Small-N and Quasi-experimental designs Flashcards
psychophysics
Example: Continuing with the same experiment, suppose that the results reveal an interaction effect between “Amount of Sleep” and “Caffeine Intake.” It means that the influence of sleep on cognitive performance is not the same for individuals with low caffeine intake as it is for those with high caffeine intake.
For instance, it could be the case that:
* With low caffeine intake, there is a significant difference in cognitive performance between 6 hours and 8 hours of sleep (indicating a main effect of “Amount of Sleep”).
* With high caffeine intake, there is no significant difference in cognitive performance between 6 hours and 8 hours of sleep (indicating a main effect of “Amount of Sleep”).
* However, the interaction effect reveals that the difference in cognitive performance between 6 hours and 8 hours of sleep is significantly greater when combined with high caffeine intake compared to low caffeine intake.
behavior-change studies
Behavior change studies are a broad category of research that aim to understand and influence human or animal behavior. These studies often utilize different experimental designs to investigate how various factors can change or influence behavior. Here are some common designs in behavior change studies.
Small-N studies
- psychophysics
- behavior-change studies
- case studies
types of behavior-change studies
- stable-baseline design (AB Design)
- reversal design (ABA or ABAB Design)
- multiple baseline design
stable-baseline design (AB Design)
Stable Baseline Design (AB Design): In this design, researchers observe a subject’s behavior during a baseline period (A) to establish a stable pattern of behavior. Then, they introduce a treatment or intervention (B) and monitor how it affects the behavior. This design helps assess whether the intervention has a significant impact by comparing it to the baseline.
reversal design (ABA / ABAB Design)
Reversal Design (ABA or ABAB Design): This design is used to assess the reversibility of the effects of an intervention. It involves an initial baseline (A), followed by a treatment phase (B), and then a return to baseline conditions (A) or a re-introduction of the treatment (B) to determine if the behavior changes are indeed due to the treatment.
Multiple baseline design
Multiple-Baseline Design: In this design, researchers implement the treatment at different times for multiple subjects, behaviors, or settings. By doing this, they can determine if the treatment consistently leads to behavioral changes across different conditions. It helps establish the treatment’s generalizability and effectiveness.
Quasi-experimental designs
experiments in which the experimenter selects rather than manipulates the independent variable, e.g. environmental variables / subject variables
A quasi-experimental design is a research design that shares similarities with both experimental and observational research but lacks some of the key characteristics of a true experimental design. It is often used in situations where researchers cannot or should not manipulate certain variables for ethical, practical, or other reasons. Quasi-experimental designs are commonly used in social sciences, education, and fields where conducting true experiments is challenging
pretest/posttest design (AB Design)
The pretest/posttest design, often referred to as an “AB design,” is a basic quasi-experimental research design used to evaluate the impact of an intervention or treatment. It involves measuring a group or individual on a dependent variable (DV) both before and after the introduction of the independent variable (IV) or treatment. This design is commonly used in educational and social sciences research to assess the effectiveness of interventions or programs.
nonequivalent control group pretest/posttest design
to have a control group that does not receive the treatment
This design involves comparing an experimental group that receives a treatment or intervention with a control group that does not, but the groups are not formed through random assignment.
interrupted time-series design
Researchers collect data at multiple time points before and after an intervention to assess the impact of the intervention over time.
→ e.g. effect of fluoride in drinking water on school performance
→ e.g. effect of new traffic measurement on number of accidents
→ e.g. effect of anti-smoking campaign on smoking behavior in teens
nonequivalent control group interrupted time-series design
The Nonequivalent Control Group Interrupted Time-Series Design (NCGITS) is a specific type of research design used to evaluate the impact of an intervention or treatment. It is a quasi-experimental design that combines elements of an interrupted time-series design and the use of nonequivalent control groups. This design is particularly useful when researchers cannot use random assignment to create equivalent control and experimental groups.
evaluate quasi-experimental designs
- aim to generate causal claims
- but internal validity is low (just as with small-n studies)
- construct validity, external validity and statistical validity are essential to evaluate the design