ResearchMethods
WHYRESEARCHMETHODS?
EventhoughtheUMATdoesnot require youtohaveknowledgeofresearchmethods,
manyofthepassagesandmuchofthedatapresentedareintheformofresearchtrialsor
investigations. If you understand some basic concepts before you go into the UMAT,
youllspendlesstimeinterpretingthemintheexamandthushavemoretimetoanswer
thequestions.
THEBASICS
Howresearchersthink
Researchers use two types of logic: generaltospecific thinking and specifictogeneral
thinking.
Generaltospecific thinking occurswhen researchers havealready discovered alawof
natureandwishtoapplytheirknowledgetoaspecificarea.Forexample,aresearcher
maywishtoknowhowfastapebblewillbefallingwhenithitsthegroundthreeseconds
afterbeingdropped.Thereisalawofphysicsfromwhichitcanbedeterminedthat:On
Earth,fallingobjectsaccelerateatarateofabout9.8m/sec2.Theresearchercoulduse
thisknowngeneralprincipletocalculatethespecificinformationthatheorsheneeds.
Here,theresearcherusesageneral principle(the acceleration ofanyobject falling on
Earth)tofindaspecificfact(thespeedofthepebble).
Specifictogeneralthinkingisadifferentkindofthinkinginordertodiscoveranewlaw
ofnature.Inthiscase,researchersexaminemanyspecificfactsandthendrawageneral
conclusionaboutwhattheyhaveseen.Forexample,aresearchermightwatchhundreds of
differentkindsoffrogsliveanddie,andmightnoticethatallofthemdevelopedfrom
tadpoles.Heorshemaythenannounceaconclusion:Allfrogsdevelopfromtadpoles.
Thistypeoflogicisspecifictogeneral.Theresearcherlooksatmanyspecificfrogstofind
ageneralruleaboutallfrogs.
Thisconclusioniscalledahypothesis,notafactoratruth,becausetheresearcherhas
notcheckedeverysinglefrogintheuniverse.Heorsheknowsthattheretheoretically
couldbeafrogsomewherethatgrowsfromsomethingelse.Butuntilheorshefindssuch
afrog,itisreasonabletothinkthatherhypothesisiscorrect.Manyhypotheses,in fact,
aresowelldocumentedthattheybecometheequivalentoflawsinnature.
In your science classes in school, you mostly learn about generaltospecific thinking.
Yourteachersexplaingeneralrulesofsciencetoyouandthenexpectyoutoapplythese
rulestoanswerquestionsandsoleproblems.ThemajorityofUMATquestionsinvolving
researchmethods,however,involvespecifictogeneralthinking.Thequestionstestyour
abilitytoseethekindsofpatternsinspecificdatathat,asaresearcher,youwoulduseto
formulateyourowngeneralhypotheses.
Howexperimentswork
Many UMAT stimuli describe experiments. Experiments help researchers do specific
togeneralthinkinginareliableandefficientway.Considerthetadpoleresearcherabove.
Inarealworldsituation,whatwouldprobablyhappen isthatheorshewould notice
some of the frogsdevelop fromtadpolesand wonderif maybetheyall did. Then she
wouldknowwhattolookforandcouldcheckallthefrogssystematically.Thisprocess
describesthetwobasicstepsofanyexperiment:
Formingahypothesis(guessingthatallfrogscomefromtadpoles)
Testingahypothesis(checkingthefrogstoseeifthisguesswasright)
Researchersareofteninterestedincauseandeffect relationships.Havingformedhis or
her hypothesis about tadpoles, a researcher might wonder what causes a tadpole to
becomeafrog.Totestcausalrelationships,aspecialkindofexperimentisneeded.Heor
shemusttestonepossiblecauseatatimeinordertoisolatewhichoneactuallyproduces
the effect in question. For example, the researcher might inject tadpoles with several
differentkindsofhormones.Someofthesetadpolesmightdie.Othersmightturninto
frogsnormally.ButafewthoseinjectedwithhormoneXforexamplemightremain
tadpolesforanindefinitetime.OnereasonableexplanationisthathormoneXinsome
wayinhibitedwhatevercausesnormalfrogdevelopment.Inotherwords,theresearcher
wouldhypothesiseacausalrelationshipbetweenhormoneXandfrogdevelopment.
The relationship between hormone X and frog development, however, would not be
demonstratedverywellifthescientistalsofeddifferentdietstodifferenttadpoles,kept
someinwarmerwater,orallowedsometohavemoreroomtoswimthanothersorifshe
did not also monitor tadpoles who were injected with no hormones at all but who
otherwisewerekeptunderthesameconditionsasthetreatedtadpoles.Why?Becauseif
theeternaltadpoleshaddietsthatdifferedfromthatoftheothers,thescientist would
notknowwhetheritwashormoneXorthespecialdietthatkepttheeternaltadpolesfrom
becomingfrogs.Moreover,ifthewaterwaswarmerthanthatoftheothers,maybeitwas
thewarmththatsomehowkeptthetadpolesfromdeveloping.Andiftheresearcherdid
notmonitoruntreatedtadpoles(acontrolgroup),heorshecouldnotbesurewhether
underthesameconditionsanormal,untreatedtadpolewouldalsoremainundeveloped.
Therefore,ascientistcreatingawelldesignedexperimentwill:
Ensurethatthereisasinglevariable(likehormoneX)thatvariesfromtesttotestor
grouptogroup
Ensurethatallotherfactors(diet,temperature,space,etc.)remainthesame
Ensurethatthereisacontrolgroup(tadpoleswhodontgetanyhormoneX)for
comparisonpurposes
WheneveryouseeanexperimentintheUMAT,youshouldaskyourself:
Whatisthefactorthatisbeingvaried?Thatiswhatisbeingtested.
Whatisthecontrolgroup?Itisthegroupthatisnotreceivingthetreatmentetc.
Whatdotheresultsshow?Whatdifferencesexistbetweentheresultsforthecontrol
groupandthosefortheothergroup(s)?Orbetweentheresultsforonetreatedgroupand
thoseforanother,differentlytreatedgroup?
DEFINITIONS
Here are some common words that are used in research methods that you should
familiariseyourselfwith.
Participants/Subjects
Participants are people / animals that are involved in the experiment. They are often
dividedintogroupsandeithergivenatreatment/interventionoraplacebo.
Population
Apopulationisawelldefinedgroupinwhichmembershaveparticularcharacteristics(for
example,maleofficeworkerslivinginAustralia).Participantsinastudyaretakenfroma
particularpopulation.
ControlGroup
The participants in the control group are used for comparison with the experimental
group.
Forexample,letssaythatDrugAisbeingtestedforitsusefulnessinpreventingheart
disease.Theparticipantsofthetrialmightbedividedintotwogroupsthecontrolgroup
andtheexperimentalgroup.TheexperimentalgroupwouldbegivenDrugAwhilethe
controlgroupwouldbegivennotreatmentoraplacebo.Attheendofthetrial,thetwo
groupsarecomparedtoseewhetherDrugAhelpedpreventheartdisease.
ExperimentalGroup
Theparticipantsintheexperimentalgroupreceivethetreatment,drugorinterventionthat
isbeinginvestigated.Seecontrolgroupforanexampleofanexperimentalgroup.
Placebo
Aplaceboisasubstancethatlookslikethetreatment/interventionbutisactuallyinert.
Forexample,whenthetreatmentisadrugpill,asugarpillcouldbetheplacebo. In
manyexperiments,theparticipants inthecontrolgrouparegivenaplacebo.
Placeboeffect
Theplaceboeffectoccurswhenthebelieforknowledgethatoneisbeingtreatedcanitself
haveaphysiologicaleffect.Forexample,somepeoplewhotakeapillthattheythinkisa
drug(butwhichactuallyhasnoactivesubstance)canimproveintheirconditionsimply
due to the beliefthat they took adrug that would work. This can confound research
findings. Singleanddoubleblind trialsareusedtoreducethedetrimentaleffectofthis
phenomenononresearch.
Experimenterbias
Experimenterbiasisbiasthatisintroducedbyanexperimenter,whoseexpectationsabout
theoutcomeof the experiment can besubtlycommunicated to theparticipantsin the
experiment.Forexample,ifanexperimenterwastestinghowgoodDrugEwasinhelping
withdepression,hisexpectationsthatDrugEwouldworkcouldbecommunicatedtothe
patients,thusaffectingtheoutcomeoftheexperiment.
Singleblindtrial
In a single blind trial, participants do not know whether theyare in the experimental
grouporthecontrolgroup.Thishelpsreduceanyerrorsduetotheplaceboeffect.
Doubleblindtrial
In a double blind trial, both participants and experimenters do not know which
participantsareinthe experimentalgroup orthe controlgroup.Thishelpsreduceany
errorsduetotheplaceboeffectaswellasanyexperimenterbias.
IndependentVariable
Anindependentvariableisthefactor(s)whoseeffectsaretobestudiedandmanipulated
inanexperiment.Forexample,ifanexperimenterisexaminingthewaydifferentfertilisers
helpaplantgrow,theindependentvariableisthetypeoffertiliser(i.e.thevariable(s)that
theexperimenterchanges).
DependentVariable
Dependent variables change in response to changes in the independent variable. For
example,ifanexperimenterisexaminingthewaydifferentfertilisershelpaplantgrow,
thedependentvariableisthegrowthoftheplants(i.e.thevariablethathelpsmeasurethe
effectsoftheindependentvariable).
Hypothesis
Ahypothesisisatentativeexplanationofobservedeventsthatcanbefurthertested.Itisa
predictionoftherelationshipbetweenanindependentvariableanddependentvariablein
anexperiment.Forexample,ahypothesiscouldbeDrugAwillreducetheriskofhaving
aheartattackinfemalesabove50yearsofage.Thiscanbetestedinanexperiment.
Itisimportanttonotethatahypothesiscanneverbeproven.Anexperimentmayshowit
to be false, but experiments cannot prove it to be absolutely true. However, a
wellcontrolledexperimentcanhelpsupportahypothesis.
Sampling
Samplingistheprocessofselectingasubgroupofparticipantsfromthepopulationsothat
thesubgroupcanbeusedtomakegeneralisationstothepopulationasawhole.Ingeneral,
thegreaterthenumberofparticipantsinthesampleandthemoreeffectivetheselection
procedure,thebetterthesamplerepresentsthepopulationasawhole.
Randomisation
Randomisation isatechniqueofassigningpatientsto experimental and controlgroups
thatisbasedonlyonchancedistribution.Byusingrandomisation,anyselectionbias is
diminished and variables between the experimental and control groups are more
effectivelycontrolled.
Questions in the UMAT most commonly focus on the concepts of sampling and
controlling variables (which include the concepts of the placebo effect and
randomisation).
SAMPLING
In a research study, it is important to have participants that are representative of the
populationunderinvestigationsothatinferencescanbemadeaboutthepopulationasa
whole.Forexample,ifadrugwasbeinginvestigated,buttheparticipantsinthestudy
werenotsimilartothepopulation,itwouldbedifficulttorecommendthatdrugtothe
population,eveniftheexperimentshowedittobesuccessful.
Letstakeanexample:
AresearcherwantstoinvestigatetheusefulnessofdrugAinhelpingmalesmokersaged
4045stopsmoking.
Independentvariable:BeinggivenDrugAorbeinggivenaplacebo
Dependentvariable:Whetherornottheparticipantstopssmoking
Inthiscase,thepopulationismalesmokersaged4045years.
Ofcourse,itisimpossibletohaveallmalesbetween40and45yearsofageasparticipants
intheexperiment.Asampleneedstobetakenfromthepopulation.Asampleisasubset
ofpeoplethatneedstohavesimilarcharacteristicstothepopulation.
Theselectionprocedureisvitalinensuringthatthesampleisrepresentativeofthe
populationunderinvestigation.Forexample,ifallofyourparticipantswerefromone
suburbinouterSydney,yoursamplewouldnotberepresentativeofallsmokersaged
4045yearsinAustralia.
Thenumberofpeopleinthesampleisalsoimportant.Forexample,5000malesmokers
aged between 40 and 45 is likely to give a more representative sample than 50 male
smokersaged40to45years.
CONTROLLING VARIABLES
Controllingextraneousvariables
One of the most important elements of experimental design is controlling extraneous
variables.Foranexperimenttobeeffective, allvariablesexceptfortheindependent
variable(theoneunderinvestigation)shouldbekeptconstant.Ifthisisdone,then
anychangesinthedependentvariablecanbeattributedtotheindependentvariable.
Letsconsideranexampleofhowanuncontrolledvariablecanimpactonourstudy.
Supposethatthesmokersinthecontrolgroupweregiventheirplaceboandtoldthat it
containedvitamins.Thoseintheexperimentalgroupweregiventheirdrugandtoldthatit
wouldhelpthemquitsmoking.Whenthestudywascompleted,thoseintheexperimental
groupwerefoundtohaveamuchgreaterrateofquittingthanthoseinthecontrolgroup.
Butwhatisthisdifferencedueto?Isitduetotheactivityofthedrug?Orisitduetothe
factthatthoseintheexperimentalgroupbelievedtheywerebeinghelped?
This example illustrates two concepts. Firstly, it demonstrates the importance of the
placeboeffectininfluencinganexperiment.Secondly,itshowsthedegreetowhich an
uncontrolledvariablecanaffecttheoutcomeofanexperiment.
Extraneous variables therefore need to be controlled and kept constant in order to
ensuretheeffectivenessofthestudy.
Experimentalandcontrolgroups
One key variable that must be controlled in an experiment is the differencesbetween
controlandexperimentalgroups.Itdifferenceisnotminimised,anydisparityinresults
between experimental and control groups could be due to participant characteristics,
ratherthananytreatment.
Ofcourse,wecannotguaranteethateachofthegroupswillbeidenticaltoeachotherin
everywayeventwinsaredifferent.Thebestthatwecandoisallocateparticipantsto
differentgroupsrandomly.Ifwedothis,thechancesarethegroupswillbesimilar.
Inourstudy,itisimportantthattheexperimentalandcontrolgrouphadparticipantsthat
weresimilarinparticularways.Forexample,ifalloftheparticipantsinonegroup had
spousesthatsmoked,andtheothergroupdidnot,thismayconfoundtheexperiment.
EXAMPLE
Inthisstudy,adultChinchillarabbitswerefittedwithspeciallymadesoftcontact lenses
in order to determine the effects of lens use on the health of corneas (the thin outer
coveringoftheeyeball).InGroup1(15rabbits),corneaswerenotfittedwithlenses.In
Groups2and3(15rabbitseach),lenseswereleftontheeyesfortwoweeksandthen
removed.Approximatelyhalfthecorneaswereexaminedimmediatelyupon removalof
the contacts, while the other half were examined five hours after lens removal. The
corneasofGroup2rabbitsweretreatedwithcatalase(anoxygenscavengingenzymeor
protein) over the twoweek period the corneas of Group 3 rabbits were treated with
3aminotriazole (a catalase inhibitor) during the period of lens use. All animals were
examined daily, and the corneas were documented photographically. Additionally,
enzymeactivityinthecorneaswasexaminedatthreestagesofthestudy:atthebeginning
of the twoweek period upon lens removal and, in some cases, five hours after lens
removal.
Inrabbitcorneasnotfittedwithcontactlenses,theactivityofxanthineoxidase(anoxygen
generatingenzymeorprotein)and catalasewerebalanced andpresentinhigh levels
both beforeand after the study. In rabbit corneas fitted with lenses, xanthineoxidase
activitywashighandcomparabletolevelsfoundinthecontrolgroup.Ingeneral,though,
catalaseactivitywasfoundtobesignificantlyreduceduponremovalofthelenses.But
whencorneasweretreatedwithcatalaseduringlensuse,thisdecreaseincatalaseactivity
wasprevented.Incontrast,aftertreatmentwiththecatalaseinhibitor3aminotriazole,the
activity of the catalase enzyme was nearly absent, and rabbits in this group suffered
extremeinflammationoftheeye.Further,forthoserabbitswhosecorneaswereexamined
5 hours afterlensremoval, imbalanceswere also found to exist between theenzymes
lactatedehydrogenaseandlysosomalhydrolase.
Question1
According to the design of the study, which group of rabbits constitutes the control
group?
A)rabbitsinGroup1
B)rabbitsinGroup2
C)rabbitsinGroup3
D)adultChinchillarabbitsfoundinnature
Answer:A
Solution:Rabbitsingroup1werenotfittedwithcontactlenses,thoughcornealenzyme
activitywasdocumentedforthisgroupofrabbits.Itisthereforethecontrolgroup and
optionAisthecorrectanswer.
Question2
Accrdingtothedesignofthestudy,themostlikelyreasonforexaminingsomecorneasimmediately
uponremovalofcontactlensesandsomecorneasfivehourslateristodetermine
A)theactivitylevelsofxanthineoxidaseandcatalase.
B)theactivitylevelsoflactatedehydrogenaseandlysosomalhydrolase.
C)whetheradditionaldisturbancesappearafterlensremoval.
D)whetheradditionaldisturbancesappearbeforelensremoval.
Answer:C
Solution: Scientists recorded enzyme activity levels immediately upon removal of the
contactlensesandfivehourslatertodeterminewhetherornotadditionaldisturbances
appearafterlensremoval.ThereforeoptionCisthebestanswer.
Question3
The results ofthisstudy suggest that prolonged useofcontact lenses can damagethe
cornea, primarily resulting from imbalances in the activity levels of which pair of
enzymes?
A)catalaseand3aminotriazole
B)xanthineoxidaseandcatalase
C)xanthineoxidaseand3aminotriazole
D)dehydrogenaseandlysosomalhydrolase
Answer:B
Solution:In healthycorneaswithoutcontact lenses,activitylevelsofxanthineoxidase
andcatalasearepresentandhigh.Wearingcontactlensesreducesthelevelsofcatalase,
resultinginanimbalanceinthecorneabetweenthetwoenzymes.ThereforeoptionBis
thecorrectanswer.
Question4
Fromtheresultsofthestudy,itcanbehypothesizedthatsupplementingcontactlensuse
withdropsofcatalase
A)isinsignificant.
B)candamagethecorneas.
C)canpreventsomedamagetothecorneas.
D)caninhibitproductionof3aminotriazole.
Answer:C
Solution:Treatingeyeswithcatalaseduringlensusepreventedthecatalaseactivitylevels
from dropping. Treating eyes with 3aminotriazole during lens use caused the activity
levelstodrop.Corneastreatedwith3aminotriazolewerefounduponconclusionofthe
studyto beinflamed. It can thereforebe hypothesized that catalasecan preventsome
damagetothecorneas.ThereforeoptionCisthecorrectanswer.
Question5
From the results of the study, which of the following statements holds true for the
relationshipbetweencatalaseandinflammationoftheeye?
A)Thereisnorelationshipbetweencatalaseandinflammation.
B)Excessiveamountsofcatalasecanleadtoinflammation.
C)Presenceofcatalasecanleadtoinflammation.
D)Absenceofcatalasecanleadtoinflammation.
Answer:D
Treatingeyeswith3aminotriazoleduringlenswearcausedtheactivitylevelsofcatalase
todrop.Theseeyeswerelaterfoundtobesufferingfrominflammation.Itcanthereforebe
hypothesizedthattheabsenceofcatalaseleadstoinflammation.ThusoptionDisthebest
answer.
Question6
Thestudysuggeststhatinthecorneasofnoncontactlenswearers,theenzymeactivityof
cornealcatalaseispresentin
A)highlevelscomparabletolevelsof3aminotriazole.
B)lowlevelscomparabletolevelsof3aminotriazole.
C)highlevelscomparabletolevelsofxanthineoxidase.
D)lowlevelscomparabletolevelsofxanthineoxidase.
Answer:C
Incorneasofthecontrolgroupofrabbits,theenzymaticactivitylevelsofcatalase and
xanthineoxidasewerepresentandhigh.ThereforeoptionCisthebestanswer.
THEIDEALSTUDY
Insummary,followingarethemaincomponentsofanidealstudy:
Goodsamplingtechnique(largenumberofparticipants,representativesample)
Randomisedallocationtocontrolandexperimentalgroups
Doubleblindproceduretocontroltheplaceboeffectandexperimenterbias
Allvariablesapartfromthosebeingmanipulatedarecontrolled
Itisworthwhilekeepingthislistinmindwhenyouarereadingthrougharesearchstudy.
Itisalsohelpfultohaveagoodunderstandingofwhattheexperimenteristestingortrying
toachievewiththestudy.
