Experimental and Quasi-Experimental Designs Flashcards
Between-subject Design
-Participants assigned to different levels of IV
-Comparisons made between different groups of participants
Within Subjects Designs
-Participants assigned to all levels of IV
-Comparisons made within the same group of participants
Between Subjects: Post-test only
-Two (or more) groups that should not be systematically different
-Random Assignment
-Introduce Independent variable
-Assess the dependent variable
Pros:
-No influence of previous measure
Cons:
-Cannot check group equivalence
-Not necessary if participants are randomly assigned
Between Subjects: Pre-test Post-test
-A pretest is given to each group prior to the introduction of the experimental manipulation
-Purpose: Are the groups actually equivalent at the beginning of the experiment
Pros:
-Controls for individual differences/groups equivalence
Cons:
-Time consuming
-Attrition more likely
-Awareness of hypotheses
Within Subjects: Concurrent Measures Design
Present the different conditions simultaneously
Within Subjects: Repeated Measures Design
-Assessed on the DV after every level of the IV
-Ex: Show 3 scary movie clips to 100 people
-Ask for each one “how afraid are you right now”
Issues with Within Subjects designs
-Order effects:
-What if the order of alone v. confederate makes a difference?
-What if the order we show the scary movies in makes a difference?
-Order is confounded with the IV
Types:
-Carryover effect- Effects carrying over from one level to another
-Practice effect
-Fatigue Effect
Solutions to order effects:
-Full counterbalancing-Using all possible combinations (with 3 conditions
How to choose between subjects or within subjects?
- Few Participants = Within
- Will participants act differently if they know about the study = Between
- Will the IV produce a relatively permanent change = Between
-Depends on what you are studying
Quasi Experimental Designs
-Used when features of an experimental design cannot be achieved
-Missing element is almost always random assignment
- Have lower internal validity than experimental, but you can improve it
Quasi: One-group Posttest-Only design
Quasi: One-group Pretest-posttest design
-Obtains a comparison by assessing participants before and after IV
-Better but still many issues:
-History
-Maturation
Quasi: Nonequivalent Control Group Design
-Compares groups:
-Common groups are a treatment group with comparison group
-Still no random assignment:
-Consequence is that differences may be systematic and selection bias can be caused
Quasi: Nonequivalent control group pretest-posttest design
-One group pretest-posttest design + nonequivalent control group design
-Compares groups and incorporates a pretest and posttest
-Groups might be a treatment group with non-equivalent control group
-Still no random assignment
Maturation
Natural changes that occur in participants over time, independent of any intervention
Creating groups
-In some cases, you can manipulate the IV
-In other cases, you can use subject variables like ethnicity
Quasi: Interrupted time series design
-Examines the dependent variable over an extended period of time, before and after the independent variable is implemented
Ex: A state stops the death penalty
-Did murders go up?
-Not the same people over time
Issues: History effects
Quasi: Control Series Design
-Interrupted time series + control/comparison group
-Find a similar population that did not receive the IV being studied
