✦ScienexySCIENCE KNOWLEDGE LIBRARY
psychology-learning · SCI-PSY-001-033

Correlation and Causation in Psychology

Correlation describes the statistical relationship between variables, whereas causation means that changes in one variable produce changes in another under specified conditions. Distinguishing the two is essential for interpreting psychological research accurately.

Explanation

A fuller explanation is being developed. Verify this concept with cited sources before relying on it.

Example or application

A worked example has not yet been added.

Conceptual Coverage

Definition

Correlation

Correlation is a statistical relationship describing how two variables vary together. It does not by itself establish that one variable causes the other.

Causation

Causation refers to a relationship in which changes in one variable bring about changes in another, subject to appropriate causal assumptions and evidence.

Types Of Correlation

NameDescription

Positive correlation

Higher values of one variable tend to occur with higher values of another.

Negative correlation

Higher values of one variable tend to occur with lower values of another.

Zero or negligible linear correlation

No meaningful linear relationship is detected, although a non-linear relationship may still exist.

Perfect positive correlation

A linear correlation coefficient of +1.

Perfect negative correlation

A linear correlation coefficient of −1.

Causal Criteria

NameDescription

Association

The proposed cause and outcome must show an appropriate relationship.

Temporal precedence

The proposed cause must occur before the outcome.

Non-spuriousness

The relationship should not be adequately explained by uncontrolled alternative factors.

Plausible mechanism

A credible process should explain how the proposed cause influences the outcome.

Common Explanations For Correlation

NameDescription

Direct causation

X influences Y.

Reverse causation

Y influences X rather than X influencing Y.

Third-variable explanation

A separate variable influences both X and Y.

Bidirectional influence

X and Y influence each other over time.

Coincidental association

An observed relationship may arise by chance, especially in small or multiple-testing contexts.

Correlation Vs Causation

Correlation

Describes statistical co-variation.

Causation

Makes a claim about the effect of changing one variable on another.

Key Difference

A correlation can exist without a direct causal relationship.

Scientific Foundation

Historical Development

Correlation became an important statistical tool in the study of individual differences and psychological measurement during the nineteenth century. Later developments in experimental design, causal inference and longitudinal research expanded methods for investigating causal relationships.

Important Contributors

NameContribution

Francis Galton

Contributed to the development of correlation methods in the study of individual differences.

Karl Pearson

Developed the Pearson product-moment correlation coefficient.

Ronald Fisher

Made major contributions to experimental design and statistical inference.

Donald Rubin

Contributed to the potential-outcomes framework for causal inference.

Judea Pearl

Advanced causal diagrams and formal approaches to causal reasoning.

Research Methods

  • Correlational studies

  • Randomized experiments

  • Longitudinal studies

  • Natural experiments

  • Quasi-experimental designs

  • Causal diagrams and statistical adjustment

Limitations

  • Correlation cannot establish causal direction by itself.

  • Unmeasured confounding variables may remain.

  • Statistical adjustment depends on assumptions about the causal structure.

  • Experimental manipulation may be unethical or impractical for some psychological questions.

Deep Understanding

What

Correlation measures how variables are statistically related; causation concerns whether an intervention or change in one variable produces a change in another.

Why

Confusing association with causation can lead to incorrect psychological explanations and ineffective decisions.

How

Researchers examine temporal order, alternative explanations, confounding variables and suitable experimental or quasi-experimental evidence.

When

This distinction is important when interpreting studies about stress, sleep, academic performance, social media use, memory and health-related behaviour.

Example

Suppose screen time and sleep problems are correlated. It is possible that screen time contributes to sleep difficulties, sleep difficulties increase screen use, or a third factor influences both.

Causal Diagram

Direct

X → Y

Reverse

Y → X

Third Variable

Z → X and Z → Y

Description

Different causal structures can produce similar observed correlations.

Mathematical Coverage

Relevance

Correlation and causal inference use statistical mathematics to describe associations, quantify uncertainty and evaluate proposed causal relationships.

Formulas

NameFormulaRangeInterpretation

Pearson correlation coefficient

r = Σ[(xᵢ − x̄)(yᵢ − ȳ)] / √[Σ(xᵢ − x̄)² × Σ(yᵢ − ȳ)²]

−1 ≤ r ≤ +1

Measures the direction and strength of a linear association between two quantitative variables.

Coefficient of determination

R² = r²

In a simple linear regression with an intercept, describes the proportion of sample outcome variance accounted for by the fitted linear model.

Simple regression

Y = β₀ + β₁X + ε

Models the expected outcome as a function of a predictor and an error term.

Worked Example

Problem

A study reports a Pearson correlation of r = 0.70 between weekly study time and examination scores.

Interpretation

The result indicates a strong positive linear association in the sample. Students who study more tend to have higher scores in the observed data.

Important Caution

The correlation does not prove that increasing study time alone causes higher scores. Prior knowledge, study quality, motivation and other factors may influence the relationship.

Calculation

In a simple linear regression with an intercept, R² = 0.70² = 0.49, meaning the fitted linear model accounts for 49% of the sample variance in scores.

Caution

The strength of a correlation does not indicate the importance of a causal mechanism. Pearson's r also measures linear association and can miss non-linear patterns.

Practical Learning

Applications

  • Investigating relationships between sleep and academic performance.

  • Studying stress and concentration.

  • Examining associations between social interaction and well-being.

  • Evaluating relationships between practice time and skill development.

  • Understanding observational psychological research.

Case Study

Title

Social media use and sleep quality

Scenario

A survey finds that people reporting more evening social media use also report poorer sleep quality.

Possible Explanations

  • Evening social media use may interfere with sleep routines.

  • Poor sleep may lead to more late-night device use.

  • Stress or workload may influence both behaviours.

  • Self-report measurement errors may affect the association.

Learning Point

A correlational finding identifies a relationship that requires further investigation; it does not identify the causal explanation by itself.

Exercise

Develop three possible explanations for a positive correlation between examination preparation time and examination performance. Identify what additional research would help distinguish them.

Critical Thinking

Misconceptions

ClaimCorrection

A strong correlation proves causation.

Even a strong correlation can arise through reverse causation, confounding or other explanations.

A zero Pearson correlation means two variables have no relationship.

They may have a non-linear or more complex relationship.

A statistically significant correlation is necessarily important.

Practical importance depends on effect size, context, measurement and uncertainty.

Debates

  • When is observational evidence sufficient to support a causal claim?

  • How should researchers handle unmeasured confounding?

  • What are the ethical boundaries of experimental manipulation in psychology?

Limitations

  • Correlation coefficients can be influenced by outliers.

  • Restricted ranges can weaken observed associations.

  • Measurement error can distort relationships.

  • Causal conclusions depend on assumptions and research design.

Critical Questions

  • Could the direction of influence be reversed?

  • What third variables might explain the association?

  • What research design would provide stronger causal evidence?

Knowledge Assessment

Mcq

QuestionOptionsAnswerExplanation

What does a positive correlation indicate?

  • One variable definitely causes another.

  • Higher values of one variable tend to occur with higher values of the other.

  • Both variables must be identical.

  • There is no statistical relationship.

Higher values of one variable tend to occur with higher values of the other.

A positive correlation describes the direction of an association, not its cause.

Which research approach is generally strongest for testing a causal effect when ethically and practically feasible?

  • An uncontrolled personal anecdote

  • A suitably designed randomized experiment

  • A single opinion survey

  • A simple comparison without controls

A suitably designed randomized experiment

Random assignment can help balance confounding factors and support causal inference when the study is well designed.

Analytical Questions

  • Explain why correlation does not necessarily imply causation.

  • Describe direct causation, reverse causation and third-variable explanations.

  • Discuss the role of random assignment in causal research.

Application Exercises

  • Interpret a hypothetical correlation of r = −0.60 between stress and sleep quality.

  • Create a causal diagram showing how a third variable could influence two correlated psychological variables.

Answer Guidance

Answers should correctly interpret the direction and strength of association and identify why alternative causal explanations must be examined.

Further Learning

Prerequisites

  • Observation, Measurement and Explanation in Psychology

  • Hypotheses and Scientific Predictions

  • Scientific Reasoning and Critical Thinking

Related Concepts

  • Psychological Statistics

  • Experimental Research Methods

  • Individual Differences and Scientific Generalization

  • The Role of Context in Psychological Explanation

Next Learning Path

Study correlation, regression, confounding, experimental design, longitudinal research and causal inference.

Sources and verification

Sources are shown as provided in the project data. Not every source has yet been independently verified. Please report errors or missing citations.

  • Source verification pending; check an authoritative textbook, university, standards body, or peer-reviewed source.

Suggest a correction or source · More in Psychology