特立尼达-演讲文章

HowtoWritetheMethodsSectionofaResearchPaper
RichardHKalletMScRRTFAARC
Introduction
B asicResearchConcepts
ContentandWritingStyleofth eMethodsSection
Subjects
EthicalConsideratio ns
Preparations
ProtocolDesign
Measuremen tsandCalculations
DataAnalysis
Summary
Th emethodssectionofaresearchpaperprovidestheinformat ionbywhichastudy’svalidityisjudged.
Therefore,i trequiresaclearandprecisedescriptionofhowanexperim entwasdone,andtherationale
hodssectionshoulddes cribewhatwas
donetoanswertheresearchquestion,de scribehowitwasdone,justifytheexperimentaldesign,an d
ore,the
methodssectionstructureshould:desc ribethematerialsusedinthestudy,explainhowthe
ma terialswerepreparedforthestudy,describetheresearch protocol,explainhowmeasure-
mentsweremadeandwha tcalculationswereperformed,andstatewhichstatistica ltestswere
lelementsofthemethodssectionarewritt en,subsequent
draftsshould
descriptionofprep arations,measurements,andtheprotocolshouldbeorgani zedchronologi-
rity,whenalargeamountofdetailmus tbepresented,informationshouldbe
alineachsectio nshouldbeorganizedby
ds:publications;research;r esearchmethodology;
clinicaltrials;laboratoryre search;writing;mauscripts,medical.[RespirCare2004; 49(10):1229–1232.
©2004DaedalusEnterprises]
Introduction
Themethodssectionisthemostimportan taspectofa
researchpaperbecauseitprovidestheinf ormationbywhich
RichardHKalletMScRRTFAARCisaffi liatedwiththeCardiovascular
ResearchInstitute,a ndwithRespiratoryCareServices,Departmentof
Anes thesia,SanFranciscoGeneralHospital,UniversityofCal ifornia,
SanFrancisco,California.
RichardHKa lletMSRRTFAARCpresentedaversionofthisarticleat
theR
ESPIRATORY
C
ARE
Journalsymposium ,“AnatomyofaResearch
Paper:ScienceWriting101,”a tthe48thInternationalRespiratoryCon-
gress,held October5–8,2002,inTampa,Florida.
ore,the
aut hormustprovideaclearandprecisedescriptionofhow
anexperimentwasdone,andtherationaleforthespecific< br>bewrittenwith
enoughinformationsothat:(1)the experimentcouldbe
repeatedbyotherstoevaluatewhe thertheresultsarere-
producible,and(2)theaudien cecanjudgewhetherthe
articleIdescribe
Corres pondence:RichardHKalletMSRRTFAARC,RespiratoryCare< br>Services,SanFranciscoGeneralHospital,NH:GA-2,10 01PotreroAv-
enue,SanFranciscoCA.94110.E-mail:r kallet@.
R
ESPIRATORY
C
ARE
•O
C TOBER
2004V
OL
49N
O
101229

< br>H
OWTO
W
RITETHE
M
ETHODS
SECTIONOFA
R
ESEARCH
P
APER
ethis< br>sectionissointimatelyrelatedtotheprinciplesofsc ien-
tificresearch,Ibeginwithareviewofbasicrese arch
concepts,andthenfollowwithadiscussionofimp ortant
pointstoincorporatewhenwritingthemethods section.
BasicResearchConcepts
Thescientific methodattemptstodiscovercause-and-
effectrelati onshipsbetweenobjects(ie,physicalmatteror
proce sses).Inthephysicalsciencesobjectsareregardedas
variables,andavariableisanythingthatcanassumedif-
atingacause-and-effectrelationship
betweeno bjectsrequiresthatvariablesareclassifiedas
inde pendent,dependent,pendent
variableisonethat,whe nmanipulated,causesachangein
iablethatchangesin responseto
thatmanipulationisreferredtoasadepen dentvariable.
Forexample,arterialoxygentensioni sadependentvari-
ablethatrespondstomanipulation sinindependentvari-
ablessuchasbarometricpressu reandoxygenconcentra-
undingorextraneousvariabl eisanything
otherthantheindependentvariableofin terestthatmay
ore,achangeina
dependentvariab lemaybeduewhollyorinparttoa
mple,achange
inm inuteventilationcanalterarterialoxygentensionby
itseffectuponalveolarcarbondioxidepartialpressure .
Evaluationofapotentialcause-effectrelationshi pbe-
tween2objectsisaccomplishedthroughthedevel opment
designissimplyastrategyto
controlandm anipulatevariablesthatprovideananswerto
therese archquestionregardingpotentialcause-and-effect
relationships.
Validityreferstothecredibilityof experimentalresults
andthedegreetowhichtheresul tscanbeappliedtothe
alvalidityrefersto
thecr edibilityofastudyandisdeterminedbythedegree
tow hichconclusionsdrawnfromanexperimentcorrectly
d escribewhatactuallytranspiredduringthestudy.
1< br>Exter-
nalvalidityreferstowhether(andtowhatd egree)the
resultsofastudycanbegeneralizedtoalar gerpopula-
tion.
1
Unfortunately,allbiolo gicalsystemsareprofoundly
complex,sosimple,unam biguous,directrelationshipsbe-
ernal
validit yofastudyisjudgedbythedegreetowhichits
outcomes canbeattributedtomanipulationofindependent
vari ablesandnottotheeffectsofconfoundingvariables.
Therefore,thestudyprotocolmustbedesignedtocontrol< br>(eg,tokeepconstant)asmanyextraneousfactorsaspos -
siblesothatanypotentialcause-and-effectrelati onship
portant
1230
toemphasizethatconfou ndingvariablescanneverbefully
rmore,theinfluenc eofthesevariables
maynotbefullyappreciatedbytho seconductingthere-
alvalidityisprimarilydetermi nedbyhow
subjectsareselectedtoparticipateinastu dyandbythe
useofrandomizationproceduresthatlimi tpotentialbiasin
howsubjectsareassignedtotreatm entgroups.
ContentandWritingStyleoftheMethodsSe ction
Historically,themethodssectionwasreferred toasthe
“materialsandmethods”toemphasizethe2dis tinctareas
thatmustbeaddressed.“Materials”refer redtowhatwas
examined(eg,humans,animals,tissuep reparations)and
alsotothevarioustreatments(eg,d rugs,gases)andin-
struments(eg,ventilators)used inthestudy.“Methods”
referredtohowsubjectsorobj ectsweremanipulatedto
answertheexperimentalques tion,howmeasurementsand
calculationsweremade,an dhowthedatawereanalyzed.
Thecomplexityofscienti ficinquirynecessitatesthat
thewritingofthemetho dsbeclearandorderlytoavoid
,itisusuallyhelpfult o
structurethemethodssectionby:
bingthemater ialsusedinthestudy
ninghowthematerialswereprepa red
bingtheresearchprotocol
ninghowmeasureme ntsweremadeandwhat
calculationswereperformed
gwhichstatisticaltestsweredonetoanalyze
thedat a
2
Second,thewritingshouldbedirectandprecis eandin
ndsentencestructuresshouldbe
avoided, aswellasdescriptionsofunimportantdetails.
Oncea llelementsofthemethodssectionarewrittendown
dur ingtheinitialdraft,subsequentdraftsshouldfocusonhowtopresentthoseelementsasclearlyandlogicallyas
ral,thedescriptionofpreparations,mea-
surem ents,andtheprotocolshouldbeorganizedchrono-
rit y,whenalargeamountofdetailmust
bepresented,info rmationshouldbepresentedinsubsec-
eachsectionan dsubsec-
tion,materialshouldalwaysbeorganizedby topicfrom
mosttoleastimportant.
Subjects
Judgingtheexternalvalidityofastudyinvolvinghu-
mansubjects(ie,towhomthestudyresultsmaybeap-
pl ied)requiresthatdescriptivedatabeprovidedregarding
thebasicdemographicprofileofthesamplepopulatio n,
includingage,gender,andpossiblytheracialcomp osition
imalsarethesubjectsofastudy,it
isimp ortanttolistspecies,weight,strain,sex,andage.
R
ESPIRATORY
C
ARE
•O
CTOBER
2004 V
OL
49N
O
10

H
OWTO
W
RITETHE
M
ETHODS
S
ECTIONOFA
R< br>ESEARCH
P
APER
Whoischosenforinclusion inastudy(aswellashow
treatmentsareassigned)inla rgemeasuredetermineswhat
limitsareplacedonthege neralizationsthatcanbemade
,whenwritingthemeth-
odssection,itisimportanttodescribewhothesubjec ts
ection
criteriaandrationaleforenrollingpa tientsintothestudy
mple,ifthestudypro-
claim stoexaminewhethernoninvasiveventilationreduces
theneedforintubationofpatientswithcardiogenicpul-< br>monaryedema,thenonewouldnotanticipatethatsurgic al
patientswithrespiratoryfailurewouldberecruit ed.
Inaddition,itisimportantwhendescribingpatie ntsto
providesomeevaluationoftheirhealthstatust hatisrel-
mple,whenexaminingtherapies
thatma yimpactmortalityinacutelyillpatients,thestudy
s ubjects’healthstatuscanbeassessedwithascoringsys-< br>temsuchastheSimplifiedAcutePhysiologyScore.
3
If
studyingpatientsinarehabilitationsettin g,thenageneral
quality-of-lifequestionnairesuch astheSicknessImpact
Profilecanbeused.
4
E thicalConsiderations
Whenworkingwithhumanoranim alsubjects,there
mustbeadeclarationthatthemedic alcenter’sinstitutional
reviewboardgoverningres earchonlivingmatterhasde-
terminedthatthestudyp rotocoladherestoethicalprinci-
tsuchapproval,no researchprojectcanbe
conductednorcanitbepublish edinareputable,peer-
reviewsciencejournal.
P reparations
Instudiesinvolvinganimalmodelsormec hanicalmod-
els,adetaileddescriptionmustbeprovi dedregardingthe
preparationsmadepriortobeginnin gtheexperimentalpro-
iesinvolvinganimalsadetail eddescription
shouldbeprovidedontheuseofsedatio nandanesthesia,
therouteofadministration,andhow itsefficacywaseval-
uated.
2
Inaddition,a llaspectsofanimalortissueprepa-
rationrequiredp riortoinitiationoftheresearchprotocol
yanimalpr eparation
ormechanicalmodeltheremustbeenoughdet ailprovided
sothatthereadercanduplicateitoreval uateitsrelevance.
Whenastudyinvolvestheuseoreva luationofdrugs,the
genericdrugnameshouldbeuseda ndthemanufacturer,
concentration,dose,andinfusi onrateshouldbespecified.
Likewise,whenmedicalga sesareused,theconcentration
andflowratesshouldb especified.
Itisworthnotingthattheintroductiono fanynovel
methodformeasuringavariable,orprepari ngdesigninga
ingonhow
unique(orunorthodox)th enewmethodis,itsvalidation
R
ESPIRATORY
C
ARE
•O
CTOBER
2004V
OL
49N
O
10
probablyshouldbeestablishedinaseparatepu blication,
publishedpriortosubmissionofthemains tudy.
ProtocolDesign
Theresearchprotocolisth esequenceofmanipulations
andmeasurementprocedur esthatmakeuptheexperiment.
Itsdescriptionshould followtheexactsequenceofhow
theprocedureswereex ecuted.
2
Typically,thisfirstin-
volvesad escriptionofbaselineconditionsandanyasso-
ciate dbaselinemeasurements,followedbythesequence
ofm anipulationsoftheindependentvariableandthesub-
sequentmeasurementofchangesinthedependentvari-
soimportanttodescribeallrelevantaspectsof
clini calmanagementnotcontrolledbytheprotocolinthe
pe ri-experimentalperiod.
Whenwritingthemethodssec tion,itisimportantto
bearinmindthattherationale orassumptionsonwhich
someproceduresarebasedmayn otalwaysbeobviousto
particularlytruewhenwriting fora
generalmedicalaudience,asopposedtomemberso fa
ore,thewritermustalwayskeepin
ionaleandas sump-
tionsonwhichexperimentalproceduresarebase dshould
bebrieflystatedinthemethodssectionand,i fnecessary,
-
everitisnotobvious,thepurposeo faprocedureshouldbe
statedinrelationshipeithert otheresearchquestionorto
gthemethodssectioninth is
styleiscalledapurpose-procedureformat.
2< br>MeasurementsandCalculations
Thenextstepinthe methodssectionistodescribewhat
variablesweremea suredandhowthosemeasurementswere
criptionofmeas urementinstrumentsshould
includethemanufacturer andmodel,calibrationproce-
dures,maybe
neces sarytojustifywhyandhowcertainvariableswere
come sparticularlyimportantwhenthe
objectoftheexperi mentcanbeapproachedonlyindi-
tially,wheneverava lueforavariableis
usedtosignifyastateorconditio n,thisshouldbestated
mple,onecouldstate:“Adequa teintra-
vascularvolumestatuswasindicatedbyacen tralvenous
pressureofՆ8mmHg.”Alistingofallcalcu lationsused
inthestudytypicallyfollowsthedescri ptionofmeasure-
ments.
DataAnalysis
Thela ststepinthemethodssectionistodescribehow
thedat awillbepresentedintheresultssection(eg,mean
vsm edian),whichstatisticaltestswillusedfortheinfer-1231

H
OWTO
W
RITETHE
M
ETHODS
S
ECTIONOFA
R
ESEARCH
P
APER
REFERENCES
entialdata,andwhatpvalueisde emedtoindicateasta-
tisticallysignificantdiffer ence.
Summary
Themethodssectionisthemostimpo rtantpartofa
researchpaperbecauseitprovidesthei nformationthe
readerneedstojudgethestudy’ingaclearandprecisedescriptionofhowanexperimentwasdone,andtherationaleforspecificexperimentalproce -
duresarecrucialaspectsofscientificwriting.
SB,NewmanTB,tomyandphysi-
:HulleySB,CummingsSR (editors).Designing
ore:William&Wilkins;1988:1– 11.

York:McGraw-Hill;1991:113–138.
JR,Le meshowS,mplifiedAcutePhys-
iologyScore(SAPSII)b asedonaEuropeanNorthAmericanmul-
1993;270(24):2 957–2963.
rM,BobbittRA,CarterWB,knessIm-
pac tProfile:developmentandfinalrevisionofahealthstatu smea-
e1981;19(8):787–805.
1232R
ESPIRATO RY
C
ARE
•O
CTOBER
2004V
OL
4 9N
O
10