Correlational Research Strategy Notes

Definition of Correlational Research: A research strategy that examines the relationship between two variables without manipulating them.
Goal: To establish whether there is a relationship between two variables (e.g. X and Y).
Key Phrase: "Correlation does not equal causation" - important to remember as correlation can indicate association but not causation.

Scatter Plots: Graphical representation of data points with X and Y coordinates to visualize relationships between variables.
Positive Correlation: Both variables increase together (e.g. knowledge of correlation and happiness).
Negative Correlation: One variable increases while the other decreases (e.g. more hours worked may correlate with lower grades).

Correlational Research: Observes natural relationships without manipulation.
Experimental Research: Assigns participants to different conditions (groups) and manipulates variables to test effects (e.g. drug vs placebo).
Differential Research: Compares groups without assigning participants to them (e.g. examining average income across different suburbs).

Definition: A statistic that quantifies the strength, direction, and consistency of a correlation.
Direction:
- Positive: As X increases, Y increases.
- Negative: As X increases, Y decreases.
Form:
- Linear Relationship: A straight-line pattern, measured by Pearson's correlation coefficient (R).
- Non-linear Relationship: Could be monotonic where the correlation is maintained, but not at a constant rate.
Strength of Correlation: Determined by the absolute value of R (closer to 1 indicates a stronger correlation).

Coefficient of Determination (R²): Indicates the proportion of variability in one variable that can be explained by the other.
Statistical Significance (P-value): Assesses whether the correlation is likely due to chance; important when evaluating small sample sizes.

Point-Biserial Correlation: Used when one variable is numeric and the other is binary (e.g. suburb vs income).
Chi-Square Test: Evaluates categorical data in contingency tables to see if observed frequencies differ from expected frequencies.

Predictive Modelling: Utilizes knowledge of variable relationships for forecasting outcomes (e.g. temperature vs bushfire severity).
Meta-Research: Validity and reliability assessments in research measures.
- High correlation with established tests indicates validity.
- Test-retest reliability assesses consistency over time.

Non-Intrusive: Reflects real-world data without manipulation.
High External Validity: Good at capturing behaviors as they naturally occur.

Lack of Causality: Cannot determine if one variable influences the other.
Third Variable Problem: A third variable may influence both variables (e.g. temperature influencing both ice cream sales and crime rates).
Directionality Problem: Unclear whether X influences Y, Y influences X, or if they are both influenced by another variable.

Used for examining relationships while controlling for other variables to understand unique contributions to the outcome variable.

Encourage thinking of personal correlations in daily life.
Visit the 'spurious correlations' website for illustrative examples and the understanding of the correlation vs causation concept.

Correlation is a significant research method that serves practical applications in various fields. Understanding its strengths and limitations is crucial for effective analysis and application.