure.Fitthenullmodel,thestraight-linemodel,andthefifth-degreepoly-nomial. Foreachmodelfit,findSSEandfindthesumofsquarederrorsaroundthetruemeans.Explain whatthisshows.
FortheScottishhillracesdata(Section 6.1.4) inthe ScotsRaces data file,itisplausiblethat the recordtimesaremorevariableathighervaluesofdistanceandclimb.(a) Usingtheidentitylinkfunction,fitthegammaGLMtothewomen’srecordtimeswith distance
For the Houses data filedescribedin Section 7.1.3, consider Y = selling price, x1 = tax bill(in dollars), and x2 = whether thehouseisnew.(a) Formthescatterplotof y and x1. DoesthenormalGLMstructureofconstantvariability in y seem appropriate?Ifnot,howdoesitseemstobeviolated?(b)
Withquantitative x and y, itcanbeinformativetoplotthedataandshowasmoothingofthe relationship withoutassuminglinearityorsomeotherfunctionalform.WithanInternetsearch, read aboutthe lowess (locallyweightedscatterplotsmoothing)methodforthis.Explainthe method,withanexample,inareportofatmost300words.
Usingformula6.13,explainheuristicallywhyvar( ˆβj) tends todecreaseas n increases.
Showthemodelmatrixfor12observationsandthreeexplanatoryvariables,thefirstofwhichis quantitative,thesecondisbinary,andthethirdhasthreecategories,suchthateachcombination of thecategoricalvariableshastwoobservations.
Tosummarizetheinteractionbetweenquantitative x and binary z, supposewecenter x at its mean andexpressthemodelas E(Y ) = β0 + β1(x − μx) + β2z + β3(x − μx) × z.Explain howtointerpret β2. Howdoesthisdifferfromtheinterpretationforthemodelwithout centering?
Thelinearmodelfortheone-waylayoutgeneralizesfortwoormorefactors.Forthetwo-way r × c layoutwithfactors A and B:(a) Writethelinearmodelformulathatassumesalackofinteraction.Explainhowtocon-struct an F statistic fortestingthatthe A main effectsarezero.(b)
Derivethe F test in Section 6.5.6 for comparingnestednormallinearmodelsasalikelihood-ratio test,usinganargumentliketheonein Section 6.4.6 for theglobal F test.
Inconductingalargenumberofsignificancetests,the false discoveryrate (FDR) isthe expectedproportionoftherejectednullhypothesesthatareerroneouslyrejected.Asimple algorithm canensureFDR ≤ α when appliedwith t independenttestsandisespeciallyuseful when
ToshowBoole’sinequality(Section 6.5.4), let B1 = E1, B2 = Ec 1 ∩ E2, B3 = Ec 1 ∩ Ec 2 ∩ E3, ...Explain why P(∪jEj) = P(∪jBj) = Σj P(Bj) ≤ Σt j=1 P(Ej).
Illustratehowbetween-groupsandwithin-groupsvariabilityaffecttheresultoftheANOVA F test of H0: μ1 = μ2 = μ3 byconstructingtwoscenarios,eachwith n1 = n2 = n3 = 4 and y2 = 10, y2 = 12, y2 = 14, suchthatonescenario has arelativelylarge P-valueandonehasarelatively small P-value.36
Supposethatthepopulationcorrelation ρ is closeto1.(a) Whatshapewouldyouexpectthesamplingdistributionof r to have?Why?(b) In1915,R.A.Fishershowedthatwithsimplerandomsampling,has anapproximate N[T(ρ), 1~(n − 3)] distribution. Usingthis,explainhowtoconstruct a confidenceintervalfor ρ. = T(r)
(a)Explainintuitivelywhyapredictionintervalfor Y is muchwiderthanaconfidenceinterval for E(Y ).(b) Explainhowpredictionintervalsfor Y at variousvaluesof x wouldlikelybeinsubstantial error if(i)thevariabilityin Y tends toincreaseas x increases, (ii) Y is highlydiscrete, suchas Y =
Considertheconfidenceinterval(6.8)for E(Y ) in thebivariatemodel.(a) Showthatat x0 = x, theintervalis the sameastheinterval y ± tα~2,n−1(s~ºn) in the formula(4.10)forthe marginal distributionof y, exceptforthe t quantilescore.(b) At x0 ≠ x, explainwhyyou can estimate E(Y ) more preciselywhen
Forindependentrandomsamplesofsizes n1 and n2 from normalpopulationswithcommon variance σ2, using Section 4.4.6 and thedefinitionofthe F distribution in Section 6.4.1, explain why S2 1 ~S2 2 has an F distribution with df 1 = n1 − 1 and df 2 = n2 − 1.
Usingthedefinitionsof T and F random variables, showthatif T has a t distribution, then T2 has an F distribution with df = 1 and df2 equal df for the t distribution.
F statistics havealternateexpressionsintermsof R2 values.(a) Showthatfortesting H0: β1 = ⋯ = βp = 0,Explain whylargervaluesof R2 yield largervalues of F.(b) Showthatfor comparing nestedlinearmodels, F = (TSS-SSE)/p SSE/[n-(p+1)] R2/p is equivalently - (1 R2)/[n (p+1)]*
Usetheequationsin Section 6.2.6 that aresolvedtofindtheleastsquaresestimatestoshow that theleastsquaresfitofamultipleregressionmodelsatisfies 1 - 5) - . - 5) + . - .).
From Section 6.1.5 and Exercise6.31,thecorrelationisa standardizedslope. Whenwefita multipleregressionmodeltostandardizedversionsof y and theexplanatoryvariables,the estimated regressionparametersarecalled standardizedregressioncoefficients.(a) Explainhowtointerpretthestandardizedcoefficientfor xj
Usinghow r2 for thebivariatemodelisbasedonthemarginalvariabilityof y and theconditional variabilityof y given x, explainwhyforaparticulartrend,we’dexpect r2 (and SrS) tobelarger when therangeof x-valuessampledislarger.
Explainwhyconditionalindependencebetween Y and X1, given X2, doesnotimplymarginal independencebetween Y and X1. Toillustratethis,sketchahypotheticalscatterplotbetween y and x1 valuesatdifferentvaluesof x2.
Foruncorrelatedrandomvariables U, V , and W, let X = U + V and Y = U + W. (a) What happenstothecorrelationbetween X and Y when weconditionon U bykeepingitfixed;(b)Explain howaspuriousassociationcanoccur.
ForrecentU.S.presidentialelections,ineachstatewealthiervoterstendtobemorelikely to voteRepublican,yetstatesthatarewealthierinanaggregatesensetendtohavesmaller proportionsvotingfortheRepublican.35 Sketchaplotofstatewidevaluesof x = median income against y =
Astudyfoundthatchildrenwhoregularlyeatbreakfastgetbettermathgradesthanthose who donoteatbreakfast.Thisresultwasbasedontheassociationbetween X = whether eat breakfast (yes,no)and Y = grade inlastmathcoursetaken.Suggestapossiblelurkingvariable and
Aresearcharticlestatesthatsubjectswhoexerciseregularlyreportedonlyhalfasmanyserious illnesses peryear,ontheaverage,asthosewhodonotexerciseregularly.Theresultssection states, “Wenextanalyzedwhetheragewasaconfoundingvariableaffectingthisassociation.”(a)
Aten-yearstudy34 of elderlypeoplefoundthatthosewhoplayedatleastoneroundofgolf everyweekweremorelikelytobealiveadecadelater.Explainhowthiscouldbespuriousby describing potentialassociationsbetweenanothervariableandbothmortalityandfrequency of playinggolf.
Youcanfitthequadraticequation E(Y ) = β0+β1x+β2x2 byfittingamultipleregressionmodel with x1 = x and x2 = x2.(a) Simulate100independentobservationsfromthemodel Y = 40.0−5.0x+0.5x2+ϵ, where X has auniformdistributionover[0,10]and ϵ ∼ N(0, 1). Plotthedataandfitthequadratic
Forthemodel E(Yi) = β0 + β1xi, considertheexpressionfor ˆβ1 in equation(6.1).(a) Specifythesamplingdistributionof ˆβ1 and showthat E( ˆβ1) = β1 and var( ˆβ1) =σ2~[Σi(xi − x)2].(b) Fromthevariance formula,explainwhymorepreciseestimationof β1 is possiblewhenthe data
FortheScottishhillracesdatawith y = women’srecordtime,wecouldimposethenatural constraintthatwhen x = distance =0, E(Y ) = 0.(a) Forthemodel E(Yi) = βxi, derivetheleastsquaresestimateof β.(b) Fitthemodeltothedata.(In R, youcanuse lm(timeW ∼ -1
Section 6.1.5 showedthatasingleoutlyingScottishraceobservationinfluenceswhetherthe correlation betweenwomen’srecordtimeandclimbis0.69or0.85. Section 5.8.3 showedthat in comparinggroups,itcanbemorerobusttouseranks.Transformthetimeandclimbvalues to
Thedistributionof X = heights(cm) ofwomenintheU.K.isapproximately N(162, 72).Conditional on X = x, suppose Y = weight(kg) hasa N(3.0 + 0.40x, 82) distribution. Simulate a millionobservationsfromthisdistribution.Approximatethecorrelation.Approximatethe regression equationforpredicting X using y.
For Y = income (thousandsofdollars)and x = education (years),randomlygenerate n = 100 observationswhen X has auniformdistributionovertheintegers10to17and,conditionalon X = x, Y has a N(−25.0 + 6.0x, 152) distribution. Usesoftwaretofindtheleastsquaresfitof the
Let β1 denote theeffectof x in theregressionequationusing x to predict y and let β∗1 denote the effectof y in theregressionequationusing y to predict x. Showthatthecorrelation r is the geometricmean(Exercise1.44)of ˆβ1 and ˆβ∗1 .
Inbivariatemodeling,supposewetreat X also asarandomvariable,with (X,Y ) havinga bivariatenormaldistributionwithcorrelation ρ. Regressionanalysisestimatestheequation E(Y S X = x) = μY + ρ(σY ~σX)x − μX).(a) Statetheequationestimatedifweinsteaduse Y to predict X.(b) If E(Y ) = β0 + β1x
Forthe bivariate normal distribution (Section 2.7.4) withcorrelation ρ, in1886FrancisGalton showedthattheconditionaldistributionof (Y S X = x) is N[μy+ρ(σY ~σX)(x−μx), σ2 Y (1−ρ2)].Showthatand useittoexplain theconceptofregressiontowardthemean. E(YX = x)-Hy x-x 0x
Ahealthysystolicbloodpressureis120 mm Hg or less.Astudytookavolunteersampleof subjectswithhigh-bloodpressureandaskedthemtowalkfor30minutesatlunchtimeeach day.Afteronemonth,theirmeansystolicbloodpressurefellfrom150to140.Explainhowthis could merelyreflectregressiontowardthemean.
and sx = sy.Showthat (ˆμi − y) = 0.70(xi − x). Explainhowa score onthefinalexamispredictedto regress towarditsmeanrelativetothedistancebetweenthemidtermexamscoreandits mean.
Withleastsquaresforthebivariatelinearmodel,equation(6.1)statesthat ˆβ0 = y − ˆβ1x.(a) Usingthisformula,show thatanalternativeexpressionforthefittedlineis (ˆμi − y) =ˆβ1(xi − x).(b) Fordataon y = final exam score and x = midterm score,suppose r =
Whenobservationsonboth x and y are standardized,showthatthebivariatepredictionequa-tion hastheform ˆμi = rxi, where r is thesamplecorrelation.
Whenthevaluesof y are multipliedbyaconstantc, fromtheirformulas,showthat sy and ˆβ1 in thebivariatelinearmodelarealsothenmultipliedbyc. Thus,showthat r = ˆβ1(sx~sy) does not dependontheunitsofmeasurement.
Analternativeparameterizationofthebivariateregressionmodelcenters x around itsmean and expressesthemodelas E(Y ) = β0 + β1(x − μx).(a) Forthisparameterization,explainhowtointerpret β0. Howdoesthisdifferfromthe interpretationforthemodelwithoutcentering?(b) Ifyoureplace μx in thisexpressionby
Showthatthecorrelation betweenavariableanditselfisnecessarily1.0.
Showthatforthenullmodel E(Yi) = β0, leastsquaresyields ˆβ0 = y.
ThestatisticianGeorgeBox,whohadanillustriousacademiccareerattheUniversityofWis-consin, isoftenquotedassaying,“Allmodelsarewrong,butsomemodelsareuseful.”Whydo youthinkthat,inpractice, (a) all modelsarewrong, (b) some modelsare not useful.
TheInternetsite www.artofstat.com/web-apps has a Guess theCorrelation app. Playthe CorrelationGame 10 times.Showthetableofguessesandactualvaluesandfindthecorrelation betweenthem.
Constructascatterplotwith5pointsthathaveacorrelationcloseto +1, andthenaddasingle pointthatchangesthecorrelationtoastrongnegativevalue.33 Explain whythissingleoutlying observationhassomuchinfluence.
The Hare data fileatthebook’swebsite32 has datafor550haresonbodymass(grams)and hind footlength(mm), bygenderofthehare.Useregressionmodelswithbodymassasthe responsevariable,interpretingcarefullyanyeffectsorinteractions.Prepareareportofatmost 300
for thepriordistributionfor β0. Showthatshrinkingβ0 far fromtheleastsquaresestimateof28.2toward0alsoforcesdramaticchangesin the otherBayesianposteriormeanestimates.(c) With τ = 10 and similarresultsaswithleastsquares,explaindifferencesbetweenthe interpretationsoftheBayesianposterior P(β1 ≤ 0)
for theeffectsoflife eventsandSESandcompareresultswithhighlydispersenormalpriorshaving τ = 10.0.Explain howtheposteriormeansfor β1 and β2 dependon τ .(b) In(a),supposeyoualsoused τ =
for β1 and β2 (and thusprecision 1~(0.20)2 = 25) yethaveahighlydispersepriorfor β0 (e.g., precision = 10−10), wecan specifya3×3 matrix A to usefortheprecision B0, wherethevaluesonthemaindiagonal are theprecisionsfor (β0, β1, β2):> A=diag(c(10^(-10),25,25))> fit.bayes
Refertothementalimpairmentexamplein Sections 6.4.2 and 6.6.2.(a) Theexamplein Section 6.6.2 used highlydispersepriors.Let’scompareresultstoin-formativepriorsbasedonsubjectivebeliefsthat β1 and β2 are notvarylarge.With the MCMCregress function in R,
FortheScottishhillsracesdata,modeltherecordtimesimultaneouslyformenandwomenby using distance,climb,andgenderasexplanatoryvariables.Prepareareportofatmost300 words,includinghighlyeditedsoftwareoutputinanappendix.Forthefinalmodelchosen,in-terpret parameterestimates,findandinterpret R2,
UselinearmodelingtoanalyzetherecordtimesformenintheScottishhillraces,withdistance and climbasexplanatoryvariables,(a) withleastsquares,(b) withBayesianmethods.Prepare a report,includinghighlyeditedsoftwareoutputinanappendix.
ConductBayesianfittingofthelinearmodelfortheScottishhillraces,usingclimbanddis-tance predictorsofwomen’srecordtimeswithouttheoutlyingHighlandFlingraceobservation.Compare theposteriormeanestimatesoftheparameterstotheleastsquaresestimates,asyou varythespreadofthenormalpriordistributions.
Apharmaceuticalclinicaltrial31 randomly assigned24patientstothreetreatmentgroups(drug A, drugC,placebo)andcomparedthemonameasureofrespiratoryability(FEV1 = forced expiratory volumein1second,inliters)after1houroftreatment.Forthe FEV data fileat the book’swebsite,fitthelinearmodelfor fev with
For72young girls sufferingfromanorexia,the Anorexia.dat file atthebook’swebsiteshows their weightsbeforeandafteranexperimentalperiod.Thegirlswererandomlyassignedto receivecognitivebehavioraltherapy(cb). familytherapy(f ), orbeinacontrolgroup(c).(a)
Refertothepreviousexample.Thedataexhibitastrongpositivecorrelationof0.84between selling priceandthetaxbill.Mightthisbeaspuriousassociation,withthesizeofthehome causally affectingbothofthem?Sketchafiguretoreflectthepotentialspuriousrelationship.Analyze
912 11732 1 no 3 365.55165430764 2 no(a) Summarizethedatawithascatterplot between y and x1, usingseparatesymbolsforthe twocategoriesof x2.(b) Fitthemodel E(Yi) = β0 + β1xi1 + β2xi2. Interprettheestimatedeffects.(c) Reportandinterpret R2 for summarizingtheglobalpredictivepower.(d)
Table6.5 showsobservationsonhomesalesinGainesville,Florida,fromthe Houses data file at thebook’swebsitefor100homes.Variableslistedaresellingprice(thousandsofdollars), size ofhouse(squarefeet),annualpropertytaxbill(dollars),numberofbedrooms,numberof
1136.3607
0.03411.651 Anova(lm(SpermTotal~CW+factor(Color) + CW:factor(Color),data=Crabs2))Anova Table(TypeIItests)Sum SqDfFvaluePr(>F)CW
cm), whichisameasureofitssize,and color (1 = dark, 2 = medium, 3 = light),whichisa measure ofadultage,darkeronesbeingolder.> summary(lm(SpermTotal~CW+factor(Color), data=Crabs2))Estimate Std.ErrortvaluePr(>|t|)(Intercept) 11.3660.63817.822
and s = 2.0. Thetwo explanatory variablesusedin the R output arethehorseshoecrab’s carapacewidth (CW, mean18.6 cm, standarddeviation
Thedataset30 Crabs2 at thebook’swebsitecomesfromastudyoffactorsthataffectsperm traits ofmalehorseshoecrabs.Aresponsevariable, SpermTotal, isthelogofthetotalnumber of sperminanejaculate.Ithas y =
Forthe Polid data filesummarizedin Table5.2, conductanANOVAtoanalyzewhethermean politicalideologyvariesbyrace.Useafollow-upmultiplecomparisonmethodwithoverall confidence level0.95toestimatedifferencesofmeansbetweenpairsofracesonpoliticalideology.Interpretresults.
Forthe UN data fileatthebook’swebsite(seeExercise1.24),constructamultipleregression modelpredictingInternetusingalltheothervariables.Usetheconceptofmulticollinearity to explainwhyadjusted R2 is notdramaticallygreaterthanwhenGDPisthesolepredictor.Compare
levelifweusetheBonferroniapproachto test thefamilyofthreeindividualeffects?Explain.(d) Aretheeffectsof tv and sport significant?Proposeanalternativemodel.
RefertothemodelfittedinthepreviousexercisetopredictcollegeGPA.(a) Test H0: β1 = β2 = β3 = 0. Reportthe P-valueandinterpret.(b) Showhowtoconductasignificancetestabouttheindividualeffectof hsgpa, adjusting for tv and sport, using α = 0.05. Interpret.(c) Istheeffectin(b)significantatthe α =
The Students data fileshowsresponsesonvariablessummarizedinExercise1.2.(a) Fitthelinearmodelusing hsgpa = high schoolGPA, tv = weeklyhourswatchingTV,and sport = weeklyhoursparticipatinginsportsaspredictorsof cogpa = college GPA.Report the predictionequation.Whatdothe P-valuessuggest?(b)
Exercise1.49gavealinkto2020U.S.statewidedataon x = percentageofpeoplewearing masks inpublicand y = percentageofpeoplewhoknowsomeonewithCovid-19symptoms.Interpretthevalueof r2 for thedata,whichis0.724,andreportthecorrelation.
Refertotheexample in Section 6.2.5 of thecrimerateinFloridacounties.(a) Explainwhatitmeanswhenwesaythesedataexhibit Simpson’sparadox. Whatcould cause thischangeinthedirectionoftheassociationbetweencrimerateandeducation when weadjustforurbanization?(b) Usingthe Florida data
Forthemodelpermittinginteractionbetweendistanceandclimbintheireffectonwomen’s record timesfortheScottishhillraces,analyzewhetheranyobservationisinfluentialinthe least squaresfit.Howdoestheparameterestimatefortheinteractiontermchangewhenyou analyze thedatawithoutit?
Refertothestudyofmentalimpairmentin Section 6.4.2.(a) Sketchafiguretoillustrateanassociationbetweenimpairmentandlifeeventsthatis spurious, explainbySES.Fittwomodelstoinvestigatewhetherthisisthecase,and summarize.(b) Section 6.4.2
Forthe Covid19 data fileatthetextwebsite:(a) Constructthetwoscatterplotsshownin Figure 6.3.(b) Findandinterpretthecorrelationbetweentimeand(i)cases,(ii) log(cases).(c) Fitthelinearmodelforthelog-transformedcountsandreportthepredictionequation.29 Explain whythepredictedcountatday x+1 equals
Inthe2000PresidentialelectionintheU.S.withDemocraticcandidateAlGoreandRepublican candidate GeorgeW.Bush,somepoliticalanalyststhoughtthatmostofthevotesinPalm BeachCounty,Florida,fortheReformpartycandidate,PatBuchanan,mayhaveactually
The Firearms2 data fileatthetextwebsiteshowsU.S.statewidedataon x = percentage of peoplewhoreportowningagunand y = firearm deathrate(annualnumberofdeaths per100,000population),from www.cdc.gov. Identifyapotentiallyinfluentialobservationfrom the
Foradvancedindustrializednations,the Firearms data fileatthetextwebsiteshowsannual homicide rates(permillionpopulation)andthenumberoffirearms(per100people),withdata takenfromWikipediaand smallarmssurvey.org.(a) Constructascatterplotandhighlightanyobservationsthatfallapartfromthegeneral trend.(b)
For theScottishhillracesdata,alinearmodelcanpredictmen’srecordtimesfromwomen’s record times.(a) Showthescatterplotandreportthepredictionequation.Predictthemen’srecordtime for theHighlandFling,forwhichtimeW=490.05minutes.(b) Findandinterpretthecorrelation.(c)
Afamilyofdistributions f(y; θ) is saidtohave monotone likelihoodratio (MLR) ifastatistic T(y) exists suchthatwhenever θ′ < θ, ℓ(θ)~ℓ(θ′) is anondecreasingfunctionof T. Forany suchfamily,themostpowerfultestof H0: θ = θ0 against Ha: θ > θ0 forms P-valuesfromvalues of T at
Foradiscretedistributionandateststatistic T with observedvalue tobs and one-sided Ha suchthatlarge T contradicts H0, mid P-value = 1 2P(T = tobs) + P(T > tobs).(a) Supposethat P(T = tj) = πj , j = 1, . ..c. Showthat E(mid P-value) = 0.50. (Hint: Show that Σj πj[(πj~2) + πj+1 + ⋯+ πc] =
find the P-valueforeachsample.Plotahistogramordensityestimate using the1,000,000 P-values.Relateittotheresultin(a).
is true.Generate1,000,000randomsamples,eachofsize n = 1500, and for Ha: π >
From Section 2.5.7, if T is acontinuousrandomvariablewith cdf F, then F(T) has theuniform distribution over[0,1].(a) Insignificancetestingwithateststatistic T, explainwhy F(T) and 1−F(T) correspond to one-sided P-values.Explainwhy 1 − F(T) also hasauniformdistributionunder
Explainwhatismeantby censored data. Giveanexampleofasituationinwhichsomeobser-vationswouldbe(a) right-censored,(b) left-censored.
Showing 2500 - 2600
of 5564
First
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Last
Step by Step Answers