Not- Quite Experimental Designs
Classic Experimental Design
—— participant are randomly assigned to IV conditions
——IV is manipulated
——DV is measured
What if you can’t randomly assign?
—–Signle-Case Designs
——Quasi-experimental designs
——–Developmental designs
Single- Case Experimental Designs
- A.K.A single subject designs or small N designs
- Use this when trying to figure out whether an experimental manipulation had an effect on single participant
- basically, you record the baseline control) behavior, than record behavior during the treatment period, to see if it changed
- problem is there are many alternative explanations for why the behavior chaged
Single-Case Experimental Designs: Reversal Designs
- record the baseline behavior, then record behavior during the treatment period, then remove to treatment to record a second baseline period
- called ABA design, or withdrawal design
- to rule out possibility that some unrelated event caused the change in behavior, you can do multiple reversals ABA design or ABA BAB design)
Single- Case Experiments: Multiple Baseline Designs
- sometimes its unethical or impossible to reverse El treatment
0n this case, you can take multiple baseline measures at diff times, and
multiple post-treatment measures at different times- The multiple baselines can be across subjects, introducing the IV manipulation at different points in time
- across=behaviors
- across situations
- interrupted timer series design
- control series design
—– try to find a comparison group, to act like a control group
Developmental Research Designs
Cross- Sectional Method
- study people of different areas of a single point in time
— Longitudinal Method
—– study the same group of people at different points in time as they grow older
COMPARISON of LONGITINAL And CROSS-SECTIONAL METHODS
- advantages of cross- sectional designs:
—— much faster and cheaper - disadvantages of cross–sectional designs
—— differences may be due to the cohort effects, not age-related change - advantages of longitudinal designs
——– can actually see how a variable changes in each person, and the group, as they age - disadvantages of longitudinal designs
——— expensivee difficult and takes many years
———- attraction can impact results
SEQUENTIAL METHOD
- do a cross- sectional study and longitudinal study at the same time
Single Case Experiments: Replications
replication= do the same single -subject experiment with additional subjects
- enhances generability, and helps rule out individual differences as casual factors
QUASI- EXPERIMENTAL DESIGNS
- you use these when you want to study the effect of an independent variable, but you aren’t able to exercise experimental control over it
— limits ability to make definitive statement about cause and effect
Quasi experimental design- One Group Post Test only design
- you have nothing to compare it too
Quasi-experimental - Nonequivalent control group pretest-post test design
-concern is that there’s section differences between groups
Quasi Experimental Design- One Group pretest post test design
- a change in one group pretest–post test design can’t rule out alternative explanations
Quasi- experimental Design; Nonequivalent control group design
- does have control group
- participants not randomly assigned
- concern is that there’s selection differences between groups
Quasi experimental design- -nonequivalent control group pretest- post test design
—- can use pretest- to see if groups were the same before start of experiment
—- can also look at look at change in scores from pretest to post test
Quasi experimental Design- Propensity score matching of nonequivalent treatment and control groups;
—- similar to matched pairs Design; measure all participants health scores match each person in experimental group with someone in control group with similar score
- Propensity score matching is statistically matching the groups on multiple variables instead of just one
State the key difference between true experimental designs and Quasi experimental designs, How does this affect how we can interpret the results?
In a true experimental random assignment is used to ensure equality within groups
Affects how we interpret results because, with quasi-experiments, confounding variables could be our causes
Why do we use single-case experiments
we use them to determine whether an experimental manipulation had an effect on a single research participant
Describe the steps involved in an ABAB design, as well as what each step is for
ABAB
A(baseline) B (treatment) A (baseline) B (treatment)
What is meant by a multiple-baseline design across subjects, or across situations? Provide Examples
Multiple baseline design
- Across Subjects: the behavior of several subjects is measured over time; for each subject, manipulation is introduced at different points in time
- Across situations: same behavior is measured in different settings, differerent times in each setting
Multiple Baselines Design across subjects
- Baselines are measured over time for each subject and manipulation is introduced at a different point in time (which rules out explanations based on chance and historical events ECT.)
Multiple baseline across behaviors
Several different behaviors are measured over time, and manipulation is applied to each of those behaviors to see if it affects those behaviors
Multiple Baselines across situations
Behavior is measured in diff settings. Manipulation introduced at different times in each setting and expectation is that change in behavior will occur only after manipulation
What is a reversal design? Describe an example and explain what the purpose is of doing it this way?
A single-case design in which the treatment is introduced after a baseline period and then with drawn during a second baseline period. It is to show the reversibility of the manipulation
Give an example of a one-group PostTest-only design, and explainw hy these results cannot be interpreted.
Participant->smoking measure->training program-> Smoking measure DV post test
Cannot be interpreted because there can be confusing variables
