Unformatted text preview: Chapter 11
QuasiExperimental Designs, Single Case Designs, and Developmental Research What is a QuasiExperimental What is a QuasiExperimental Design? This is a study used to investigate an independent variable when a proper control is not possible. QuasiExperimental Designs
QuasiExperimental Designs Program Evaluation tests the effects that certain programs have on groups or individuals. Example the D.A.R.E. program Steps found in Program Evaluation
Steps found in Program Evaluation
needs assessment “Is there a problem that needs to be addressed in a target population?” Example “Do low SES families have more violence within the home?” program theory assessment “Does the proposed program address the needs of the target population?” Example “Will a domestic violence campaign raise the awareness of and the desire to lower the amount of violence in the home?” process evaluation “Is the program reaching the target population? Is the staff administering the program properly providing the service?” Example “Are the low SES families seeing the TV commercials and reading the pamphlets we are handing out? Are the workers actively trying to pass out the pamphlets?” outcome evaluation “Is the goal of the program being achieved?” Example “Is the number of reported cases of domestic violence declining since the program started?” efficiency assessment “Did the benefits of the program outweigh the costs?” Example “Did the money and effort we spent on materials go to good use?” QuasiExperimental Designs assesses the impact of the IVs effect on the DV but has less control than a true experimental design this will lower the confidence that the IV causes changes in the DV. QuasiExperimental Designs
QuasiExperimental Designs Onegroup posttest only design has no control group; very poor ability to determine internal validity. Example p.211. (sitting next to someone on a bench) Onegroup pretestposttest design effect of IV is inferred from the pretestposttest difference in a single group. Example p.211212 (relaxation training and smoking) Problems with onegroup pretestposttest design History effects any event that occurs between the first and second manipulation; confounds the manipulation.
Maturation effects changes that occur systematically over time. Testing when just taking the pretest changes the participant’s behavior.
Instrument Decay characteristics of measuring the behaviors changes as the experiment progresses.
Regression toward the mean initially, participants are selected based on their extreme high or low scores; when they are tested again, their scores have gravitated to the mean. Nonequivalent control group design the control group selected for the experiment is not equivalent to the experimental group. The assignment to groups is not random. Selection differences the differences between the participants in the two groups confound the results. Nonequivalent control group pretest
posttest design just like the design above except that a pretest is given; this improves the ability to see an IVDV relationship. The assignment to groups is not random (which keeps it from being a true experiment).
The two groups are most likely not equivalent Interrupted time series design a series of measurements are made over an extended period of time both before and after the treatments are introduced.
Control series design an interrupted time series design with a control group. Single Case Designs one subject is used and changes in behavior are recorded with the presence of an IV Traditionally used in the research on reinforcement (B. F. Skinner). A “baseline” (a period of time without the IV) is used as a control. Reversal Designs demonstrates the effect of the IV through reversibility (p.204206) also called an ABA design. Multiple Baseline Designs measure baseline several times prior to administering the IV if a change is observed only after the IV is present, then it is likely that the IV caused the change. Replications in single case designs running the single case design on multiple people; if the results are the same, then it is likely that the IV causes changes in the DV. *the following definition will be used for the developmental research designs. Cohort a group of people born at about the same time and who are exposed to the same societal events. Developmental research Developmental research designs Crosssectional method different cohort groups are measured at one point in time. cheap and takes little time
not as descriptive as the other two methods. Longitudinal method one cohort group is measured as they grow older. very descriptive
expensive and takes a very long time Sequential method starts with cross
sectional method and then participants are tested at least one more time. combines the features of both cross
sectional and longitudinal methods. ...
View Full Document
- Spring '11
- nonequivalent control group, low SES families, QuasiExperimental Designs