Week 7 Flashcards
Quasi Experiment
A research design that approximates an experimental design, but it is not a true experiment because it does not include random assignment.
What are two goals one keeps in mind when designing a Quasi
Figure out comparisons; make them as equivalent as possible, and thank about alternative explanations and rule them out.
History
Another event occurs after the treatment that accounts for the effect.
Maturation
Participants naturally evolve, get bored, tired, etc, between first and second assesment.
Regression to the Mean
The tendency for extreme data points to regress back towards the average.
Testing
Asking people twice (or more), changes their answers..
Attrition/Mortality
When people group out of the study, and differ from those who do not.
Differential Attrition and Differential History
DA: Those in the treatment group are more likely to drop out than the control group.
DH: Something happens to one of your groups and not another.
Threats to internal validity in Pretest-Posttest Non-Equivalent Group Designs.
Control Series Design
Interrupted time series design with control groups.
Mediator
The psychological reason why an effect occurred. IV-Mediator-DV.
Moderator
A variable that changes the magnitude of a known effect when present or absent - tells you when an event will occur.
Factorial Designs
An experiment design with more than one IV.
Factor
An independent variable in a factorial design.
Cell
The conditions in a factorial design. (2x2 = four cells).
Main Effect
Effect of each factor averaging across the other factors.
Interaction
A pattern of means in which the influence of one factor depends on the level of another factor. Similar to a moderator.
Marginal Means
Use to compute main effects - mean of single IV effects - end of column/row.
Simple Main Effect
The effect of one IV at particular levels of another. More specific than a main effect.
Interaction Effect
Differences between differences (simple main effects) - If there is no difference, there is no interaction.
Mixed Factorial Designs
Manipulate one IV using between-subjects design, and another using within-subjects design.
Nonmanipulated Independent Variable Factorial Designs
Measured but not manipulated - usually participant variables (ex: self-esteem).
Factor Analysis
Organize the variables into a smaller number of clusters, such that they are strongly correlated within each cluster, but weakly correlated between.
Correlation Matrix
Correlation (r) between every possible pair of variables in the study.
Multiple Regression
Measuring several IV (all possible causes of DV) resulting in an equation that expresses DV as an additive combination of the IV - shows if an IV makes a contribution to a DV over and above contributions made by another IV.
