Experimental Research Take some action and observe its effects Take some action and observe its effects Extension of natural science to social science.

ExperimentalExperimentalResearchResearch

Experimental ResearchExperimental Research

Take some action and observe its effectsTake some action and observe its effects Extension of natural science to social scienceExtension of natural science to social science Best for limited and well defined conceptsBest for limited and well defined concepts Useful for hypothesis testing - need theoryUseful for hypothesis testing - need theory Focus on determining causation, not just Focus on determining causation, not just

description description

Components of Components of ExperimentExperiment

Three components:Three components: Independent and dependent variablesIndependent and dependent variables

Effects of stimulus on some outcome variableEffects of stimulus on some outcome variable Pretesting and posttestingPretesting and posttesting

Ability to assess change before and after Ability to assess change before and after manipulationmanipulation

Experimental and control groupsExperimental and control groups Comparison group that does not get stimulusComparison group that does not get stimulus

Experimental and Control Experimental and Control GroupsGroups

Must be as similar as possible.Must be as similar as possible.

Control group represents what the Control group represents what the experimental group would have been like experimental group would have been like had it not been exposed to the had it not been exposed to the experimental stimulus.experimental stimulus.

Selecting SubjectsSelecting Subjects

Probability samplingProbability sampling

RandomizationRandomization Most statistics used to analyze results assume Most statistics used to analyze results assume

randomization of subjects.randomization of subjects. Randomization only makes sense if you have a Randomization only makes sense if you have a

reasonably large pool of subjects.reasonably large pool of subjects.

Pre-Experimental Pre-Experimental DesignsDesigns

On-Shot Case StudyOn-Shot Case Study

One Group Pretest-One Group Pretest-

Posttest DesignPosttest Design

Static Group Static Group

ComparisonComparison

True Experimental True Experimental DesignDesign

Solomon Four-Group Solomon Four-Group DesignDesign

Classic Design may sensitize subjectsClassic Design may sensitize subjects

More complex experimental designsMore complex experimental designs

Posttest-only Control Group Posttest-only Control Group DesignDesign

Includes Groups 3 and 4 of the Solomon Includes Groups 3 and 4 of the Solomon design.design.

With proper randomization, only these With proper randomization, only these groups are needed to control the problems groups are needed to control the problems of internal invalidity and the interaction of internal invalidity and the interaction between testing and stimulus.between testing and stimulus.

Other Design Other Design ConsiderationsConsiderations

Double blind - no experimenter biasDouble blind - no experimenter bias

Subject selection - convenience or Subject selection - convenience or representativerepresentative Generalizability vs. explanatory powerGeneralizability vs. explanatory power Comparability of experimental and control Comparability of experimental and control

groupsgroups Probability sampling for representativenessProbability sampling for representativeness Randomization over matching for Randomization over matching for

equivalenceequivalence

Threats to Validity in Threats to Validity in ExperimentsExperiments

History - intervening eventHistory - intervening event

Maturation - people changeMaturation - people change

Testing - respond to measuresTesting - respond to measures

Instrumentation - change measuresInstrumentation - change measures

Regression - Regress to meanRegression - Regress to mean

Selection biases - incomparable groupsSelection biases - incomparable groups

Experimental mortality - Drop out of studyExperimental mortality - Drop out of study

Diffusion of treatment - contamination of controlDiffusion of treatment - contamination of control

Compensation - advantage control groupCompensation - advantage control group

Compensatory rivalry - control group competes harderCompensatory rivalry - control group competes harder

Demoralization - control group may give upDemoralization - control group may give up

+ External Threats to Validity / Interactions with Stimulus+ External Threats to Validity / Interactions with Stimulus

Quasi-Experimental Design

"Natural" Experiments"Natural" Experiments

Important social scientific experiments Important social scientific experiments occur outside controlled settings and in the occur outside controlled settings and in the course of normal social events. course of normal social events.

Raise validity issues because researcher Raise validity issues because researcher must take things as they occur.must take things as they occur.

Time and Survey Time and Survey DesignDesign

Extending logic of Experimentation to Extending logic of Experimentation to SurveysSurveys Static designs:Static designs:

Cross-sectional studyCross-sectional study Longitudinal designs:Longitudinal designs:

Trend studiesTrend studies Cohort studiesCohort studies Panel studies Panel studies

Experimental MethodExperimental Method

Strengths:Strengths:

Isolation of the experimental variable over Isolation of the experimental variable over time.time.

Experiments can be replicated several Experiments can be replicated several times using different groups of subjects.times using different groups of subjects.

Weaknesses:Weaknesses:

Artificiality of laboratory setting.Artificiality of laboratory setting.

Social processes that occur in a lab might Social processes that occur in a lab might not occur in a more natural social setting.not occur in a more natural social setting.