Overview
This lecture covers the basics of correlation and experimentation, explaining the differences between the two, key statistical concepts, and common research biases to be aware of in scientific studies.
Correlation
- A correlation measures how two variables are related or associated with each other.
- Correlation does not imply causation; correlational studies cannot determine cause and effect.
- Variables are anything measurable in a study (e.g., smoking, pregnancy).
- Correlational studies are used when experiments would be unethical or impractical.
- Data from correlational studies is often displayed on a scatter plot, with each dot as a data point.
- A positive correlation means as one variable increases, so does the other.
- A negative correlation means as one variable increases, the other decreases.
- No correlation shows no consistent relationship between variables.
- The correlation coefficient (+1: strongest positive, 0: no correlation, –1: strongest negative) quantifies relationship strength.
- The strongest correlation is the number closest to 1 or –1, regardless of sign.
- Common misconceptions arise from confirmation bias (believing patterns that aren’t real).
- Regression toward the mean: extreme results are likely to return to average on subsequent trials.
- Third variable problem: an unmeasured variable may actually cause the observed correlation.
- Directionality problem: unclear which variable influences the other.
Experimentation
- Experiments manipulate one or more independent variables to observe effects on dependent variables.
- Only experiments can establish cause-and-effect relationships.
- The independent variable is manipulated; the dependent variable is measured for change.
- Random assignment places participants into control or experimental groups to reduce confounding variables.
- Experimental group receives the independent variable; control group does not.
- Placebo (inactive treatment) is used in control groups to prevent participant bias.
- Single blind: participants don't know group assignment; double blind: both participants and researchers are unaware.
- Placebo effect: improvements due to belief in treatment, not the treatment itself.
- Nocebo effect: experiencing negative side effects just because they are expected.
- Reliable research yields consistent results; valid research accurately measures what it intends to.
Key Terms & Definitions
- Correlation — measure of relationship between two variables.
- Variable — anything measurable in a study.
- Scatter plot — graph showing data points for two variables.
- Correlation coefficient — numerical index of relationship strength (from –1 to +1).
- Confirmation bias — tendency to seek out or interpret information in a way that confirms one's beliefs.
- Regression toward the mean — tendency for extreme results to revert to average.
- Third variable problem — unmeasured variable causing the observed correlation.
- Directionality problem — uncertainty about which variable affects the other.
- Independent variable — variable manipulated by the experimenter.
- Dependent variable — variable measured for effect.
- Random assignment — randomly placing participants into groups.
- Experimental group — group receiving the independent variable.
- Control group — group not receiving the independent variable.
- Placebo — inactive treatment to measure psychological effects.
- Single blind — participant unaware of group assignment.
- Double blind — both participant and researcher unaware of group assignment.
- Placebo effect — observed improvement due to belief in treatment.
- Nocebo effect — negative effects experienced due to expectations.
- Reliability — consistency of a measure.
- Validity — accuracy of a measure.
Action Items / Next Steps
- Write down the key differences between correlation and experimentation.
- Practice identifying independent and dependent variables from study examples.
- Review scatter plots to recognize positive, negative, and no correlation.
- Complete assigned textbook reading examples on common research misconceptions.