Research Methods for Research Methods for the Social Sciences:the Social Sciences:Research Design Part IIResearch Design Part II
Tuesday, February 02, 2010 Tuesday, February 02, 2010
Data Collection & ProceduresData Collection & Procedures
Division on Addictions, Cambridge Health Division on Addictions, Cambridge Health AllianceAlliance
Harvard Medical SchoolHarvard Medical School
Class ObjectivesClass Objectives
Review:Review:– SamplingSampling– Qualitative vs. QuantitativeQualitative vs. Quantitative– Data collection proceduresData collection procedures– Your research ideasYour research ideas
SamplingSampling Selecting a group of Selecting a group of
people from a larger people from a larger population.population.
Why?Why?– SizeSize– TimeTime– $$$$$$
Sampling errorSampling error– Sampling biasSampling bias
Types of samplesTypes of samples
NonprobabilityNonprobability– ConvenienceConvenience– QuotaQuota– Purposive or judgmental Purposive or judgmental – Snowball Snowball
Types of samplesTypes of samples
ProbabilityProbability– Simple randomSimple random
Select at random using random number Select at random using random number generatorgenerator
Same chanceSame chance
– SystematicSystematic 1,000/100 = 101,000/100 = 10thth
Types of samplesTypes of samples– StratifiedStratified
Divide population into strataDivide population into strata Sample = population on a specific characteristicSample = population on a specific characteristic
– E.g., AGEE.g., AGE
ClusterCluster
Examples of sampling
Special populations– College students -> Other populations – Treatment Seekers -> Everyone with a
specific problem
Who’s collecting the data?Who’s collecting the data?
SecondarySecondary– Data collected by othersData collected by others
E.g., Researchers, E.g., Researchers, governmentgovernment
– Strengths:Strengths: Save time & moneySave time & money
– Weaknesses:Weaknesses: No control over study No control over study
design or data collectiondesign or data collection
Who’s collecting the data?Who’s collecting the data?
PrimaryPrimary– Data you collect on Data you collect on
your ownyour own– CollaborationsCollaborations– StrengthsStrengths
You control study You control study design, sampling, data design, sampling, data collectioncollection
– WeaknessWeakness Time &”MoneyTime &”Money
Sample sizeSample size
Power analysisPower analysis1.1. Sample size n = ???Sample size n = ???
2.2. Ability to measure effectsAbility to measure effects
What kind of data are you What kind of data are you collectioncollection
Qualitative Quantitative
Experimental Correlation SurveyInteractive Interview
NaturalisticObservation
CaseStudy
Written Description
What type of data are you What type of data are you collecting?collecting?
QualitativeQualitative Systematic but nonuniformSystematic but nonuniform Field work or participant observationField work or participant observation
– Natural observationNatural observation– Case studyCase study– Interactive interviewsInteractive interviews– Written descriptions by participantsWritten descriptions by participants
Open-endedOpen-ended Record everythingRecord everything
Natural ObservationNatural Observation
Narrative recordNarrative record Video/audio recordingsVideo/audio recordings SamplingSampling
– TimeTime– EventEvent– SituationSituation
After you will categorize and code After you will categorize and code data for analysisdata for analysis
Case StudyCase Study
Review: Biographical Review: Biographical data; Medical records; data; Medical records; Family history; Family history; Observations; Interviews; Observations; Interviews; Psychological testsPsychological tests
Look for variables that Look for variables that might help explain might help explain behavior your are behavior your are studying.studying.
Interactive Interview/focus Interactive Interview/focus groupgroup
Similar to natural observationSimilar to natural observation– InteractInteract
Ask specific questions and lead Ask specific questions and lead discussiondiscussion
Data collection methods similar to Data collection methods similar to natural observationnatural observation
Written descriptionWritten description
JournalJournal Long narrativeLong narrative Open-ended survey responsesOpen-ended survey responses
CodingCoding
Simple coding: Counting the number Simple coding: Counting the number of behaviors in a recordingof behaviors in a recording
More complex coding: CategorizationMore complex coding: Categorization
Qualitative Qualitative
Content Analysis of open ended Content Analysis of open ended responsesresponses– Missouri Self-excludersMissouri Self-excluders– Question: Why did you self-exclude?Question: Why did you self-exclude?– ““Because my wife made me.”Because my wife made me.”– ““I was sick of worrying about finances.”I was sick of worrying about finances.”
New computer programs can New computer programs can categorize for you.categorize for you.
What kind of data are you What kind of data are you collectioncollection
Qualitative Quantitative
Experimental Correlation SurveyInteractive Interview
NaturalisticObservation
CaseStudy
Written Description
What type of data are you What type of data are you collecting?collecting?
QuantitativeQuantitative– You are attempting to measure how You are attempting to measure how
large an effect is.large an effect is. ExperimentalExperimental CorrelationsCorrelations SurveySurvey
– Two types of data collectionTwo types of data collection Self-report of behaviorSelf-report of behavior Actual behaviorActual behavior
Modes of Data Collection:Modes of Data Collection:Self-reportSelf-report
Questionnaires/ Questionnaires/ SurveysSurveys
– Pen and paperPen and paper in personin person mailedmailed
– Internet-based Internet-based surveysurvey
– Random digit Random digit dialingdialing
Modes of Data Collection:Modes of Data Collection:Self-reportSelf-report
InterviewsInterviews– Computer Computer
Assisted Personal Assisted Personal Interviewing Interviewing (CAPI)(CAPI)
– Telephone Telephone interviewsinterviews
– Face to face or Face to face or personal personal interviewsinterviews
– Focus groupsFocus groups
Computer Assisted Computer Assisted Telephone InterviewTelephone Interview
Surveying BehaviorSurveying Behavior
StrengthsStrengths– Specific and quantifiableSpecific and quantifiable– ReplicableReplicable– Larger sample size compared to qual.Larger sample size compared to qual.– GeneralizableGeneralizable
LimitationsLimitations– Self-reportSelf-report
Allows biasesAllows biases
Modes of Data Collection: Modes of Data Collection: Observing BehaviorObserving Behavior
Implicit Association TestImplicit Association Test– Computer programComputer program
Categorize objectsCategorize objects Faster = stronger associationFaster = stronger association
Computers and InternetComputers and Internet– Allow us to observe actual Internet Allow us to observe actual Internet
gambling behaviorgambling behavior– Could also be used for online shopping Could also be used for online shopping
behaviorbehavior– What else?What else?
Example of IATExample of IAT
When will you collect your When will you collect your datadata
LongitudinalLongitudinal– Multiple data pointsMultiple data points
Cross-sectionalCross-sectional– One data pointOne data point
When is it too much?When is it too much?
Only collect data you Only collect data you will usewill use
Time = CompensationTime = Compensation Time = WorkTime = Work Every 10 minutes = 7% Every 10 minutes = 7%
drop in RR (McCarty et al. drop in RR (McCarty et al. 2006)2006)
Response RatesResponse Rates
As the response rate As the response rate Greater likelihood results will Greater likelihood results will
represent the total population (i.e., represent the total population (i.e., generalizability)generalizability)
The higher the response rate, the The higher the response rate, the more accurate the survey.more accurate the survey.
Response RatesResponse Rates
Striving to achieve the highest rate Striving to achieve the highest rate you canyou can
Low rates, might affect your ability to Low rates, might affect your ability to generalize to the population as a generalize to the population as a wholewhole
At least 70% At least 70%
RecapRecap
Each methodology and data Each methodology and data collection method has strengths and collection method has strengths and weaknessesweaknesses
Consider using a combinationConsider using a combination– Focus group to test survey questionsFocus group to test survey questions– Journal to help recall and periodic Journal to help recall and periodic
surveyssurveys– Close ended and open ended questionsClose ended and open ended questions
Recap contRecap cont
Research design, Sampling, and Data Research design, Sampling, and Data collectioncollection– trade off b/w your hypotheses, variables, trade off b/w your hypotheses, variables,
and reality. and reality. – Balance: time; money; and effortBalance: time; money; and effort
Primary SourcesPrimary Sources
Thames Valley University Disseration Thames Valley University Disseration Guide: Guide: http://brent.tvu.ac.uk/dissguide/hm1u3/hm1u3fra.htm
Research Methods in Social RelationsResearch Methods in Social Relations