Comprehensive Study Notes: Correlation, Causation, Experimental Design, and Research Ethics

Variable: anything that can vary and is measurable.
Scatterplot: a graph where each dot represents the values of two variables.
- Slope indicates the direction of the relationship.
- Amount of scatter indicates the strength of the correlation.
Correlation types:
- Positive correlation: two sets of scores tend to rise or fall together.
- Negative correlation: two sets of scores relate inversely (one goes up as the other goes down).
- Correlation coefficient (r): measures the strength and direction.
- Perfect positive correlation: r = +1.000
- No relationship: r = 0.000
- Perfect negative correlation: r = -1.000
Statistics reveal relationships that might be missed with casual observation.

Illusory correlation: perceiving a relationship where none exists or perceiving a stronger-than-actual relationship.
Regression toward the mean: extreme scores tend to move toward the average on retesting; extraordinary happenings are followed by more ordinary ones.
Important point: Correlational data do not establish causation.

Core idea: Correlation does not prove causation.
Possible interpretations of a correlation between X and Y:
1. X causes Y.
2. Y causes X.
3. A third variable Z causes both X and Y.
4. A bidirectional relationship.
Always consider alternative explanations.

Purpose: To establish cause and effect by manipulating factors.
Experimental group: receives the treatment (version of independent variable).
Control group: does not receive the treatment; serves as comparison.
Random assignment: assigns participants to conditions by chance to equalize groups and minimize preexisting differences.
Independent variable (IV): the factor that is manipulated (the cause).
Dependent variable (DV): the outcome measured; may change due to the IV (the effect).
Confounding variables: other factors that might influence the DV; controlled by random assignment.
Operational definitions: precise procedures for manipulating the IV and measuring the DV, crucial for replication.

Placebo effect: experimental results caused by expectations rather than the actual treatment.
Double-blind procedure: neither participants nor researchers know who receives treatment or placebo; controls for experimenter bias and placebo effects.

Descriptive (naturalistic) methods:
- Purpose: Observe and record behavior.
- Weaknesses: No manipulation; limited causation inference.
Correlational methods:
- Purpose: Detect naturally occurring relationships; assess prediction.
- Weaknesses: Cannot specify cause and effect.
Experimental methods:
- Purpose: Explore cause and effect.
- Weaknesses: Sometimes not feasible or ethical; results may not generalize.

Laboratory conditions simplify reality to isolate theoretical principles, not to recreate exact everyday behavior.
Generalization: Principles derived in the lab often apply to real-life situations.

Why study animals? To understand learning, neural mechanisms, and contribute to human disease treatments; shared biological processes among species.
Human participant protections:
- Informed consent: participants must agree after being told enough about the study.
- Protection from harm: minimize risks.
- Confidentiality: protect participants’ information.
- Debriefing: explain the study afterward.
Institutional Review Boards (IRBs): review proposals to safeguard participants.
Accountability and integrity: honesty, replication, and avoidance of fraud are core scientific values.

Correlation identifies relationships but doesn't establish cause and effect.
Experiments (with IV manipulation, random assignment, confound control) are necessary to demonstrate causation.
Good science requires rigorous methods, transparent reporting, replication, and ethical safeguards.