Study Design
In experimental studies, is the intervention the independent or dependent variable?
The intervention is the independent variable.
What's the difference between probability and non-probability sampling?
Probability sampling involves random selection of participants.
What are the four criteria for causation?
1. Temporality
2. Plausibility
3. Covariance
4. Non-spuriousness
Do risk and protective factor studies involve an intervention?
No, risk and protective factors studies are cross-sectional surveys in which there is no intervention.
A researcher wants to survey individuals who have experienced homelessness and have attempted suicide. What type of sampling strategy should they use?
Purposive sampling
Can a cross-sectional risk and protective factor study establish causation?
No, because data is collected at a single point in time. A researcher must use a pre-test/post-test design or a longitudinal design to establish that the independent variable occurs before the dependent variable.
A researcher surveys 4,305 high school students about risk factors that lead to teen pregnancy. What are the independent and dependent variables?
The risk factors are the independent variables and teen pregnancy is the dependent variable.
A sample of people were selected at random from the CSU faculty directory and sent a questionnaire asking about their lifestyle and health over the past 4 weeks.
What type of sampling strategy is this?
Probability sampling - it's a simple random sample
Which study design is better able to account for confounding variables -- quasi-experimental studies or randomized controlled trials?
Randomized controlled trials because random assignment of participants to a control and intervention group increases internal validity.
A school administrator wants to find out if students' emotional regulation can be enhanced through after-school programming. She surveys 45 students who are involved in after-school programming and 23 who are not. What type of study design is this?
This is a quasi-experimental design because the researcher has an intervention and a comparison group.
At a local high school, 4 participants are selected at random from each class to participate in a research study.
What type of sampling is this?
Stratified random sampling.
A researcher hypothesizes that depression scores will correlate to increased substance use, however, they do not find a relationship between these two variables. What criteria of causality do they fail?
Co-variance. There has to be a relationship between the independent and dependent variable.
A researcher is working with 16 veterans, 8 of whom were randomly selected to receive a service dog and 8 who are still waiting to receive a service dog. They conduct pre-test and post-test surveys with both groups to determine if receiving a service dog reduces depression and anxiety. What type of study design is this?
This is a true experimental design that uses a waitlist control group.
Which type of sampling reduces sampling bias?
Non-probability sampling involves random selection of participants which reduces the potential for bias.
At intake and after termination of services, a clinician evaluates PTSD symptomology with a convenience sample of 13 clients who do 8 weeks of talk therapy and 14 clients who do 8 weeks of cognitive behavioral therapy (CBT). The clinician finds that individuals who are doing CBT have improved outcomes as compared to those who are not. Can the clinician establish that CBT caused the decrease in PTSD symptomology?
No, he cannot establish causation.
1. Temporal order - Yes, because multiple evaluations
2. Plausibility - Yes, the research literature supports CBT as an effective intervention for PTSD.
3. Covariance - Yes, there is a relationship between CBT and decreased PTSD symptomology.
4. Non-spuriousness - No, the researcher used a convenience sample and a comparison group which does not control for confounding variables.