6. Correlational Research

Introduction to Correlational Research

• Correlational research investigates the association or relationship between two variables.

• Two variables are measured from the same individual (paired data).

• Correlation does not imply causation (e.g., film crews and global warming).

• Goal: Understand if changes in one variable relate to changes in another, not to prove cause.

Data in Correlational Studies

• Data are presented as paired scores (X and Y) for each subject.

• Represented through tables and scatter plots to visualize relationship patterns.

• Paired data indicate the direction and form of the relationship.

Quantifying Relationships

• Direction: positive (variables move in same direction) or negative (variables move in opposite directions).

• Form: linear (straight line) or monotonic (consistent direction without necessarily being straight).

• Strength: ranges from -1 to +1.

  • Values closer to -1 or +1 indicate a stronger relationship.

  • Values closer to 0 indicate a weak or no relationship.

• Significance: whether the relationship is strong enough to be unlikely due to chance (p-value < alpha).

• Pearson's r coefficient measures the strength and direction for linear relationships between continuous variables.

• Crucially, this coefficient measures association, not cause-and-effect.

Specialized Correlation Methods

• Point-biserial correlation: for relationships between nominal data and continuous data.

• Chi-square test for independence or phi-coefficient: for relationships between two non-numeric (categorical) variables.

• Summary: Non-numeric data require alternative correlational methods.

Applications

• Prediction: using regression models to predict one variable from another.

• Reliability and Validity assessment:

  • Test-retest reliability

  • Concurrent validity

  • Item-total correlation

• Theory Evaluation: assess how variables relate to support or refute theoretical constructs.

• Interpretation: understanding the significance, direction, and strength of observed relationships.

Caveats: Correlation vs. Causation

• Primary Question: "Is this causation?"

• Answer: Correlation alone does not automatically imply causality.

• It's essential to distinguish between association and cause-and-effect, considering the study design.

Strengths

• Excellent for preliminary research to understand variable relationships, especially when variables are difficult or unethical to manipulate (e.g., amphetamine use and brain size).

• Studies variables in their natural state, leading to high external validity.

Weaknesses

• Third-variable problem: An unmeasured variable might influence both measured variables, creating a spurious relationship (e.g., a pre-existing health condition affects both fitness program participation and sick days).

• Directionality problem: It's unclear which variable is causing the change in the other (which is the predictor, which is the effect).

• Cannot establish cause-and-effect relationships.

Summary of Strengths & Weaknesses

• Strengths: Describes relationships, non-intrusive, natural behaviors, high external validity.

• Weaknesses: Cannot assess causality, third-variable problem, directionality problem, low internal validity.

Multiple Regression

• Used to evaluate relationships involving two or more predictor variables and one outcome variable (e.g., how anxiety, IQ, and stress relate to motivation).

• Methodology:

  • Enter: all predictors entered simultaneously.

  • Backward: starts with all predictors, then removes the least significant ones.

• The model describes how a dependent variable (Y) relates to multiple independent variables ($X<em>1, X</em>2, ext{…}$).

• $R^2$ indicates the proportion of variance in Y explained by the model.

• Coefficients describe the average change in Y for a one-unit increase in a predictor, while holding other predictors constant.

Correlational Study Design Example

• Phenomenon: Student academic and personal pressures affect health.

• Research Task: Design a correlational study.

• Possible Design Elements:

  • Variables: Psychological distress (dependent); academic pressure, personal pressure (independent); control variables (social support, study hours, demographics).

  • Measurement: Scales for distress, academic pressure, personal pressure (ordinal/interval).

  • Sampling: Convenience sampling of students (> 100 participants).

  • Procedure: Informed consent, online/paper surveys, confidentiality, anonymous IDs.

  • Analysis: Correlation analysis, potentially multiple regression to assess individual predictor contributions.

  • Ethical Considerations: Informed consent, data confidentiality, participant well-being.