Validity and Threats
The degree to which results can be generalized to other populations.
What is external validity?
The gold standard of research design.
What is a Randomized Controlled Trial (RCT)?
The hallmark of a true experimental design.
What is randomization?
A study design that approximates experimental design but does not include randomization.
What is a Quasi-Experimental Design?
The model that emphasizes multiple levels of influence, including policy and community levels.
What is the Socioecological Model (SEM)?
The confidence that results are due to the intervention and not something else.
What is internal validity?
This design includes both a pretest and posttest with random assignment.
What is a Pretest–Posttest Randomized Controlled Trial?
A method where participants are randomized within strata (e.g., gender, age group).
What is stratified randomization?
This design uses naturally formed groups and includes both pretest and posttest.
What is a Nonequivalent Groups Pretest–Posttest Design?
Experimental designs are often poorly suited for interventions at this level.
What is the community or policy level?
Variables other than the independent variable that may affect the dependent variable.
What are confounding variables?
This experimental design does not include a pretest.
What is a Posttest-Only Control Group Design?
When exposure to one treatment affects performance in later treatments.
What are carryover effects?
A nonequivalent groups design that does not include baseline measurement.
What is a Nonequivalent Control Group Posttest-Only Design?
A major reason randomization is not used in quasi-experimental designs.
What is ethical or practical constraints?
A threat that occurs when participants naturally change over time (e.g., aging).
What is maturation?
A design used when a characteristic strongly related to the outcome must be balanced before randomization.
What is a Matched-Pairs Design?
A technique used to reduce carryover effects by varying the order of treatments.
What is counterbalancing?
A design that allows the control group to later receive the intervention.
What is a Switching Replication Design?
A quasi-experimental design where all clusters eventually receive the intervention but not randomly.
What is a Nonrandomized Stepped-Wedge Design?
A threat caused when measurement tools or observers change over time.
What is instrumentation?
A design where clusters cross over sequentially from control to intervention randomly.
What is a Randomized Stepped-Wedge Design?
A design where each subject serves as their own control but lacks randomization.
What is a Repeated-Measures Design?
A version of switching replication where the first group continues treatment after the second group begins.
What is a Switching-Replication-Without-Treatment-Removal Design?
The primary difference between experimental and quasi-experimental designs.
What is the presence or absence of randomization?