Lecture 2 Research methods Flashcards
What is the scientific method?
A formal, systematic approach to identify reliable and valid explanations for phenomena.
Evidence-based medicine relies upon the scientific method.
What is the relationship between evidence-based medicine and the scientific method?
Evidence-based medicine integrates best research evidence with clinical expertise and patient values.
What is the first step in designing a study?
Formulating a research question.
Example research question regarding coffee and health?
Do people who drink coffee tend to have better health?
What are the main types of study designs?
- Experimental study
- Observational study
- Case study
What characterizes an experimental study?
Random assignment and manipulation of independent variable in a highly controlled setting.
Advantages of experimental studies?
- Can allow for stronger causal inferences.
Disadvantages of experimental studies?
- Random assignment not always possible
- Some things can be difficult to manipulate
- Manipulations are not always reliable/valid
- Not always ecologically valid
What is an observational study?
Observing variables in the real world and quantifying their associations.
Advantages of observational studies?
- Typically large samples
- Ecologically valid
- Measurement tends to be strong
Disadvantages of observational studies?
Difficult to draw strong causal inferences (i.e., correlation does not equal causation).
What are cross-sectional designs?
People are assessed at one point in time.
What is a longitudinal design?
People are assessed over time.
What are quasi-experimental designs?
Leveraging naturally-occurring differences between groups without random assignment.
What is a case study?
Closely studying one person to find out as much as possible.
Advantages of case studies?
- Great depth
- Can inform our understanding of psychological phenomena
- Sometimes necessary
Disadvantages of case studies?
Unknown generalizability.
What must be met to infer causality?
- Evidence must be consistent
- Cause must appear before outcome
- Dose-response relationship must exist
- Strength of association must suggest causality
- Incidence/prevalence must drop when causal factor is removed
What are types of health measurements?
- Genes
- Biomarkers (e.g., cortisol)
- Neuroimaging
- Self-report
- Behavioral/life outcomes
What does reliability refer to in health measures?
Consistency of a measure.
What does validity refer to in health measures?
Accuracy of a measure.
Can an unreliable measure be valid?
No.
What is the null hypothesis in significance testing?
The effect is not true.
What is the alternative hypothesis?
The effect is true.
