Notes on Correlational Approach in Research
Introduction
The correlational approach is defined as a non-experimental research method that studies the relationship between two or more variables and assesses the statistical relationship between them with little or no effort to control extraneous variables (APA).
Francis Galton (1888–1911) is regarded as the founder of the idea of correlation; English psychologist and statistician.
In daily life, we observe associations: when one aspect increases, another may increase or decrease, or vice versa.
Examples:
Temperature rise → ice cream sales increase.
Higher occupational stress among employees → performance may decrease.
History and Background
Roots of correlation:
Mathematics (probability): Blaise Pascal and Pierre de Fermat.
Astronomy: astronomers measuring planetary positions; averaging measurements to reduce errors.
Francis Galton (1888): founder of correlation; heredity and height experiments showed tall parents → tall children but not as tall on average; short parents → short children but not as short on average; introduced the idea of regression to the mean and correlation.
Karl Pearson: Galton’s student; sought a unit-free measure between -1 and +1; introduced the correlation coefficient r (Pearson product-moment correlation):
This r is a unit-free measure bounded by -1 and +1.
Charles Spearman (1904): used correlation to study intelligence; developed factor analysis.
1920s–1930s: correlation used in social psychology to study attitudes and traits; Thurstone developed attitude scales based on correlational techniques.
Mid 20th century: correlation central to survey research in studying prejudice, attitudes, and behavior; Richard LaPiere’s experiment explored whether attitudes matched behavior.
Characteristics
Non-Experimental Correlational research: a non-experimental method where investigators do not modify factors; they analyze and examine relationships among variables without changing them.
Backward-Looking Correlational study: looks back at historical information to observe past trends; may identify long-term patterns; links observed may shift in future years.
Dynamic Correlational study: relationships between two factors are not static and can evolve over time due to various causes; a negative relationship previously observed may become positive later.
Advantages
Useful when exploratory manipulation is inappropriate or unethical (e.g., certain studies in humans).
Quick identification of the statistical link between two variables using a suitable methodology.
Generally shorter in time and cheaper than experimental research; advantageous with limited funds or small research teams and a small number of variables.
Allows data collection via brief surveys or multiple approaches, enabling rapid data gathering from a small sample.
Types of Correlation
Direction of Change:
Positive correlation: Variables move in the same direction (both increase or both decrease).
Negative correlation: Variables move in opposite directions (one increases, other decreases).
Zero correlation: No apparent relationship.
Ratio of Variables:
Linear correlation: Constant rate of change.
Non-linear (curvilinear) correlation: Non-constant rate of change.
Number of Variables Involved:
Simple correlation: Between two variables.
Partial correlation: Between two variables, holding others constant.
Multiple correlation: Among three or more variables simultaneously.
Based on the ratio of variables:
Linear correlation: constant rate of change in one variable with respect to the other.
Non-linear (curvilinear) correlation: the relationship is not a constant rate of change.
Based on the number of variables involved:
Simple correlation: between two variables.
Partial correlation: between two variables while holding other variables constant.
Multiple correlation: among three or more variables simultaneously.
Scales Used in Correlational Research
Correlational studies often use psychological scales to measure variables, with emphasis on validity and reliability (psychometrics).
Nominal Scale: categories with no inherent order (e.g., gender, religion, yes/no).
Correlation tests: Chi-square, Phi coefficient.
Ordinal Scale: ordered categories, distances not necessarily equal (e.g., ranks, satisfaction levels).
Correlation tests: Spearman’s rank correlation.
Interval Scale: ordered, equal distances, but no true zero (e.g., Celsius temperature, IQ scores).
Correlation tests: Pearson’s r, regression.
Ratio Scale: ordered, equal distances, and a true zero (e.g., age, income, height).
Correlation tests: Pearson’s r, regression, path analysis.
Examples of Scales in Social/Applied Psychology
Likert Scale: measures attitudes (e.g., 1 = strongly disagree to 5 = strongly agree).
Beck Depression Inventory (BDI): measures depression levels.
Rosenberg Self-Esteem Scale: measures self-esteem.
Perceived Stress Scale (PSS): measures stress levels.
These scales produce quantitative data that can be correlated with each other to understand patterns of relationships (correlation) and, with additional analyses, potential causation considerations.
Correlation and Causation
Correlation indicates the direction and strength of the relationship between two or more variables but does not imply causation.
Causation requires demonstrating a cause-effect relationship with evidence; correlation alone may be due to third variables or alternative explanations.
Example: Correlation between Sense of Purpose and FOMO can exist, but establishing causation would require additional analysis (e.g., controlling for third variables such as screen time, self-esteem) and potentially factor analysis.
Examples
Correlation study: illness representations and coping styles in caregivers for individuals with schizophrenia; result: slightly correlated; method: correlational descriptive.
Motivation and Emotion – The effect of Stress on Health (Holmes & Rahe, 1967): psychosomatic research; results: strong positive correlation between stress and illness.
A study about babies crying and being held (APA dictionary): holding more correlates with less crying; negative correlation.
Applications
Correlational approaches are widely used in psychology, education, health, and social sciences because they allow researchers to explore natural relationships without manipulating variables.
Main applications:
1) Identifying relationships between variables (e.g., relationship between social media use and self-esteem in teenagers).
2) Predicting outcomes (strong correlations can help predict the other variable; e.g., high correlation between college entrance scores and academic performance for admissions predictions).
3) Studying variables that can’t be manipulated ethically or practically (age, IQ, trauma history) – e.g., childhood adversity and adult mental health.
4) Exploring health and clinical patterns (e.g., relationship between stress and blood pressure).
5) Early-stage research (exploratory phase to detect patterns before experiments, e.g., loneliness and depression leading to further experiments).
6) Educational and workplace applications (to improve learning and productivity; e.g., correlation between study habits and exam performance; correlation between job satisfaction and employee turnover).
Limitations
No Causality: correlation does not prove causation; two related factors may be causally linked in either direction or due to a third variable.
Third Variable Problem: hidden factors may explain the relationship (e.g., ice cream sales and drownings both rise in summer due to temperature).
Directionality Problem: cannot determine which variable influences the other.
Over-interpretation Risk: findings may be exaggerated as implying causation when they do not.
Limited Control: researchers observe variables as they occur naturally; confounding factors may distort results (e.g., when studying screen time and academic performance, factors like parental involvement, teaching quality, or socio-economic background may be responsible).
Summary / Conclusion
The correlation approach studies the relationship between variables without manipulating them.
Foundational figures include Sir Francis Galton, Karl Pearson, and Charles Spearman, who advanced its scientific application.
Widely used across psychology, education, health, and social sciences to identify patterns and make predictions.
It has limitations: correlation cannot establish causation and may be influenced by third variables or other factors.
Overall, it is a valuable exploratory tool when used cautiously.