Research Design Flashcards
Mediating Variable
Responsible for the relationship between variables
IV affects the mediator, then mediator affects dv
Moderating Variable
Affects the direction and strength of the relationship between the IV and DV
Quasi-Experimental vs True Experimental research
True experiment = random assignment
Quasi-experiment = pre-existing groups or single group *NO CONTROL of assignment of subjects
Simple Random Sampling
Every member of population has equal chance of being selected (random)
Selection of one member has no effect on selection of another member of the population
Reduces bias
Stratified Random Sampling
Population varies in terms of the characteristics
(E.g. gender, age, education level, SES etc.)
Divide population and randomly sample each statrum
Cluster Sampling
Sample from clusters of individuals when unable to access to entire population of interest
E.g., random selection of patients from 10 inpatient programs
Confounding Variable
Extraneous/irrelevant variable that introduces Systematic error
Threatens internal validity of research design
Ways to Limit the Effects of Confounds
Random assignment
Hold extraneous variable constant
Match subjects on extraneous variable
Include extraneous variable as additional IP (blocking)
Statistical control via ANCOVA
Systematic vs. Random Error
Systematic = confounding variable
Random error = unpredictable fluctuations in subjects/condition/measuring instruments
Internal and External Validity
Research, not psychometrics
Internal Validity = there is a causal relationship between IV and DVs
External Validity = the causal relationship generalizes to other people, settings, times, and operations
Threats To Internal Validity: Extraneous Variables
Maturation (examinee)
History (external events)
Testing / Practice effects
Instrumentation
Regression toward the mean
Selection (assignment)
Attrition
***Threats To External Validity
Testing / Treatment Interaction
Selection / Treatment Interaction
Reactivity
Multiple Treatment Interference (carryover effects)
Reactivity: Evaluation Apprehension
Subject changes their behavior in order to avoid negative evaluation
[Threat To EXTERNAL Validity]
Reactivity: Demand Characteristics
Cues in experimental setting that:
- Inform subjects of the purpose of the study
- Suggest what behaviors are expected
[Threat To EXTERNAL Validity]
Reactivity: Experimenter Expectancies
Experimenter unintentionally provides subjects with cues (demand characteristics)
Experimenter behavior that leads to biased results
[Threat To EXTERNAL Validity]
Ways to Limit Effects of Reactivity
Deception
Single-or Double-blind technique
Unobtrusive measures
Multiple Treatment Interference:
Carryover Effects
When different levels of IV affect one another
Order effects
Practice effects
[Threat To EXTERNAL Validity]
Between-Group Designs
Different levels of IV administered to a different group or subject
Simplest: one IV with two levels and two groups
Between-Groups: Factorial design
2+ IV’s
Allows for analysis of Main effects and Interaction Effects
e.g. two-way ANOVA, three-way ANOVA
Between-Groups: Main Effect
Effect of one IV on the dv
*disregards effects of all other IV’s
Between-Groups: Interaction
Interaction occurs when the effects of an IV differ at different levels of another IV
e. g. effects of intervention levels differ for different sx severity levels
* When there is an interaction, main effects must be interpreted in light of interaction
Within-Subjects Designs (Repeated Measures)
All levels of the IV are administered sequentially to all subjects
Disadvantages:
- carryover effects
- pre/post test autocorrelation confound
Mixed Designs
Combines between-groups and within-subject designs
Requires at least one between-groups IV and
one-within subjects IV
Single-Subject Design
Includes at least one baseline phase and one treatment phase
Subject acts as their own no-treatment control
DV is measured repeatedly at regular intervals throughout baseline and treatment phases
Single-Subject: AB Design
Single baseline phase (A)
Single treatment phase (B)
*multiple dv measurements in each phase
Single-Subject: Reversal Designs (ABA, ABAB…)
Withdrawal of treatment during second and subsequent baseline phases
Advantage: controls for history, maturation and other threats to internal validity
Single-Subject: Multiple Baseline Design
Used when reversal design is is inappropriate for ethical or practical reasons (e.g withdrawing useful tx)
IV sequentially applied to the same subject to create multiple baselines (IV is never withdrawn)
E.g. sequentially measuring different behaviors/ settings / tasks
Interrupted Time Series Design
One Group
DV measured at regular intervals before and after tx
dv NOT measured during tx
Single-Subject Reversal Designs (ABA, ABAB…):
DV baseline levels
DV returns to original baseline level following a treatment phase
= Evidence that IV is having a true effect
Solomon four-group design
Evaluates effects of pretest on internal and external validity
Treats pretest as IV
