Skip to main content

Statistics Unlocking The Power Of Data 4Th Lock Solutions Manual

Page 1


ISBN: 9781394336623

Section1.1Solutions

1.1 (a)Thecasesarethepeoplewhoareaskedthequestion.

(b)Thevariableiswhethereachpersonsupportsthelawornot.Itiscategorical.

1.2 (a)Thecasesarethe100stocks.

(b)Thevariableisthepercentagechange,whichisanumericalquantity,foreachofthestocks.Itis quantitative.

1.3 (a)Thecasesaretheteenagersinthesample.

(b)Thevariableistheresult(yesorno)indicatingwhethereachteenagereatsatleastfiveservingsaday offruitsandvegetables.Itiscategorical.

1.4 (a)Thecasesarethebunchesofbananasinthesample.

(b)Thevariableisthenumberofdaysuntilthebananasgobad.Itisquantitative.

1.5 (a)The10beamsthatweretested.

(b)Theforceatwhicheachbeambroke.Itisquantitative.

1.6 (a)Thecasesarecountriesoftheworld.

(b)Thevariableiswhetherornottheliteracyrateisover75%.Itiscategorical.

1.7 Sinceweexpectthenumberofyearssmokingcigarettestoimpactlungcapacity,wethinkofthenumber ofyearssmokingastheexplanatoryvariableandthelungcapacityastheresponsevariable.

1.8 Sinceweexpecttheamountoffertilizerusedtoimpacttheyield(andnottheotherwayaround),we thinkoftheamountoffertilizerastheexplanatoryvariableandtheyieldofthecropastheresponsevariable.

1.9 Ingestingmorealcoholicdrinkswillcausethelevelofalcoholinthebloodtoincrease,sothenumberof drinksistheexplanatoryvariableandbloodalcoholcontentistheresponse.

1.10 Theworldrecordtimewillcontinuetodecreaseastheyearsgobysoweexpecttheyeartoimpact marathonrecordtime.Wethinkoftheyearastheexplanatoryvariableandtherecordtimeastheresponse variable.

1.11 (a) Year and HigherSAT arecategorical.Theothersixvariablesareallquantitative,although Siblings mightbeclassifiedaseithercategoricalorquantitative.

(b)Therearemanypossibleanswers,suchas“Whatproportionofthestudentsarefirstyearstudents?” or“Whatistheaverageweightofthesestudents?”

(c)Therearemanypossibleanswers,suchas“Doseniorsseemtoweighmorethanfirstyearstudents?” or“DostudentswithhighVerbalSATscoresseemtoalsohavehighMathSATscores?”

1.12 (a)Inadditiontotheidentificationcolumn, Country,thereare24variables.Weseethat Developed isacategoricalvariable,whiletheother23variablesareallquantitative.

(b)Therearemanypossibleanswers,suchas“Whatistheaveragelifeexpectancyforallcountriesofthe world?”or“Whatproportionofcountriesaredeveloped?”

(c)Therearemanypossibleanswers,suchas“Docountrieswithagreaterlandareahavealargerpercent rural?”or“Docountriesthatspendarelativelylargeamountonthemilitaryspendarelativelysmall amountonhealthcare?”or“Dodevelopedcountrieshavealongerlifeexpectancythandeveloping countries?”

1.13 (a)Theexplanatoryvariableisthetypeofformatinwhichthestoryispresented.Itiscategorical, withthreecategories(audio,illustrated,andanimated).

(b)Theresponsevariableisthemeasureofbrainconnectivity.Itisquantitative.

(c)Thecasesarethefour-year-oldsinthestudy,sothereare27cases.

1.14 (a)ThecasesarefemalegamersinGreatBritain.

(b)Therearethreevariablesmentioned:Whetherthegamershadreceivedobscenemessages(categorical), howmanyhoursaweektheyplayed(quantitative),andwhethertheyfelttherewereenoughstrong femalecharactersingames(categorical).

(c)Thereare1151casesand3variables,sothedatasetwillhave1151rowsand3columns.

1.15 (a)Thecasesarethestudentsinthecollegephysicsclass.

(b)Thereare3variables:Whetherthestudentwasassignedtoanactiveorpassivelearningclass,the measureofhowmuchthestudentthoughttheylearned,andthescoreofactuallearningonthetest. Theclassassignmentiscategorical,whiletheothertwoarequantitative.

(c)Theclassassignment(activeorpassive)istheexplanatoryvariable,whilethetwomeasuresoflearning areresponsevariables.

(d)Thestudentsinthephysicsclassarethecases,sothereare154+142=296cases.Therearethree variables,sothedatasetwillhave296rowsand3columns.

1.16 Thereareatleasttwovariables.Onevariableiswhetherornotthespiderengagedinmock-sex.This variableiscategoricalandtheexplanatoryvariable.Anothervariableislengthoftimetoreachthepointof realmatingoncethespiderisfullymature.Thisvariableisquantitativeandtheresponsevariable.

1.17 Theindividualcasesarethelakesfromwhichwatersamplesweretaken.Foreachlakeinthesample, werecordtheconcentrationofestrogeninthewaterandthefertilityleveloffish.Botharequantitative variables.

1.18 Therearetwovariables.Onevariableindicatesthepresenceorabsenceofthegenevariantandthe secondvariableindicateswhichofthethreeethnicgroupstheindividualbelongsto.Bothvariablesare categorical.

1.19 (a)Thereare10cases,correspondingtothe10cities.Thetwovariablesarepopulation,whichis quantitative,andthehemispherethecityisin,whichiscategorical.

(b)Weneedtwocolumns,oneforeachvariable.Thecolumnscanbeineitherorder.Seethetable.

1.20 (a)Thereare7cases,representingthesevenpigeons.Therearetwovariables.Oneisthesexofthe pigeon,whichiscategorical,andtheotheristhespeedofthepigeon,whichisquantitative.

(b)Thedatasetwillhave7rowsand2columns.Seethetable.(Thesevencasescanbelistedinany order.)

Sex Speed

Hen 1676

Hen 1452

Hen 1449

Cock 1458

Cock 1435

Cock 1418

Cock 1413

1.21 (a)Thecasesarethehomingpigeons,sothereare1412ofthem.

(b)Thereare4variables.Twoarecategorical(loft,sex)andtwoarequantitative(distance,speed).

(c)Thedatasetwillhave1412rows(oneforeachpigeon)and4columns(oneforeachvariable).

1.22 Onevariableiswhethereachmalewasfedahigh-fatdietoranormaldiet.Thisistheexplanatory variableanditiscategorical.Theresponsevariableiswhetherornotthedaughtersdevelopedmetabolic syndrome,whichisalsocategorical.

1.23 Onevariableiswhethertheyoungfemalemicelivedinanenrichedenvironmentornot.Thisisthe explanatoryvariableanditiscategorical.Theresponsevariableishowfasttheoffspringlearnedtonavigate mazesandisquantitative.

1.24 (a)Themodeoftransportation(e.g.,bus,car,walk)givesacategoricalvalueforeachstudent.

(b)Theansweraboutallergies(yesorno)givesacategoricalvalueforeachstudent.

(c)Theproportionofstudentsinthesamplewhoarevegetariansisasinglenumber,basedonallofthe students,anddoesnotgiveavalueforeachindividualstudent.Itisnotavariableforthedataset.

(d)Thenumberofhoursworkedatapaidjobgivesaquantitativevalueforeachstudent.

(e)Thedifferenceinhoursofsleeponschoolnightsandnon-schoolnightsgivesaquantitativevaluefor eachstudent.

(f)Themaximumtimetogettoschoolisasinglenumber,basedonallofthestudents,anddoesnotgive avalueforeachindividualstudent.Itisnotavariableforthedataset.

(g)Thedesiredsuperpowerchosenfromalistgivesacategoricalvalueforeachstudent.

1.25 (a)Thetotaltuitionandfeesisaquantitativevalueforeachschool.

(b)ThenumberofschoolsintheNortheastisasinglevalue,basedonalloftheschools,anddoesnotgive avalueforeachindividualschool.Itisnotavariableforthedataset.

(c)Thetypeofschool(public,private,forprofit)isacategoricalvalueforeachschool.

(d)Thenumberofundergraduatesisaquantitativevalueforeachschool.

(e)Thepercentageofpart-timestudentsisaquantitativevalueforeachschool.

(f)Theschoolwiththehighestaveragefacultysalaryisasinglevalue,basedonalloftheschools,and doesnotgiveavalueforeachindividualschool.Itisnotavariableforthedataset.

1.26 Inthefirststudy,thecasesarethestudents.Theonlyvariableiswhetherornotthestudenthas smokedahookah.Thisisacategoricalvariable.

Inthesecondstudy,thecasesarethepeopleinahookahbar.Thevariablesarethelengthofthesession, thefrequencyofpuffing,andthedepthofinhalation.Allarequantitative.

Inthethirdstudy,thecasesarethesmokesamples,andthevariablesaretheamountoftar,nicotine,and heavymetals.Allthreevariablesarequantitative.

1.27 (a)Thisdescriptionofthestudymentionssixvariables:age,nosevolume,nosesurfacearea,nose height,nosewidth,andgender.

(b)Oneofthevariables(gender)iscategorical,andtheotherfivearequantitative.

(c)Therearesixvariablessothedatasetwillhavesixcolumns.The859participantsarethecases,sothe datasetwillhave859rows.

1.28 (a)Thecasesarethe47participants.

(b)Thedescriptionofthestudyincludesthreedifferentvariables:thescoreontheno-distractionstest, thescoreonthetestwhiletexting,andwhetherornotthestudentconsideredhimorherselftobe goodatmultitasking.Thetwotestscorevariablesarequantitativeandthemultitaskingvariableis categorical.

(c)Thedatasetwouldhave47rows(oneforeachparticipant)andthreecolumns(oneforeachofthethree variables).

1.29 (a)Thecasesarethe40peoplewithinsomniawhowereincludedinthestudy.

(b)Therearetwovariables.Oneiswhichgroupthepersonisassignedto,eithertherapyornot,andthe otheriswhetherornotthepersonreportedsleepimprovements.Botharecategorical.

(c)Thedatasetwouldhavetwocolumns,oneforeachofthetwovariables,and40rows,oneforeachof thepeopleinthestudy.

1.30 Ifwesimplyrecordageinyearsandincomeindollars,thevariablesarequantitative.Often,however, inasurvey,wedon’taskfortheexactagebutratherwhatagecategorytheparticipantfallsin(20 29, 30 39,etc.).Similarly,weoftendon’taskforexactincomebutforanincomecategory(lessthan$10,000, between$10,000and$25,000,etc.).Ifweaskparticipantswhatcategorytheyareinforeachvariable,then thevariablesarecategorical.

1.31 Wecouldsamplepeopleeligibletovoteandaskthemeachtheirpoliticalpartyandwhethertheyvoted inthelastelection.Thecaseswouldbepeopleeligibletovotethatwecollectdatafrom.Thevariables wouldbepoliticalpartyandwhetherornotthepersonvotedinthelastelection.Alternatively,wecould askwhethereachpersonplanstovoteinanupcomingelection.

1.32 Wecouldsurveyasampleofpeopleandasktheirhouseholdincomeandmeasurehappinessinsome way,suchasaskinghowhappytheyareonascaleof1–10.Thecaseswouldbethepeoplewecollectdata from.Thevariablesinthiscasewouldbehouseholdincomeandhappinessrating,althoughanytwovariables measuringwealthandhappinessarepossible.

1.33 Answerswillvary.

Section1.2Solutions

1.34 Thisisasample,becauseonlyasubsetoffisharemeasured.

1.35 Thisisapopulation,becauseallcustomersareaccountedfor.

1.36 Thisisapopulation,becauseallregisteredvehiclesareaccountedfor.

1.37 Thisisasample,becauseonlyasubsetofcollegestudentsweresentthequestionnaire.

1.38 Thesampleisthe120peopleinterviewed.Thepopulationmightbeallpeopleinthattownorall peoplethatgotothemallinthattownoravarietyofothergroupslargerthanandcontainingthe120 peopleinthesample.

1.39 ThesampleisthefivehundredCanadianadultsthatwereaskedthequestion;thepopulationisall Canadianadults.

1.40 Thesampleisthe100customerssurveyed;thepopulationisallcustomersofthecellphonecarrier.

1.41 Thesampleisthe1000householdswhichhavedataboxesattachedtothetelevisions.Thepopulation isallUShouseholdswithtelevisions.

1.42 (a)Thesampleisthe100collegestudentswhowereaskedthequestion.

(b)ThepopulationweareinterestedinisallAmericans.

(c)Apopulationwecangeneralizeto,givenoursample,iscollegestudents.

1.43 (a)Thesampleisthe10selectedtwitteraccounts.

(b)Thetargetpopulationisalltwitteraccounts.

(c)Thepopulationwecangeneralizeto,giventhesample,isonlytwitteraccountsofthisauthor’sfollowers, sincethisisthepopulationfromwhichthesamplewasdrawn.

1.44 (a)Thesampleisthe1500peoplewhowerecontacted.

(b)ThepopulationweareinterestedinisallresidentsoftheUS.

(c)Apopulationwecangeneralizeto,givenoursample,isresidentsofMinnesota.

1.45 (a)Thesampleisthegirlswhoareontheselectedbasketballteams.

(b)Thepopulationweareinterestedinisallfemalehighschoolstudents.

(c)Apopulationwecangeneralizeto,givenoursample,isfemalehighschoolstudentswhoareona basketballteam.

1.46 Yes,thisisarandomsamplefromthepopulation.

1.47 Yes,thisisrandomsamplefromthepopulation.

1.48 No,thisisnotarandomsample,becausesomeemployeesmaybemorelikelythanotherstoactually completethesurvey.

1.49 No,thisisnotarandomsample,becausecertainsegmentsofthepopulation(e.g.,thosenotattending college)cannotbeselected.

1.50 No,thisisnotarandomsample.Wemightthinkwecanpickouta“representativesample”,butwe probablycan’t.Weneedtoletarandomnumbergeneratordoitforus.

1.51 No,thisisnotarandomsample,thisisavolunteersample,sincetheonlypeopleinthesampleare thosethatself-selecttorespondtotheonlinepoll.

1.52 ThissampleisdefinitelybiasedbecauseonlystudentswhoareatthelibraryonaFridaynightcanbe selected.Therandomsampleshouldbefromallstudents.

1.53 Thisisbiasedbecausethewaythequestioniswordedisnotatallobjective.Althoughthesampleis arandomsample,thewordingbiasmaydistorttheresults.

1.54 Thissampleisbiasedbecausetaking10applesoffthetopisnotarandomsample.Theapplesonthe bottomofthetruckloadareprobablymorelikelytobebruised.

1.55 Fromthedescription,itappearsthatthismethodofdatacollectionisnotbiased.

1.56 Thissampleisbiasedbecauseitisavolunteersurveyinwhichpeoplechoosetoparticipateornot. Mostlikely,thepeopletakingthetimetorespondtotheemailwillhavestrongeropinionsthantherestof thestudentbody.

1.57 BecausethiswasarandomsampleofparentsinKansasCity,theresultcanbegeneralizedtoallparents inKansasCity.

1.58 (a)No,thesampleisalmostcertainlynotrepresentative,sinceitisavolunteersampleandonly includespeoplewhovisitthatwebsiteandwhochosetoparticipateinthepoll.

(b)No,itisnotappropriatetogeneralizesincethesampleisnotrepresentative.

1.59 (a)Yes,thesampleislikelytoberepresentativesinceitisarandomsample.

(b)Yes,sincethesampleisarandomsample,wecangeneralizetothepopulationofallCanadianconsumers.

1.60 (a)Thesampleisthe1, 000USadultsthatwerecontacted.TheintendedpopulationisallUSadults.

(b)Yes,itisreasonabletogeneralizesincethesamplewasselectedrandomly.

1.61 (a)Thesampleisthe800peoplewhoparticipatedinthesurvey.TheintendedpopulationisallUS smartphoneusers.

(b)Thecasesarethepeoplewhoparticipatedinthesurvey.Therearetwovariables:Whetherornota fooddeliveryappwasusedinthelastmonth,andwhichapp(ifany)wasused.Botharecategorical.

1.62 (a)The457studentsinthesampleisnotalargerpopulation.Wewouldknowtheexactanswersfor thisgroupsononeedtogeneralize.

(b)ThesamplewasrandomlyselectedfromamongallPennsylvaniahighschoolseniorswhoparticipated intheCensusatSchoolproject,sothatwouldbeareasonablepopulationtogeneralizeresultsto.

(c)Theremightbesomethingspecialaboutschools(orstudents)whoparticipateintheCensusatSchool project,soitmightnotbereasonabletogeneralizetheresultsfromthatsampletoallPennsylvania highschoolseniors.

(d)ItwouldnotbereasonabletogeneralizefromasampleofonlyPennsylvaniastudentstostudentsfrom allstates.

1.63 (a)Thereisnosamplingbiashere,sincewearetoldthatarandomsamplewasused.Theresults arebeingmanipulatedbasedonlyontheorderoftheoptions(thewordingofthequestion)sothisis anexampleofwordingbias.

(b)Weseethatpeoplearemorelikelytoselectthefirstoptiongiven,soyoushouldaskyourfriendto choosebetween“Option W andOption Q,“withOption W presentedfirst.

1.64 (a)Yes,thereisprobablysamplingbiassincethegrouphasnotsharedthemethod.No,wecannot knowifitisappropriatetogeneralize,anditisprobablynotappropriatetodoso.Thesurveymight havebeenaskedofchildren,forexample.Thiswouldcertainlycreatesamplingbias.

(b)Yes,thewaythequestionwaswordedalmostcertainlybiasedtheresults.Abetterwaymightbeto askanopen-endedquestionsuchasdescribehowchocolatemilkismade,

1.65 (a)Theindividualcasesaretheover6000restroompatronswhowereobserved.Thedescription makesitclearthatatleastthreevariablesarerecorded.Oneiswhetherornotthepersonwashed theirhands,anotheristhegenderoftheindividual,andathirdisthelocationoftheobservation.All threearecategorical.

(b)Inaphonesurvey,peoplearelikelytorepresentthemselvesinthebestlightandnotalwaysgive completelyhonestanswers.Thatiswhyitisimportanttoalsofindotherwaysofcollectingdata,such asthismethodofobservingpeople’sactualhabitsintherestroom.

1.66 (a)Thesampleisthesurveyparticipants,thepopulationisallprofessorsattheUniversityofNebraska.

(b)No,wecannotconcludethatthesampleofsurveyrespondersisnotrepresentativeofprofessorsatthe UniversityofNebraskasincewearenotgivenenoughinformationtodecideonewayortheother.

(c)No,the94%isbasedonselfdescriptions,whichcanbe(andinthiscase,probablyare)biased.

1.67 No.Thisisavolunteersample,andthereisreasontobelievetheparticipantsarenotrepresentativeof thepopulation.Forexample,somemaychoosetoparticipatebecausetheyLIKEalcoholand/ormarijuana, andthoseinthesamplemaytendtohavemoreexperiencewiththesesubstancesthantheoverallpopulation. Inaddition,theadvertisementsforthestudywereairedonrockradiostationsinSydney,soonlythosepeople wholistentorockradiostationsinSydneywouldhearabouttheoptiontoparticipate.

1.68 Yes!Thesampleisarandomsamplesowecanbequiteconfidentthatitisprobablyarepresentative sample.

1.69 (a)Thisisnotasimplerandomsamplefromthepopulation,sinceonlythosewhosawandwanted toclickandcompletethesurveywereincluded.

(b)Theseresultscouldalsohavebeenbiasedbyhowthesurveywasconstructed.Thewordingofthe questionsmightalsointroducebias.

1.70 Thestudygivenfoundarelationshipinasampleofrats.Thisrelationshipmaynotgeneralizetothe humanpopulation.

1.71 Thesampleofplanesthatreturnfrombombingmissionswasbiased.Morebulletholeswerefoundin thewingsandtailbecauseplanesthatwereshotinotherregionsweremorelikelytocrashandnotreturn.

1.72 (a)ThepopulationintheCPSisallUSresidents.(Alsoacceptable:UScitizens,UShouseholds...) (b)ThepopulationintheCESsurveyisallnon-farmbusinessesandgovernmentagenciesintheUS.

(c)i.TheCESsurveywouldbemorerelevant,becausethequestionpertainstocompanies.

ii.TheCPSwouldbemorerelevant,becausethequestionpertainstoAmericanpeople.

iii.TheCPSwouldbemorerelevant,becausethequestionpertainstopeople,notbusinesses.

1.73 (a)SincetheNHANESsampleisdrawnfromallpeopleintheUS,thatisthepopulationwecan generalizeto.

(b)SincetheNHAMCSsampleisdrawnfrompatientsinemergencyroomsintheUS,wecangeneralize theresultstoallemergencyroompatientsintheUS.

(c)i.NHANES:Thequestionaboutanassociationbetweenbeingoverweightanddevelopingdiabetes appliestoallpeopleintheUS,notjustthosewhovisitanemergencyroom.

ii.NHAMCS:Thisquestionasksspecificallyaboutthetypeofinjuryforpeoplewhogotoan emergencyroom.

iii.NHAMCS:Thisquestionofaveragewaitingtimeonlyappliestoemergencyroompatients.

iv.NHANES:ThisquestionisaskingaboutallUSresidents.Notethattheproportionwouldbe equaltooneforthepeoplesampledinNAMCSsincetheyonlygetintothesampleiftheyvisit anemergencyroom!

1.74 Answerswillvary.Seethetechnologynotestoseehowtousespecifictechnologytoselectarandom sample.

1.75 Answerswillvary.Seethetechnologynotestoseehowtousespecifictechnologytoselectarandom sample.

Section1.3Solutions

1.76 Theuseof“improves”impliesthisisacausalassociation.

1.77 Since“nolinkisfound”thereisneitherassociationnorcausation.

1.78 Thephrase“leadstodeaths”indicatesacausalassociation.

1.79 Thephrase“morelikely”indicatesanassociation,butthereisnoclaimthatwealth causes peopleto lie,cheatorsteal.

1.80 Thephrase“tendtobemoreeducated”indicatesanassociation,butthereisnoclaimthatowninga catcausesmoreeducation(orthatbettereducationcausespeopletoprefercats).

1.81 Thestatementsimplythateatingmorefiberwillcausepeopletoloseweight,sothisisacausal association.

1.82 Onepossibleconfoundingvariableistemperature(orseason).Morepeopleeaticecream,andgo swimming,inwarmweather.Otheranswersarepossible.Rememberthataconfoundingvariableshouldbe associatedwithbothofthevariablesofinterest.

1.83 Onepossibleconfoundingvariableispopulation.Increasingpopulationintheworldovertimemay meanmorebeefandmoreporkisconsumed.Otheranswersarepossible.Rememberthataconfounding variableshouldbeassociatedwithbothofthevariablesofinterest.

1.84 Onepossibleconfoundingvariableiswealth.Peoplewhoownayachtarelikelywealthyandcanafford asportscar.Otheranswersarepossible.Rememberthataconfoundingvariableshouldbeassociatedwith bothofthevariablesofinterest.

1.85 Onepossibleconfoundingvariableissnowinthewinter.Whenthereismoresnow,salesofboth toboggansandmittenswillbehigher.Rememberthataconfoundingvariableshouldbeassociatedwith bothofthevariablesofinterest.

1.86 Onepossibleconfoundingvariableisnumberofcars(andalsonumberofpeople).Iftherearelotsof cars,therewillbemorepavementandmoreairpollution.Rememberthataconfoundingvariableshouldbe associatedwithbothofthevariablesofinterest.

1.87 Onepossibleconfoundingvariableisgender.Malesusuallyhaveshorterhairandaretaller.Other answersarepossible.Rememberthataconfoundingvariableshouldbeassociatedwithbothofthevariables ofinterest.

1.88 Wearenotmanipulatinganyvariablesinthisstudy,weareonlycollectinginformation(ricepreference andmetabolism)astheyexist.Thisisanobservationalstudy.

1.89 Weareactivelymanipulatingtheexplanatoryvariable(playingmusicornot),sothisisanexperiment.

1.90 Weareactivelymanipulatingtheexplanatoryvariable(plantingtreesornot),sothisisanexperiment.

1.91 Wearenotmanipulatinganyvariablesinthisstudy,weareonlymeasuringthings(omega-3oilsand wateracidity)astheyexist.Thisisanobservationalstudy.

1.92 Datawerecollectedafterthefactfromsprinters,marathonrunners,andnon-athletes.Nogeneswere manipulated.Thesedatacamefromanobservationalstudy.

1.93 Thepenguinsinthisstudywererandomlyassignedtogeteitherametaloranelectronictagsothisis anexperiment.

1.94 Allthreestudiesareexperimentssincethescientistsactivelycontrolthetreatment(tearsorsalt solution).

1.95 Apossibleconfoundingvariableisamountofsnowandiceontheroads.Whenmoresnowandicehas fallen,moresaltwillbeneeded and morepeoplewillhaveaccidents.Noticethattheconfoundingvariable hasanassociationwith both thevariablesofinterest.

1.96 Ageorgradelevel!Certainly,studentsinsixthgradecanreadsubstantiallybetterthanstudentsin firstgradeandtheyarealsosubstantiallytaller.Gradelevelisinfluencingbothofthevariablesanditisa confoundingvariable.Ifwelookatindividualgradesoneatatime,theassociationcouldeasilydisappear.

1.97 (a)Thecasesarepeople.Thetwovariablesarewhetherornotthepersonisagolferandhowlong thepersonlives.

(b)Novariablesweremanipulated,sothisresultcomesfromanobservationalstudy.

(c)Aconfoundingvariableisavariablethatisassociatedwiththeothervariablesandmayhelpexplain theassociation.Wealthorincomeisoneobviousconfoundingvariable,sincegolfingisarelatively expensivesportandthereforeassociatedwithwealthandwealthisalsoassociatedwithlivinglonger. Otherconfoundingvariablesarealsopossible.

1.98 (a)Thecasesarethe1, 773participants.

(b)Theexplanatoryvariableisthegrouptowhichtheparticipantisassigned.Itiscategorical.

(c)Thisisanexperimentsincethegroupisrandomlyassignedbytheresearchers.

(d)Thereare1, 773casesandfourvariables(groupassignmentandallthreeattitudemeasures),sothe datasetwillhave1773rowsand4columns.

1.99 Yes,thisstudyprovidesevidencethatloudermusiccausespeopletodrinkmorebeer,becausethe explanatoryvariable(volumeofmusic)wasrandomlydeterminedbytheresearchersandanassociationwas found.

1.100 Study1isarandomized,controlledexperiment,whileStudy2isobservational.Therefore,Study1 providesbetterevidenceofcausation,whileStudy2ismorepronetoconfoundingvariables(e.g.,people whoeatmorenutsmayalsobemorehealthyinotherways).

1.101 (a)Yes.Becausethestudyinmicewasarandomizedexperiment,wecanconcludecausation.

(b)No.Sinceitappearsthatthestudyinhumanswasanobservationalstudy,itisnotappropriate toconcludecausation.Althoughtheheadlinemaystillbetrueforhumans,wecannotmakethis conclusionbasedonthestudydescribed.

1.102 (a)Sincethestudentsweren’tassignedtoonetypeofdietoranother,thisisanobservationalstudy.

(b)No,wecannotconcludeacause-and-effectrelationshipsincetheresultdoesnotcomefromarandomized experiment.

(c)Aconfoundingvariableisavariablethatisassociatedwiththeothervariablesandmayhelpexplain theassociation.Inthiscase,therearemanypossibleconfoundingvariables.Oneisexercise,since studentswhoexercisemayalsotendtoeatwellandmayalsohavelowerratesofdepression.

1.103 (a)Thetwovariablesareamountofironinthesoilandamountofpotassiuminspinachgrownin thatsoil.Bothvariablesarequantitative.

(b)Amountofironinthesoilistheexplanatoryvariablesinceitappearstoaffecttheamountofpotassium inthespinach.

(c)Sincethereisnoindicationthattheamountofironinthesoilwascontrolledbytheresearchers,the dataappeartocomefromanobservationalstudy.

(d)No.Sincethedatadonotcomefromanexperiment,wecannotassumecausation.

(e)Aconfoundingvariableisavariablethatmightaffectboththeamountofironinthesoilandtheamount ofpotassiuminthespinachleaves.Examplesmightbetheamountofrainfallorthetemperatureof thesoilorotherchemicalsinthesoil.

1.104 (a)Thecasesarethechildrenbeingstudiedandthereare19ofthem.Therearethreevariables: Hoursspentreading,Hoursofscreentime,andConnectivityinthebrain.

(b)Allthreevariablesarequantitative.Itappearsthathoursspentreadingandhoursofscreentimeare theexplanatoryvariablesandconnectivityinthebrainistheresponsevariable.

(c)Thisisanobservationalstudy,sincenovariableswerecontrolled.

(d)No,wecannotconcludecausationsincethiswasnotanexperiment.

(e)Aconfoundingvariableaffectsbothvariablesinanassociationandmighthelpexplaintheapparent association.Inthiscase,onepossibleconfoundingvariablemightbetheconnectivityintheparents’ brains,sinceparentswithhigherbrainconnectivitymightbemorelikelytoreadtotheirkidsandless likelytohavethemspendtimeonscreens,andchildrenofparentswithhighbrainconnectivitymight bemorelikelytoinherithighbrainconnectivity.Otheranswersarepossible.

1.105 (a)Thisisanobservationalstudysincenoneofthevariableswascontrolled.

(b)Thereareseveralpossibleanswers.Wealthisone,sincewealthhasbeenshowntobeassociatedwith betterhealthandwealthypeoplemightfinditeasiertotakevacations.Jobstressmightbeanother sincejobstressmightharmhealthandjobstressmightalsomakeitdifficulttotakevacations.

(c)Theresultsofthisstudyindicateanassociationbutnotacausationassociation.Thusthefirsttwo statementsareappropriateandthesecondtwoarenot.(i)Yes(ii)Yes(iii)No(iv)No.

(d)No.“WanttobeHealthier?TakeaVacation!”impliescause-and-effect,whichisnotanappropriate conclusionfromthisobservationalstudy.

1.106 No,thiswasanobservationalstudyandallowedformanyconfoundingvariables.

1.107 (a)Thecasesarethe2,623schoolchildren.

(b)Theexplanatoryvariableistheamountofgreeneryaroundtheschools.

(c)Theresponsevariableisthescoreonthememoryandattentiontests.

(d)Yes,theheadlineimpliesthatmoregreenspacecauseskidstobesmarter.

(e)Novariablesweremanipulatedsothisisanobservationalstudy.

(f)No!Sincethisisnotanexperiment,wecannotconcludecausation.

(g)Thesocioeconomicstatusofthechildrenisapossibleconfoundingvariable,sinceitislikelytoeffect boththeamountofgreenspaceandalsothetestscores.Thereareotherpossibleanswers.

1.108 (a)ThecasesaretheDanishmenwhowereincludedinthestudy.

(b)Theexplanatoryvariableiswhetherornotthepersonhasbeeninthehospitalforaninfection.Since theansweriseitheryesorno,thisisacategoricalvariable.

(c)TheresponsevariableistheIQscorefortheperson.Wecanfindanaveragesothisisaquantitative variable.

(d)Yes,theheadlineimpliesthatinfectionslowerIQ,whichiscausation.

(e)Theexplanatoryvariablewasnotmanipulated,sothisisanobservationalstudy.

(f)No,sincethiswasnotanexperiment,itisnotappropriatetoconcludecausation.Therearemany possiblelurkingvariablesthatmightimpactbothIQandthelikelihoodofhavinganinfection.

1.109 (a)Theexplanatoryvariableisamountofleisuretimespentsittingandtheresponsevariableis whetherornotthepersongetscancer.

(b)Thisisanobservationalstudybecausetheexplanatoryvariablewasnotrandomlyassigned.

(c)No,wecannotconcludespendingmoreleisuretimesittingcausescancerinwomenbecausethisisan observationalstudy.

(d)No,wealsocannotconcludethatspendingmoreleisuretimesittingdoesnotcausecancerinwomen; becausethiswasanobservationalstudywecanmakenoconclusionsaboutcausality.Sittingmayor maynotcausecancer.

1.110 (a)Theexplanatoryvariableiswhetherornottheparticipantsweregivenaccesstofoodanddrink after10pm,orjustwater.Theresponsevariablesarereactiontimeandnumberofattentionlapses.

(b)Thisisarandomizedexperimentbecausetheexplanatoryvariablewasrandomlyassigned.

(c)Yes,wecanconcludethateatinglateatnightworsensreactiontimeandincreasesattentionlapsesfor sleepdeprivedpeople.

(d)No,therearenotlikelytobeconfoundingvariables,becausethiswasarandomizedexperiment.

1.111 (a)Thisisanobservationalstudybecausewecannotrandomlydeterminewhenachildlearnsto talk.

(b)No,wecannotconcludethatearlylanguageskillsreducepreschooltantrumsbecausethisisanobservationalstudy.

(c)Therearemanypossibleconfoundingvariables,suchastheintelligenceofthechildorparentalinvolvement.

1.112 (a)No,becauseitisanobservationalstudy,notarandomizedexperiment.Peoplewerenotrandomlyassignedtobuyorganicfoodornot,butratherchoseforthemselves,sothegroupsalmost certainlydifferedtobeginwith.

(b)Answerswillvary,butpossibleanswersmayincludeincome(organicfoodcostsmoremoneyand wealthierpeoplearegenerallyhealthier)orotherhealth-consciousbehaviors(peoplewhocareabout eatingorganicmayeathealthieringeneral).

(c)No,becausewedonothaveevidenceagainstalternativeexplanation(ii);wecannotdeterminewhether explanation(i)or(ii)(orboth)aredrivingtheobserveddifference.

1.113 (a)Yes,becausethiswasarandomizedexperiment;itwasrandomlydeterminedwhichfliesate organicandwhichateconventionalfood.

(b)Answertopart(a)wasnotno.

(c)Yes,becausewehaveevidenceagainstbothofthecompetingalternativeexplanations,wehaveconvincingevidencethateatingorganicfoodreallydoesmakefruitflieslivelonger.

1.114 (a)Sinceallthechildrenarepresentedwithallthreeformats,thisisamatchedpairsexperiment (althoughitmightmoreaccuratelybecalledamatchedtriplesexperiment!)

(b)Yes,wecanconcludecausationsincetheresultscomefromanexperiment.

(c)Theformats(andthestoriesused)shouldbepresentedtothechildreninrandomorder.(Thisishow theexperimentwasconducted.)

1.115 (a)Theexplanatoryvariableiswhetherthepersonjusthadafullnightofsleepor24hoursofbeing awake.Theresponsevariableisabilitytorecognizefacialexpressions.

(b)Thisisarandomizedexperiment,amatchedpairsexperimentbecauseeachpersonreceivedboth treatments.

(c)Yes,wecanconcludethatmissinganightofsleephinderstheabilitytorecognizefacialexpressions, becausetheexplanatoryvariablewasrandomlyassigned.

(d)No,wecannotconcludethatbetterqualityofREMsleepimprovesabilitytorecognizefacialexpressions,becausetheexplanatoryvariableinthiscase(qualityofREMsleep)wasnotrandomlyassigned.

1.116 (a)No,wecannotconcludethatdrinkingdietsodacausesweightgainbecausethestudywas observationalandpronetoconfoundingvariables.Forexample,itispossiblethatseniorswhodrink moredietsodadosobecausetheyknowtheyarepronetogainingweight.

(b)Neitherstudyisperfect.Thestudyonseniorcitizensisobservationalandthestudyonratsmaynot generalizetohumans.However,togetherthesestudiesprovideamoreconvincingcasethatdietsoda cancauseweightgain.Opinionscanvaryonhowstrongthisevidenceforcausationis.

1.117 (a)Itisanobservationalstudysincenooneassignedsomepeopletoliveinacityandsometolive inthecountry.

(b)No,sincewecanneverconcludefromanobservationalstudythatthereisacausalassociation.

(c)The2011studyisalsoanobservationalstudy,since,again,nooneassignedsomepeopletoliveina cityandsometoliveinthecountry.

(d)Theexplanatoryvariableiswhetherornottheparticipantlivesinthecityorthecountry,whichisa categoricalvariable.Theresponsevariableislevelofactivityinstresscentersofthebrain,whichis quantitative.

(e)No!Theresultscomefromanobservationalstudy,sowecannotconcludeacausalrelationship.

1.118 Theexplanatoryvariablesarethetypeof payment and sex.Onlythetypeofpaymentcanberandomly assigned.Thenumberof items orderedand cost areresponsevariables.

1.119 (a)Theexplanatoryvariableiswhetherornotthepersonhadagoodnight’ssleeporissleepdeprived.Theresponsevariableisattractivenessrating.

(b)Sincetheexplanatoryvariablewasactivelymanipulated,thisisanexperiment.Thetwotreatments arewell-restedandsleep-deprived.Sinceall23subjectswerephotographedwithbothtreatments,this isamatchedpairsexperiment.

(c)Yes,wecanconcludethatsleep-deprivationcausespeopletolooklessattractive,becausethisisan experiment.

1.120 (a)Werandomlydividetheparticipantsintotwogroupsof25each.Halfwillbegivenfluoxetine andhalfwillgetaplacebo.

(b)Theplacebopillswilllookexactlylikethefluoxetinepillsandwillbetakenthesameway,butthey willnothaveanyactiveingredients.

(c)Thepatientswon’tknowwhoisgettingwhichtypeofpill(thefluoxetineortheplacebo)andthe peopletreatingthepatientsandadministeringthequestionnairewon’tknowwhoisinwhichgroup.

1.121 (a)Theexplanatoryvariableisamountofsleepandtheresponsevariableisgrowthinheight.

(b)Wewouldtakeasampleofchildrenandrandomlydividethemintotwogroups.Onegroupwouldget lotsofsleepandtheotherwouldbedeprivedofsleep.Thenaftersometimepassed,wewouldcompare theamountofheightincreaseforthechildreninthetwogroups.

(c)Anexperimentisnecessaryinordertoverifyacauseandeffectrelationship,butitwoulddefinitely notbeappropriatetorandomlyassignsomeofthekidstobesleep-deprivedforlongperiodsoftime justforthepurposesoftheexperiment!

1.122 (a)Randomlyassign25peopletocarbo-loadand25peopletonotcarbo-loadandthenmeasure eachperson’sathleticperformancethefollowingday.

(b)Wewouldhaveeachpersoncarbo-loadandnotcarbo-load,ondifferentdays(preferablydifferent weeks).Theorderwouldberandomlydetermined,sosomepeoplewouldcarbo-loadfirstandother peoplewouldcarbo-loadsecond.Inbothcasesathleticperformancewouldbemeasuredthefollowing dayandwewouldlookatthedifferenceinperformanceforeachpersonbetweenthetwotreatments.

(c)Thematchedpairsexperimentisprobablybetterbecauseweareabletocomparethedifferenteffects forthesameperson.Itismoreprecisecomparingoneperson’sathleticperformanceundertwodifferent treatments,ratherthandifferentpeople’sathleticperformanceundertwodifferenttreatments.

1.123 (a)Randomlydividethestudentsintotwogroupsof20studentseach.Onegroupgetsalcoholand theothergetswater.Measurereactiontimeforstudentsinbothgroups.

(b)Measurereactiontimeforall40studentsbothways:afterdrinkingalcoholandafterdrinkingwater. Dothetestsonseparatedaysandrandomizetheorderinwhichthestudentsaregiventhedifferent treatments.Measurethedifferenceinreactiontimeforeachstudent.

1.124 Answerswillvary.Example:Thetotalamountofpizzaconsumedandthetotalamountofcheese consumed,peryear,overthelastcentury.Eatingmorepizzacausespeopletoeatmorecheese,butthe overallriseinpopulationisalsoaconfoundingvariable.

Chapter1 CollectingData

ChapterOverview

Webeginthischapterbyintroducingthebasicstructureofdatasets,withcasesandvariables.Inthefirst section,wealsoincludemotivatingexamplestogetstudentsthinkingaboutthekindsofquestionsthatcan beansweredusingdata.Thegoalisalsotobuildenthusiasmforhowpowerfulandfunstatisticscanbe.

Therestofthechapterisaboutcollectingdata.Thesecondsectionisaboutsamplingandthethirdabout experimentsandobservationalstudies.Inthesesections,wehavechosen not toincludealonglistofmanyof thestandarddefinitions.(Wedon’t,forexample,formallydefinenon-responsebiasorstratifiedsampling.) Thefocus,instead,isonhelpingstudentsthinkcriticallyabouthowdataarecollected,andonhelping studentsunderstandhowthecollectionmethodinfluencesthepossiblescopeofinference.Werevisitthese topicsregularlythroughouttherestofthebooktocontinuetoreinforcetheseideas.

1.1TheStructureofData

KeyConcepts

• Datasetshavecasesandvariables

• Variablescanbeclassifiedaseitherquantitativeorcategorical

• Wecanusedatatoinvestigateinterestingquestions

• Ininvestigatingrelationshipsbetweenvariables,wecansometimesidentifyanexplanatory variableandaresponsevariable

Timing

Oneclassorless.Ifyourstudentshaveweakerbackgrounds,youmightdelaytalkingaboutexplanatoryand responsevariablesuntilalatersection.Thissectionisdesignedtobecoveredinthesameclassinwhichyou gooveryoursyllabus,andthematerialcanshrinkorexpandtofilltheallottedtime.

ClassNotes

Datasetshavecasesandvariables.Agreatwaytosetthetoneforthecourseistobeginthefirstclass bycollectingdatadirectlyfromthestudents.Thisdatasetcanthenprovideexamplestouseinclassall semesterlong.Itcanalsobeaperfectjumpingoffpointtotalkaboutcasesandvariables.Seemoredetails intheclassactivitydiscussionbelow.Youwillalsowanttohavesomeotherexamplesreadytogo,where youcanaskthestudentstoidentifythecasesandthevariables.Suggestionsaregivenbelow.

Variablescanbeclassifiedaseitherquantitativeorcategorical.Foreachoftheexamplesusedto identifycasesandvariables,havethestudentsidentifyeachofthevariablesascategoricalorquantitative. Again,suggestedexamplesaregivenbelow.Makesurestudentsknowthatalloftheanalysisthatfollowsin theentirebookdependsonfirstidentifyingvariablesascategoricalorquantitative.Itwillpayofflaterto spendtimeonthisnow!

Wecanusedatatoinvestigateinterestingquestions.Trytofindafewexamplesthatyourstudents willbeinterestedin,tobegintogenerateinterestinthesubject.Someideasaregivenbelow,oruseexamples fromthesection.Anotherideaistopullsomerecentheadlines1 andhavethestudentstalkabouthowto usedatatoinvestigatethequestionofinterest.Getthemtoidentifycasesandvariables,andtobeginto recognizehowpowerfuldataanalysiscanbe!

Ininvestigatingrelationshipsbetweenvariables,wecansometimesidentifyanexplanatory variableandaresponsevariable.Thisconceptislessimportantthantheothers(atthispoint)andcan bedelayeduntilalatersectionifdesired.Theconceptishelpful,though,ingettingthestudentsthinking aboutrelationshipsandcanfitinwellwiththeearlierexamples.Italsoleadsnicelyintothediscussionof experimentsthatwillcomeinSection1.3.

1 AgoodsourceforlinkstorecentnewsarticlesofinterestforteachingstatisticsistheChanceNewsprojectat http://test.causeweb.org/wiki/chance/index.php/Main Page

ClassActivity

Havestudentsanswerquestionsonastudentsurvey.Wegivetwooptionsbelow,butfeelfreetomodifythem toworkforyourstudents.Besuretohaveatleasttwoquantitativeandatleasttwocategoricalvariables. Themorevariables,thebetter!

Tellyourstudentsthatstatisticsisallaboutdataandweneedtocollectdata!Givethemtimeduringclass tofilloutthesurvey(orhavethemdoitonline2 outsideofclass).Aftercollectingthedata,havethem gothroughthevariablesandidentifyeachascategoricalorquantitative.Also,havethemcomeupwith questionstheymightliketoanswer,usingthesevariables.(Ifyouaskthe”Pulse”question,besuretohave awaytotime60secondswhiletheycounttheirpulse.)IfyouhavemultiplesectionsofyourIntroStats course,itisgreattocombinethedatafromallthesectionsintoonelargefile,sincethatincreasesanonymity.

Onceyouhavethedata,youcanuseitforclassexamplesthroughoutthesemester.Usingdataabout YOURstudentsmakesitmorerealtothem.Forprivacyreasons,though,don’tsharethefulldatasetwith thestudents-justsharesummarystatisticsandgraphsasneeded.

• Option1: Haveastudentsurveyreadytogotocollectdatainthefirstclass.Oneoptionisgivenas theSection1.1ActivityintheInstructorResources.Youcanusethequestionsthereormodifythem inanywaythatworksforyou.Thekeyistohaveenoughvariablesthatyoucanusethedatafor interestinganalysislaterinthecourse.

• Option2: Havethestudentsthemselvescreatethestudentsurvey!Duringthefirstclass,askthem whattheywouldliketoasktheirpeersandhavethemdecidethequestionstoputonthesurvey.Use thesesuggestionstocreateastudentsurveytobringtothesecondclass(orsetituponline)andhave themfillout.ThisideaisalsogiveninmoredetailintheActivityhandout.

ClassExamples

Inadditiontotheclassexamplesthatarisefromdiscussingthestudentsurvey,hereareacoupleother suggestions.SeealsothePowerPointslidesandtheClassExampleshandout.Also,checkthetextexercises foradditionalideas.Therearemoreexercisesthanneededandmanyarequiteinterestingandwouldmake greatclassexamples.Thedatasetsthatcomewiththebookalsoprovideagoodsourceforgettingyour studentscomfortablethinkingaboutcasesandvariables.

Example HollywoodMovies2011

Hereisasmallpartofthedatasetthatincludesinformationonall136moviestocomeout ofHollywoodin2011.(Thefulldatasetisavailablein HollywoodMovies2011.) Rating isaudienceratingona100-pointscale, Budget isbudgetinmillionsofdollars,and Opening isopeningweekendgross,inmillionsofdollars.

Title LeadStudioRatingGenreBudgetOpening Insidious Sony65Horror1.513.27 HarryPotterandDeathlyHallowsP2 WarnerBros92Fantasy125.0169.19 Bridesmaids Relativity77Comedy32.526.25 TheHelp DreamWorks91Drama25.026.04 HorribleBosses WarnerBros72Comedy35.028.30 Transformers:DarkoftheMoon DreamWorks67Action195.0197.85

2 Forexample,youcansetupasurveyrelativelyeasilyatwww.surveymonkey.com.

(a) Whatarethecasesinthisdataset?

(b) Whatarethevariables?Identifyeachasquantitativeorcategorical:

(c) Nameaquestionwemightaskaboutthisdatasetthatisaboutasinglevariable.Name aquestionwemightaskthatisaboutarelationshipbetweentwoofthevariables.

Solution (a) Themoviesarethecases.

(b) Therearefivevariables.(Notethat“Title”isjustanidentifiercolumnandisnota variable.)“LeadStudio”and“Genre”arebothcategoricalwhile“Rating”,“Budget”, and“Opening”areallquantitative.

(c) Therearemanypossibleanswers.Foraquestionaboutasinglevariable:Whatproportionarecomediesorwhatisanaverageaudienceratingorwhichmoviehadthe largestbudget?Letthestudentssuggestideas!Forarelationshipbetweentwoofthe variables:Whichgenreofmoviehasthehighestbudget?Isaudienceratingrelated toopeningweekendgross?Isoneofthestudiosmorelikelytomakehorrorfilms? Domovieswithahigherbudgettendtogethigheraudienceratings?Youalsomight usesomeofthesequestions(orstudentsuggestions)toaskaboutwhichwouldbethe explanatoryvariableandwhichtheresponsevariable.

Example YogurtandWeightLoss Caneatingayogurtadaycausepeopletoloseweight(ascomparedtonoteatingyogurt)? Inastudyinvestigatingthis:

(a) Whatarethecases?

(b) Whatarethevariablesandiseachquantitativeorcategorical?

(c) Whatmightadatasetlooklikethatmighthelpusanswerthisquestion?Makeup somedatatoshowthefirstcoupleofrowsofsuchadataset.

(d) Whichvariableistheexplanatoryvariable?Whichistheresponsevariable?

Solution (a) Thecasesarethepeopleinvolvedinthestudy.

(b) Onevariableiswhetherornotthepersoneatsayogurtaday,whichiscategorical. Anothervariableistheamountofweightloss,whichisquantitative.(Otherreasonable studentresponsesmightbe:Didthepersonloseweight(yes/no),whichiscategorical orrecordweightsastwoquantitativevariablesbeforeandafterthestudy.)

(c) Studentshaveasurprisinglyhardtimewiththis,andaskingquestionssuchasthis reallyseemstohelptheminternalizetheideaofcasesandvariables.Wehave:

(d) Sinceweareinterestedintheeffectofyogurtonweightloss,theexplanatoryvariable iswhetherornotthepersoneatsayogurtadayandtheresponsevariableistheweight loss.

1.1.THESTRUCTUREOFDATA

ExerciseNotes

Exercises1-6,11,12,13,14,15,16,17,18,19,22,23,26,27,28,29,30,31,32,33askstudentstoidentify casesandcategorical/quantitativevariables.

Exercises11,12,13,14,15,16,17,18,19,20,21,22,23,26,27,28,29,30,31,32,33askstudentstothink abouthowdataisusedtoansweraquestion.

Exercises7-10,13,14,15,16,22,23askstudentstoidentifyexplanatoryandresponsevariables.

Exercises30,31,32,and33askstudentstothinkmoredeeplyabouthowdataisrecorded.Exercises30, 31,32,and33areopen-ended.

Therearenoexercisesrequiringtechnology,althoughstudentsmightchoosetolookattheonlinedatasets forExercises11and12.

SuggestedExercises:

SkillBuilders:1-10odd(oreven);Exercises:12,16,22,27,30

1.2SamplingfromaPopulation

KeyConcepts

• Samplescomefrompopulations

• Samplescanbebiased

• Ifasampleisrepresentativeofapopulation,wecanuseinformationfromasampleto inferinformationaboutapopulation

• Randomsamplingisthemosteffectivewaytogetarepresentativesample

Timing

Oneclass.

SectionOverview

Themaingoalofthissectionistogetstudentstorecognizeandthinkaboutpossiblesourcesofbiasin sampling,andtorecognizehowharditistoavoidbias.Wedo not givealaundrylistofterms(suchas systematicsampling,stratifiedsampling,non-responsebias,etc.)butratherjustencouragestudentsto thinkcriticallyabouthowdataarecollected.Thisapproachmaybedifferentthanthatoftentakeninother books.Attheendofthesection,studentsshouldknow:

• Randomsamplingisverygood!

• Randomdoesnotmean“haphazard”,andpeoplecan’tdorandomwithoutusinganobjectiverandomizationmethod.

• Volunteersamplesaregenerallynotrepresentativesamples.

• Therearemanyformsofbias.

• Thestatisticalinferencewetalkaboutintherestofthebookdependsonhavinggoodsampling methods.

ClassNotes

Samplescomefrompopulations.Thisisaneasyconceptforthestudentsandanimportantonefor themtointernalize.TheactivitybelowinvolvingLincoln’sGettysburgAddressprovidesagreatwaytoget studentsthinkingaboutsamplingfromapopulationandisanexcellentwaytostarttheclass.Youwillalso wantacouplemorequickexamplesofsamplesfrompopulations.Suggestionsaregivenbelow.

Samplescanbebiased.IfyouusetheLincolnactivity,thatwillprobablyprovideanimmediateexample ofbiasedsamplingsincestudentsarelikelytoover-samplethelargewords.Givesomemoreveryobvious examplesofbiasedsamples(suchassamplingstudentsinthelibraryonaFridaynight)andthensomemore subtlewaysinwhichasamplemightgivebiasedresults.Theexampleinthebookaskingwhetherparents wouldchoosetohavechildrenagainisagreatone.Includeatleastoneexampleofavolunteersampleand atleastoneexampleofwordingbias.Seeexamplesbelow.

Ifasampleisrepresentativeofapopulation,wecanuseinformationfromasampletoinfer informationaboutapopulation.Wetalkaboutinferencehereinaveryinformalway,justtohelppreparestudentsforwhatistocome.Theonlygoalistogetthemtothinkgenerallyaboutusinginformation fromasampletoestimateinformationforapopulation.Examplesarebeloworuseanyrecentsurveyto talkaboutgeneralizingtheresultstothebroaderpopulation.Ifthesurveyisfromarandomsample(asare allthesurveysputoutbythePewResearchCenteronpewresearch.org),allthebetter.

Randomsamplingisthemosteffectivewaytogetarepresentativesample.Itisimportantforthe studentstorecognizethattheycan’tdorandomontheirownandthatrandomdoesnotmean“haphazard.” Arandomsampleisfoundonlybypullingnamesoutofahatorusingarandomnumbergeneratoror tableorsomeequivalentmethod.Youmayormaynotchoosetoshowyourstudentshowtousearandom numbergeneratorortable–thisisanoptionaltopic.Eitherway,though,thestudentsshouldrecognizethe importanceofselectingarandomsample. StatKey canbeusedtodemonstraterandomsamples,usingthe SamplingDistributionsoptions.Iftime,returntotheLincolnactivityfromthestartofclassandrepeatthe activityusingtruerandomsamples.Comparetheresultstothoseobtainedwiththeirbiasedsamples.

ClassActivity

GivestudentsacopyofLincoln’sGettysburgAddressandaskthemeachtoselectasampleof10wordsto estimatetheaveragenumberoflettersinallwordsinthespeech.(ThisisavailableontheActivityhandout forthissection.)Havethemthenrecordthenumberoflettersineachoftheir10wordsandcomputethe averagelengthfortheirsample.

Thisisagoodopportunitytopointoutthepopulation(all268wordsinLincoln’sGettysburgAddress)and thedifferentsamplesthatstudentsselectedfromthatpopulation.Havethestudentsallputtheiraverages onadotplotontheboard,oryoucanentertheaveragesintoadatasetasthestudentsgoaroundtheroom andannouncetheirs.Comparetheresultingdistributiontothecorrectaveragenumberoflettersforall268 words,whichis4.3.

Itislikelythatalmostallthestudentaveragesaretoohigh.Talkaboutthefactthattheyaremorelikelyto samplelargewordswhilemostlyignoringtheverysmallwords.Discussthefactthatitisvirtuallyimpossible forhumanstotakea“random”sample,evenwhenweknowabouttheproblemandareactivelytryingto compensate.Weneedtouseanobjectivemethod,suchasarandomnumbergenerator.

Ifyouhavetime,itisgreattorepeatthisexerciseattheendofclass,usingarandomnumbergeneratoror tabletoselectrandom samplesofsize10fromthelistofall268words.Thedistributionstudentsgetusing thismethodshouldbecenterednearthepopulationmeanof4.3.Thisreallydrivesthepointhomeofhow harditisforhumanstodo“random”.

ClassExamples

Inadditiontotheclassexamplesthatarisefromtheactivity,hereareacoupleothersuggestions.Seealso thePowerPointslidesandtheClassExampleshandout.Also,checkthetextexercisesforadditionalideas. Therearemoreexercisesthanneededandmanyarequiteinterestingandwouldmakegreatclassexamples. Websiteswithrecentpublicopinionpolls,suchaspewresearch.orgorgallup.comorrasmussenreports.com, alsocanprovidegreatcurrentexamplestoillustratesampling.

Example NumberofCellPhoneCallsperDay PrincetonSurveyResearchreportsonasurveyof1,917cellphoneusersintheUS,conducted

inMay2010,asking“Onanaverageday,abouthowmanyphonecallsdoyoumakeand receiveonyourcellphone?”

(a) Whatisthesampleinthisstudy?

(b) Whatisthepopulation?

(c) Isthevariableinthisstudyquantitativeorcategorical?

Solution (a) The1,917peoplewhowereaskedthequestion.

(b) AllcellphoneusersintheUS

(c) Quantitative

Example DrivingwithaPetonyourLap

Over30,000peopleparticipatedinanonlinepolloncnn.comconductedinApril2012asking ”Haveyoueverdrivenwithapetonyourlap?”Weseethat34%oftheparticipantsanswered yesand66%answeredno.

(a) Whatisthesample?

(b) Canweconcludethat34%ofalldrivershavedrivenwithapetontheirlap?

(c) Explainwhyitisnotappropriatetogeneralizetheseresultstoalldrivers,oreventoall driverswhovisitcnn.com.Whatistheproblemwiththismethodofdatacollection?

(d) Howmightweselectasampleofpeoplethatwouldgiveusresultsthatwecangeneralize toabroaderpopulation?

(e) Isthevariableinthisstudyquantitativeorcategorical?

Solution (a) The30,000peoplewhoparticipatedinthepoll.

(b) No!Thisisavolunteersample,andvolunteersamplesareusuallybiased.Itislikely thatpeoplewhohavedrivenwithapetontheirlaparemorelikelytorespondtothe poll.

(c) Volunteersamplesaregenerallybiasedsincepeoplewhofeelstronglyaboutthesubject aremuchmorelikelytoparticipate.

(d) UseaRANDOMSAMPLE!

(e) Categorical

Example SamplingBias

Youwishtoestimatethenumberofhoursaweekthatstudentsatyouruniversityspend studying.Foreachsamplingmethodbelow,indicatewhethertheresultsarelikelytobe representativeoftheentirestudentbodyorwhetherthemethodislikelytogivebiased results.

(a) AskstudentsyoufindatthelibraryonaFridaynight.

(b) AskstudentswhoareatapartyonaTuesdaynight.

(c) Askarandomsampleofstudents.

(d) AskthestudentsinOrganicChemistry.

(e) Askallthestudentsononeoftheschool’sathleticteams.

Solution

(f) Sendanemailouttoallstudentsandusealltheresponsesyouget.

(g) Askfriendsthatyouknow,tryingtogetareasonablyrepresentativegroup.

(a) Biased

(b) Biased

(c) Representative

(d) Biased

(e) Biased

(f) Biased,sincethisisavolunteersample

(g) Biased,sincebiascanoccurwhenpeopletrypickwhattheythinkisrepresentative

ExerciseNotes

Exercises34-41,42-45,57,66,72,73askstudentstoidentifysamplesandpopulations.

Exercises42-45,52-56,58,65,66,67,69askstudentstoidentifywhenasampleislikelytobebiased.

Exercises46-51,59,60,58,74,75discussrandomsampling.

Exercises74and75requirestudentstousearandomnumbergeneratororarandomnumbertable.Exercises 70and71askstudentstothinkmorecreativelyaboutwhenwecangeneralizefromasampletoapopulation.

SuggestedExercises:

SkillBuilders:34-56odd(oreven);Exercises:58,59,67,68

1.3ExperimentsandObservationalStudies

KeyConcepts

• Associationdoesnotimplycausation!

• Confoundingvariablescanexplainanapparentassociationbetweentwoothervariables

• Distinguishingbetweenexperimentsandobservationalstudies

• Randomizedexperimentsprovidetheonlymethodthatallowsustoinfercausality

• Aspectsofawell-designedexperiment

Timing

Oneortwoclasses.

SectionOverview

Thismaterialwillpermeatetherestofthecourseandstudentsespeciallyneedtorecognizewhentheycan andcannotconcludethatthereisacausalrelationship.However,themorespecificmaterialattheendof thesection(placebos,blinding,matchedpairs)canbecoveredeitherveryquicklyorindetail,dependingon theinterestsoftheinstructor.

ClassNotes

Associationdoesnotimplycausation! Starttheclassbydiscussingthedifferencebetweenanassociationanda causal association.Studentsshouldbeabletorecognizewhenastatementimpliesacausal relationship.Thedistinctioncanbesubtlesometimes,soincludeseveralexamples.Usetheexamplesbelow orpullsomerecentheadlinesandaskwhichappeartoimplycausation.Youmightwanttosharewith thestudentsthisquotefromStephenJayGould:“Theinvalidassumptionthatcorrelationimpliescauseis probablyamongthetwoorthreemostseriousandcommonerrorsofhumanreasoning.”

Confoundingvariablescanexplainanapparentassociationbetweentwoothervariables.Defineaconfoundingvariableasathirdvariablethatisassociatedwithboththeexplanatoryvariableand theresponsevariableinarelationship.(Inthistext,wedon’tdifferentiatebetweenconfoundingvariables, lurkingvariables,confoundingfactors,etc.)Youwantthestudentstorecognizethatconfoundingvariables areoftenpresentinobservationalstudiesandthataconfoundingvariablecanbetherealreasonbehindan apparentassociation.Giveseveralassociations(suchasthosesuggestedbelowintheclassexamples)and askthestudentstosuggestconfoundingvariables.Theyhaveasurprisinglyhardtimewiththis.

Distinguishingbetweenexperimentsandobservationalstudies.Aslongasstudentscanremember that,inanexperiment,theexplanatoryvariableisactivelymanipulatedand,inanobservationalstudy,it isn’t,theyprobablywon’thaveahardtimewiththis.Again,acoupleshortexampleswilldothetrick.

Randomizedexperimentsprovidetheonlymethodthatallowsustoinfercausality.Thisisthe mostimportantconceptofthesection.Youwillwanttomakethepointstronglyinthisclass,andthen

reinforceitthroughouttherestofthesemester.Itisworthasking,wheneverwefindanassociationinthe remainderofthecourse,whetherwecanconcludethatitisacausalassociation.Emphasizetheimportance ofassigningtheexplanatoryvariablerandomly.Theclassactivityrecommendedbelowwillhelptomake thispoint.

Aspectsofawell-designedexperiment.Thispartofthesectioncantakeanotherfullclassorcan becoveredquickly,dependingonthecourse.Itisworthtalkingatleastquicklyabouttheuseofplacebos andblinding,sincethestudentswillseethosewordsinmanyoftheexamplesandexercisesthroughoutthe restofthebook.Also,studentsfindtheplaceboeffectveryinteresting!Therearesomefunexamplesand informationinthePowerPointslides.Itisfinetodelaydiscussionofamatchedpairsdesignuntillaterin thecourseifdesired.

ClassActivity

Thisactivity,onconductingarandomizedexperiment,shouldcomerightafteryouhavetalkedaboutrandomizedexperimentsintheclass(andbeforetalkingaboutplacebos,matchedpairs,etc.)Weconductan experimenttoinvestigatethequestion: Isjust10secondsofexerciseenoughtoincreasepulserates?

Thetreatmentgroupsareexercisevssittingandtheresponsevariableispulserate.

Randomlydividetheclassintotwogroups,usingsomelegitimaterandomizationmethod(suchasarandom numbergeneratororrandomnumbertable,orputtingallstudentnamesinahatanddrawingouthalf,or flippingacoinforeachstudenttodeterminethegroup.)Itisvaluableforthestudentstoseetheactual randomizationhappen.

Havethestudentsinonegroupexercisewhileyoucountout10seconds.(Leteachpersonchoosehowto exercise:runinplace,jumpingjacks,orwhatever.Thisispartofthefun!)Havetheothergroupremain sitting.Assoonasthe10secondsisover,haveallstudentsfindtheirpulseandthengivethemthestartand stopsignalasyoutimeoneminuteforthemastheycounttheirpulse.Collectthedatafromthem(using theclassactivityhandoutifdesired)forlateruse,orenteritrightawayintoadatasetthatwillallowyou tocomparemeansortolookatside-by-sidedotplots.

Youcanalsousethisactivitytoconnecttootherclassconcepts.Askthestudentstoidentifythecasesand thevariables,andtoindicatewhethervariablesarecategoricalorquantitative.Havestudentsidentifythe explanatoryvariableandtheresponsevariable.Discusspossibleconfoundingvariablesifwehad not assigned groupsrandomlybutinsteadhadletstudentspickwhichtreatmentgrouptheywishedtojoin.Finally,ask studentswhyamatchedpairsdesignmightbeanevenmoreeffectivewaytocollectdatatoaddressthe questionofinterest.

ClassExamples

Hereareacouplesuggestionsforclassexamples.SeealsothePowerPointslidesandtheClassExamples handout.Also,checkthetestexercises,includingReviewExercises,foradditionalideas.Therearemore exercisesthanneededandmanyarequiteinterestingandwouldmakegreatclassexamples.

Example AssociationorCausation?

Foreachofthefollowingstatements,indicatewhetherthestatementimpliescausationor justindicatesassociationwithoutcausation.

(a) Ifyoustudymore,yourgradewillimproveinthiscourse.

(b) Wanttoloseweight?Drinktea!

(c) Agingofthebraintendstobedelayedinpeoplewithacollegeeducation.

Solution (a) Causation.”Willimprove”impliescausation.

(b) Causation.Theimplicationisthatdrinkingteawillcauseyoutoloseweight.

(c) Associationwithoutcausation.Thestatementisnotclaimingthatacollegeeducation willcauseadelayinagingofthebrain.Infact,therearemanyconfoundingfactors.

Example TelevisionSetsandLifeExpectancy

Thereisaverystrongpositiveassociationbetweentelevisionsetsperpersoninacountry andthelifeexpectancyofthepeopleinthatcountry,whichmeanspeopleincountrieswith lotsofTVstendtolivelonger.

(a) Doesthismeanwecanimprovethelifeexpectancyofpeopleindevelopingcountries bysendingthemlotsoftelevisionsets?

(b) Describeapossibleconfoundingvariableinthisassociation.

Solution (a) No!Thisstrongassociationisnotcausation.

(b) Thewealthofthecountriesisaconfoundingvariable.Wealthwillaffectboththe numberoftelevisionsetsandthelifeexpectancy.

Example DoesKindergartenLeadtoCrime?

AstatelegislatorinNewHampshireclaimedinJuly2012thatkindergartenprogramslead tomorecrime.Henotedasevidencethatthelargestofallthecommunitiesinhiscounty wastheonlycommunitywithakindergartenprogramandalsohadthemostcrime.

(a) Whatcommonmistakeisthelegislatormaking?

(b) Describeapossibleconfoundingvariableinthisassociation.

Solution (a) Thelegislatorisassumingassociationimpliescausation.Thisaverycommonmistake! JustbecausethereisastrongassociationdoesNOTmeanthatoneiscausingthe other.

(b) Thepopulationofthecommunities.Largercommunitiesaremorelikelytohavea kindergartenprogramandarealsomorelikelytohavemorecrime.

Example ExerciseandDepression

Ineachcasebelow,wedescribeastudytodeterminewhetherexercisehelpsincreasecertain moodenhancingchemicalsinthebrain.Indicatewhethereachdescribesanexperimentor anobservationalstudy.

(a) Wecontactarandomsampleof100peopleandrecordhowmucheachpersonexercises andalsomeasurethechemicalsinthebrainforeachperson.

(b) Usingarandomsampleof100people,werandomlyassignhalfofthemtoparticipatein aregularexerciseprogramforasix-weekperiodwhiletheotherhalfmakesnochanges. Attheendofthetimeperiod,wemeasurethebrainchemicals.

Solution (a) Observationalstudy.Novariablesweremanipulated. (b) Experiment.Theexperimentersactivelyimposedtheexplanatoryvariable(exerciseor not).

Example CaffeineandLearning

Doesingestingcaffeinehelpmicelearnthewaythroughamazefaster?Describeawelldesignedrandomizedexperimenttotestthis.Givealldetails,includingtheuseofplacebos andblinding.Wehave20micetoworkwith.

Solution

Werandomlydividethemiceintotwogroupsof10miceeach.Wegivehalfthemicecaffeine intheirfoodwhiletheothergroupgetsidenticalfoodwithoutcaffeine(placebo).Wemake thestudydouble-blindbynottellingthemiceoranyoneinteractingwiththemice(for example,whoeveristimingthemazeruns)whichmicearegettingthecaffeinatedfoodand whicharenot.Wemeasurethetimeittakesforthemicetolearnthemazeandcompare theresultsbetweenthetwogroups.

Example IstheDominantHandStronger?

DescribeaMatchedPairsexperimenttoseeifrighthandstrengthisgreaterthanleft-hand strengthinright-handedpeople.Youhaveahand-griptomeasurestrengthand30righthandedpeopletouseinyourstudy.Besuretoincludehowrandomizationisusedinyour design.

Solution Wearecomparingright-handstrengthtoleft-handstrengthandinamatchedpairsexperiment,allparticipantsaremeasuredforboth,sowewillmeasurethestrengthinboththe leftandrighthandforallparticipants.Werandomizetheorderinwhichtheparticipants havehandstrengthmeasured:somerandomlyassignedtomeasurethestrengthintheleft handfirstandsomerandomlyassignedtomeasurestrengthintherighthandfirst.Then wecomparetherightandlefthandstrengthforeachindividual.

ExerciseNotes

Exercises76-81,99,100,101,106,107,108,109,110,111,115,116,117,119,124requireanunderstanding ofassociationversuscausation.

Exercises82-87,95,96,107,110,111,91requireanunderstandingofconfoundingvariables.

Exercises88-91,92-94,99,100,101,106,107,108,109,110,111,115,116,117,119asktodistinguish betweenexperimentsandobservationalstudies.

Exercises115,118,120,121,122,123askstudentstothinkaboutaspectsofwell-designedexperiments.

Exercises115,119,122,123aretheonlyexercisesaboutmatchedpairsdesign.Exercises116,120,121, 122,123areopen-endedproblemsaskingstudentstodescribeanwell-designedexperiment.Exercise124is harderandrequiresstudentstohaveadeeperunderstandingoftheideas. Noneoftheexercisesinthissectionrequiretechnology.

SuggestedExercises:

SkillBuilders:76-91odd(oreven);Exercises:96,99,101,110,111,117,120,122

Turn static files into dynamic content formats.

Create a flipbook
Statistics Unlocking The Power Of Data 4Th Lock Solutions Manual by dferdinan - Issuu