Revision of SPSS (2016) - Discovering · PDF fileRevision of IBM SPSS Statistics IBM SPSS Statistics™ (SPSS from now on) ... an experimental condition and a different group of participants
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Revision of IBM SPSS Statistics IBMSPSSStatistics™(SPSSfromnowon)isacomputerpackagedesigntocaterforyoureverystatisticalneed.Itcarriesoutanalysesthattakehoursbyhandinamatterofseconds.Here’salistofwhatSPSScanandcannotdo:
SPSSwill:• Docomplexstatisticalanalysisinafewseconds.• Producegraphs.• Saveyoutime.• Doanythingstatisticalthatyouaskittodo.SPSSwon’t:• Decide what test needs to be done in a givencircumstance.• Passyourexamforyou.• Maketea.So,althoughSPSSisaverypowerfultool,itworksonlywithinthelimits of your own statistical knowledge. In many respects it isratherstupidbecauseifyouprovideitwithdataandaskittorunacompletelymeaninglessanalysis,itwillhappilyoblige.Therefore,youstillneedtouseyourbraintoinitiatethecorrectanalysis.
Getting started with SPSS ThereareseveralexcellenttextsthatgiveintroductionstothegeneralenvironmentwithinwhichSPSSoperates.Inmyentirelyunbiasedopinion,thebestisField(2013)!SPSSmainlyusestwowindows:thedataeditor(thisiswhereyouinputyourdataandcarryoutstatisticalfunctions)andtheviewer (this iswhere the results of any analysisappear).
OnceSPSShasbeenactivated,astart-upwindowwillappear,whichallowsyoutoselectvariousoptions.Ifyoualreadyhaveadata fileondisk thatyouwouldliketoopenthenselectOpenanexistingdatasourcebyclickingonthe sothatitlookslike :thisisthedefault option. In the space underneath this optiontherewillbealistofrecentlyuseddatafilesthatyoucanselectwiththemouse.Ifyouwanttoopenadatafilethatisn’tinthelistthensimplyselectMoreFiles…withthemouseandclickon .
WhenyoufirstloadSPSSitwillprovideablankdataeditor(showingthe‘dataview’)withthetitleNewData.Wheninputtinganewsetofdata,youmustinputyourdatainalogicalway.TheSPSSdataeditorisarrangedsuchthateachrowrepresentsdatafromoneparticipantwhileeachcolumnrepresentsavariable.Thereisnodiscriminationbetweenindependent and dependent variables, both types should be placed in a separate column. It follows from thisarrangementthatanyvariablemeasuredwiththesameparticipants(arepeatedmeasure)shouldberepresentedbyseveralcolumns(eachcolumnrepresentingoneleveloftherepeatedmeasuresvariable).However,whenabetween-groupdesignwasused(e.g.differentparticipantswereassignedtoeachleveloftheindependentvariable)thedatawillberepresentedbytwocolumns:onethathasthevaluesofthedependentvariableandonethatisacodingvariableindicatingtowhichgrouptheparticipantbelonged.Thisideawillbecomeclearerasyoulearnabouthowtocarryoutspecificprocedures.
Thefirststepinenteringyourdataistocreatesomevariablesusingthe‘variableview’ofthedataeditor,andthentoinput your data using the ‘data view’ of the data editor.We’ll go through these two steps byworking through anexample.
The highlighted cell is the cell that is currently active
This shows that we are currently in the ‘Data View’
We can click here to switch to the ‘Variable View’
This area displays the value of the currently active cell
Beforeweinputanydataintothedataeditor,weneedtocreatethevariables.Tocreatevariablesweusethe‘VariableView’ of the data editor. To access this view click on the ‘Variable View’ tab at the bottom of the data editor (
Youcanenteranameinthiscolumnforeachvariable.Thisnamewillappearatthetopofthecorrespondingcolumninthedataview,andhelpsyoutoidentifyvariablesinthedataview.Therearecertainsymbolsyoucan’tuse(mainlysymbolsthathaveotherusesinSPSSsuchas+,−,$,&),andyoucan’tusespaces.(Manypeopleusea‘hard’spaceinvariablenames,whichreplacesthe space with an underscore; for example, Andy_Field instead of Andy Field.) If you use acharacter that SPSS doesn’t like you’ll get an errormessage saying that the variable name isinvalidwhenyouclickonadifferentcell,ortrytomoveoffthecellusingthearrowkeys.
Tobeginwith,imaginethatafilmcompanydirectorwasinterestedinwhethertherewasreallysuchathingasa‘chickflick’(afilmthattypicallyappealstowomenmorethanmen).Hetook20menand20womenandshowedhalfofeachsampleafilmthatwassupposedtobea‘chickflick’(BridgetJones’Diary),andtheotherhalfofeachsampleafilmthatdidn’t fall into the category of ‘chick flick’ (Memento, a brilliant film by the way). In all cases he measured theirphysiologicalarousalasanindicatorofhowmuchtheyenjoyedthefilm.
Weneed topay attention to the ruleof thedataeditor that scores fromdifferent peopleappear in different rowsbecausethisisnothowthedataarelaidoutinthetableabove(tosavespace).Rememberthateachscoreabovecomesfromadifferentperson,thereforewewillneed40rows(not10asabove).ToenterthedataaboveintotheSPSSdataeditorweneedtocreateseveralvariables.IfwebeginwiththevariableArousal,weshouldfollowthesesteps:
Acodingvariables(alsoknownasagroupingvariable)isatypeofvariablethatyouwilluseonnumerousoccasions:itisavariableconsistingofaseriesofnumbersthatrepresentlevelsofanindependentvariable.Inexperiments,codingvariables are used to represent independent variables that have been measured between-groups (i.e. differentparticipantswereassignedtodifferentgroups).So,ifyouweretorunanexperimentwithonegroupofparticipantsinanexperimentalconditionandadifferentgroupofparticipantsinacontrolgroup,youmightassigntheexperimentalgroupacodeof1,andthecontrolgroupacodeof0.Whenyoucometoputthedataintothedataeditor,thenyouwouldcreateavariable(whichyoumightcallgroup)andtypeinthevalue1foranyparticipantsintheexperimentalgroup,anda0 foranyparticipant inthecontrolgroup.This tells thecomputerthatallof thecasesthathavebeen
Hiccupscanbeaseriousproblem:CharlesOsborneapparentlygotacaseofhiccupswhileslaughteringahog(well,whowouldn’t?)thatlasted67years.Peoplehavemanymethodsforstoppinghiccups(asurprise,holdingyourbreath),butactuallymedical sciencehasput itscollectivemind to the task too.Theofficial treatmentmethods include tongue-pullingmanoeuvres,massageofthecarotidartery,and,believeitornot,digitalrectalmassage(Fesmire,1988).Let’ssaywewantedtoputdigitalrectalmassagetothetest(asacureofhiccupsImean).Wetook15hiccupsufferers,andduringaboutofhiccupsadministeredeachofthethreeprocedures(inrandomorderandatintervalsof5minutes)aftertakingabaselineofhowmanyhiccupstheyhadperminute.Wecountedthenumberofhiccupsintheminuteaftereachprocedure.
Table2:Hiccupsdata
Baseline TonguePulling CarotidArteryMassage
DigitalRectalMassage
15 9 7 2
13 18 7 4
9 17 5 4
7 15 10 5
11 18 7 4
14 8 10 3
20 3 7 3
9 16 12 3
17 10 9 4
19 10 8 4
3 14 11 4
13 22 6 4
20 4 13 4
14 16 11 2
13 12 8 3
These data can be entered as above because each person took part in all four conditions – therefore, each rowrepresentsaparticularperson,andeachcolumnrepresentsthenumberofhiccupsthattheyhadduringaparticulartreatment.
Does and Don’ts of Presenting Data Theadventofcomputershasallowedpeopletospendvastamountsoftimeproducingverysnazzylookinggraphs,butIhopetoconvinceyouthatsnazzyisnotalwaysbest!Tuftehaswrittenextensivelyabouthowdatashouldbepresented(e.g.,Tufte,2001).SomeofbookslookatvarioustypesofvisualdisplaysanddiscussessomeofTufte’sadvice(Field,2013,2016);Briefly,Tuftepointsoutthatgraphsshould,amongstotherthings:
ü Showthedata
ü Inducethereadertothinkaboutthedatabeingpresented(ratherthansomeotheraspectofthegraph,likehownicethecolourschemeis).
ü Avoiddistortingthedata
ü Presentmanynumberswithminimumink!
ü Makelargedatasetscoherent
ü Encouragethereadertocomparedifferentpiecesofdata
û Thebarshavea3-Deffect:Neveruse3-Dona2-Dgraphbecauseallitdoesisobscurethedata.Inparticular,itmakesithardtoseethevaluesofthebarsbecauseofthe3-Deffect.Thisgraphisagreatexamplebecausethe3-Deffectmakestheerrorbarsalmostimpossibletoread.
Whenyoufirstusethechartbuildertodrawagraphyouwillseeadialogboxthatseemstosignalan impendingapocalypse. Infact,SPSSisjusthelpfully(?!)remindingyouthatfortheChartBuildertowork,youneedtohavesetthelevel of measurement correctly for eachvariable. That is, when you defined eachvariableyoumusthavesetthemcorrectlytobeScale,OrdinalorNominal.This isbecauseSPSS needs to knowwhether variables are categorical (nominal) or continuous (scale)when it
createsthegraphs.Ifyouhavebeendiligentandsetthesepropertieswhenyouenteredthedatathensimplyclickon tomakethedialogdisappear.Ifyouforgot to set the level of measurement for any variables then click on
Graphing means: bar charts and error bars Howyoucreatebarcharts inSPSSdepends largelyonhowyoucollectedyourdata(whetherthemeanscomefromindependentcasesandare,therefore,independent,orcamefromthesamecasesandsoarerelated).Forthisreason,wewilllookatavarietyofsituations.Figure8showsthevariousoptionsinthechartbuilderundertheoption‘bar’.GivenwhatI’vesaidabove,avoidthe3-Doptions.Themainoptionsthatyou’llusewillbe:
Figure9showsthecompletedChartBuilderforthebarchartandsomeotheroptions.The‘elementproperties’dialogboxshouldappearwhenyouselectthetypeofgraphyouwant,butifitdoesn’tclickon intheChartBuilder.Therearethreeimportantfeaturesofthisdialogbox.Thefirstisthat,bydefault,thebarswilldisplaythemeanvalue.Thisisfine,butjustnotethatyoucanplotothersummarystatisticssuchasthemedianormode.Second,justbecauseyou’veselectedasimplebarchartdoesn’tmeanthatyouhavetohaveabarchart.Also,youcanaskSPSStoadderrorbarstoyourbarcharttocreateanerrorbarchartbyselecting .Youhaveachoiceofwhatyourerrorbars represent. Normally, error bars show the 95% confidence interval, and I have selected this option (
anddragit into .Thiswillmeanthatbarsrepresentingmalesandfemaleswillbedisplayedindifferentcolours.Asintheprevioussection,selecterrorbarsinthepropertiesdialogboxandclickon toapplythemtotheChartBuilder.Figure11showsthecompletedChartBuilder.Clickon toproducethegraph.
representourindependentvariable(i.e.whatwemanipulated—inthiscasethetypeofintervention).SPSSusesthesetemporarynamesbecause it doesn’t knowwhatourparticular variables represent, butwe should change them tosomethingmorehelpful!Justclickon togetridofthisdialogbox.
Weneedtoeditsomeofthepropertiesofthegraph.Figure15showstheoptionsthatneedtobeset:ifyoucan’tseethisdialogboxthenclickon intheChartBuilder.IntheleftpanelofFigure15justnotethatIhaveselectedtodisplayerrorbars(seetheprevioustwosectionsformoreinformation).ThemiddlepanelisaccessedbyclickingonX-Axis1(Bar1)inthelistlabelledEditPropertiesofwhichallowsustoeditpropertiesofthehorizontalaxis.ThefirstthingweneedtodoisgivetheaxisatitleandIhavetypedInterventioninthespacelabelledAxisLabel.Thislabelwillappearonthegraph.Also,wecanchangetheorderofourvariablesifwewanttobyselectingavariableinthelistlabelledOrderandmovingitupordownusing and .Ifwechangeourmindaboutdisplayingoneofourvariablesthenwecanalsoremoveitfromthelistbyselectingitandclickingon .Clickon forthesechangestotakeeffect.TherightpanelofFigure15isaccessedbyclickingonY-Axis1(Bar1)inthelistlabelledEditPropertiesofwhichallowsustoeditpropertiesoftheverticalaxis.ThemainchangethatIhavemadehereistogivetheaxisalabelsothatthefinalgraphhasausefuldescriptionontheaxis(bydefaultitwilljustsayMean,whichisn’tveryhelpful).Ihavetyped‘MeanNumberofHiccupsPerMinute’intheboxlabelledAxisLabel.Alsonotethatyoucanusethisdialogboxtosetthescaleoftheverticalaxis(theminimumvalue,maximumvalueandthemajorincrement,whichishowoftenamarkismadeontheaxis).MostlyyoucanletSPSSconstructthescaleautomaticallyanditwillbefairlysensible—andevenifit’snotyoucanedititlater.Clickon toapplythechanges.
Figure 16: Completed Chart Builder for a repeated-measuresgraph
Figure 17: Bar chart of the mean number of hiccups atbaselineandaftervariousinterventions
Figure16showsthecompletedChartBuilder.Clickon toproducethegraph.TheresultingbarchartinFigure17displays themean (and the confidence intervalof thosemeans)numberofhiccupsatbaselineandafter the threeinterventions.NotethattheaxislabelsthatItypedinhaveappearedonthegraph.Theerrorbarsongraphsofrepeated-measuresdesignsaren’tactuallycorrect(seeField,2013,Chapter9)butyou’llhavetoreadthebookifyouwanttosortthatproblemout!Wecanconcludethattheamountofhiccupsaftertonguepullingwasaboutthesameasatbaseline;however,carotidarterymassagereducedhiccups,butnotbyasmuchasagoodoldfashioneddigital-rectalmassage.Themoralhereis:ifyouhavehiccups,findsomethingdigitalandgoamuseyourselfforafewminutes.
TheChartBuildermightnotbeabletodochartsformultiplerepeated-measuresvariables,but itcangraphwhat isknownasamixeddesign.Thisisadesigninwhichyouhaveoneormoreindependentvariablesmeasuredusingdifferentgroups, and one ormore independent variablesmeasured using the same sample. Basically, the Chart Builder canproduceagraphprovidedyouhaveonlyonevariablethatwasarepeatedmeasure.
Weallliketotext-message.Whatwillhappentothechildren,though?Notonlywilltheydevelopsuper-sizedthumbs,theymightnotlearncorrectwrittenEnglish.Imagineweconductedanexperimentinwhichagroupof25childrenwasencouragedtosendtextmessagesontheirmobilephonesoverasix-monthperiod.Asecondgroupof25childrenwasforbidden fromsending textmessages for thesameperiod.Toensure thatkids in this lattergroupdidn’tuse theirphones, thisgroupwasgivenarmbandsthatadministeredpainfulshocks in thepresenceofmicrowaves (like thoseemittedfromphones).Theoutcomewasascoreonagrammaticaltest(asapercentage)thatwasmeasuredbothbeforeand after the intervention. The first independent variablewas, therefore, textmessageuse (textmessagers versuscontrols)andthesecondindependentvariablewasthetimeatwhichgrammaticalabilitywasassessed(baselineoraftersixmonths).ThedataareinthefileTextMessages.sav.
Tographthesedata,followtheprocedureforgraphingrelatedmeans.Ourrepeated-measuresvariableistime(whethergrammaticalabilitywasmeasuredatbaselineorsixmonths)andisrepresentedinthedatafilebytwocolumns,oneforthebaselinedataandtheotherforthefollow-updata.IntheChartBuilderselectthesetwovariablessimultaneouslybyclickingononeandthenholdingdowntheCtrlkeyonthekeyboardandclickingontheother.Whentheyarebothhighlighted click on either one and drag it into as shown in Figure 18. The second variable (whetherchildrentextmessagedornot)wasmeasuredusingdifferentchildrenandsoisrepresentedinthedatafilebyagrouping
Task 2 Statisticsandmathsanxietyarecommonandaffectpeople’sperformanceonmathsandstatsassignments;womeninparticular can lack confidence in mathematics (Field, 2010). Zhang, Schmader, and Hall (2013) did a study did anintriguing study inwhich students completedamaths test inwhich someput theirownnameon the testbooklet,whereasothersweregivenabookletthatalreadyhadeitheramaleorfemalenameon.Participantsinthelattertwoconditionsweretoldthatthenamesonthebookletsweretoprotecttheiranonymityandthattheywouldbeusingthisotherperson’snameforthepurposeofthetest.Womenwhocompletedthetestusingadifferentnameperformedbetterthanthosewhocompletedthetestusingtheirownname.(Therewerenosucheffectsformen.)Zhangetal.concludedthatperformingunderadifferentnamefreedwomenfromfearsofself-evaluation,allowingthemtoperformbetter.
Task 3 According to some highly unscientific research done by a UK department store chain and reported inMarie Claremagazine(http://ow.ly/9Dxvy)shoppingisgoodforyou:theyfoundthattheaveragewomenspends150minutesandwalks 2.6 miles when she shops, burning off around 385 calories. In contrast, men spend only about 50 minutesshopping,covering1.5miles.Thiswasbasedonstrappingapedometeronamere10participants.AlthoughIdon’thavetheactualdata,somesimulateddatabasedonthesemeansarebelow.EnterthesedataintoSPSSandsavethemasShoppingExercise.sav.
This document is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 InternationalLicense,basicallyyoucanuseitforteachingandnon-profitactivitiesbutnotmeddlewithitwithoutpermissionfromtheauthor.