The correlational method is a technique used to measure the statistical likelihood of two behaviors relating to each other.
Psychologists are interested in determining whether two behaviors tend to occur together. One way to make this assessment is to use correlations. Generally, two measurements are said to be associated if when the value of one increases, so does the other; this is termed a positive correlation. If, by contrast, one value increases systematically as the other decreases, this is called a negative correlation. Correlational studies are often used in psychology research. Although these studies can suggest a relationship between two variables, finding a correlation does not proves that one variable is causally related to another.
For example, the number of correct answers on a student's test is often positively correlated to the number of hours spent studying. Students who produce more correct answers have generally spent more hours studying. Conversely, fewer correct answers occur with fewer hours spent studying. In this example, correlation and causation are aligned.
A psychologist could also examine whether the number of wrong answers on a test is associated with study time. This pattern is likely to produce a negative correlation: a greater number of wrong answers is associated with less study time. The value of one variable (wrong answers) increases as the other (hours spent studying) decreases.
Correlations allow an assessment of whether two variables are systematically related within a group of people. An individual may show behavior that differs from the group. For example, a given student might study for many hours and still not perform well on a test. This does not mean that study time and test grades are not related; it only means that exceptions exist for individuals.
It is critical to remember that correlational approaches do not allow people to make statements about causation. Thus, greater study time may not necessarily cause higher grades. Students who are interested in a particular subject, such as history, may do better on a test because of their increased interest. They may study more because they like world events, have an intuitive sense of dates, and understand the material. In this situation, academic interest may be more important than study time.
One limitation of the correlational method is that although one variable (study time) may have a causal impact on the other (test scores), researchers cannot know for certain whether a third factor (interest in the material) may be the most critical element associated with both. When a third element is responsible for both variables (increase in study time and increase in grades), psychologists refer to this as the third variable problem.
See also Research methodology ; Scientific method ; Statistics in psychology .
Altman, Douglas G. Statistics with Confidence: Confidence Intervals and Statistical Guidelines. London: BMJ Books, 2011.
Hastie, Trevor, et al. The Elements of Statistical Learning: Data Mining, Inference, and Prediction. New York: Springer, 2009.
Privitera, Gregory J. Statistics for the Behavioral Sciences. Thousand Oaks, CA: SAGE, 2012.
Spiegel, Murray R., et al. Probability and Statistics. New York: McGraw-Hill, 2009.
Cioffi-Revilla, Claudio. “Computational Social Science.” Computational Statistics 2, no. 3 (May/June 2010):259–71.