Correlation uses statistical evidence to identify relationships. The two variables can be vastly different from each other. However, correlation ignores those differences and highlights the similar occurrences between them. According to York University, the biometric study of inheritance was greatly influenced by correlation analysis. This way, researchers could associate personality traits that can span a family's generations. The American Medical Association reported in 2006 that when female patients with depression become symptom free, their children are less likely to be diagnosed with depression. Researchers would have correlated the incidence of depression in both mother and child in order to eventually reach that conclusion.
Correlation theory defines the patterned occurrence of any two variables. Although it is mathematical in theory, its practice is universally applicable. Researchers can literally focus on anything from height and weight to the possible relationship of genetically inherited family traits. Its examination of organically occurring patterns relieves correlation practitioners from a stuffy scientific experiment while providing a numerically precise result. Although correlation method does not establish a causal relationship, it does suggest possible avenues of further clinical examination of variables that may not otherwise be tested.
Patterns of Occurrence
Degree of Relationship
Correlation analysis concludes with a mathematical coefficient between +1.00 and -1.00. The "plus" stands for increased occurrence of both variables; the "minus" stands for an increase in one variable with a decrease in the other. Any number between those polar ends represents the degree of relation between the two variables. Correlation produces an objective statistic to represent a relationship. Even a 0.00 correlation indicates no degree of relationship. That's the precision of correlation.
Direction of Relationship
Practitioners of correlation can not only identify a statistical relationship between two variables but also establish an inverse relationship. The increase in one variable can cause a decrease in the other. This is called negative correlation. Zero correlation is when one variable's occurrence has no influence on the second variable. For example, a growing teenager with a "hollow leg" eats like a horse but never gains a pound: a negative correlation between calorie intake and weight gain. Researchers have generally found a positive correlation to suggest that taller people tend to be heavier.
Correlation is statistic-based, therefore it can examine relationships without requiring unethical experiments in order to obtain evidence. The depression example fits. Using correlation, researchers can spare a child the grief of experiments with depression for the purpose of analysis. In this way, correlation is observant. It doesn't seek to isolate variables within scientific methodology. Any experiment which would, by definition, endanger the two patient-variables is clinically unnecessary with correlation.
- Research Methods in Psychology: Section 8.5 - Correlation
- Simply Psychology: Correlation; Saul Mcleod; 2008
- York University: The Correlation Coefficient; Ronald A. Fisher; 1925
- National Institute of Mental Health: Depression Rates Are Lower in Children Whose Mothers Are Successfully Treated; May 2006
- Photo Credit Thinkstock/Comstock/Getty Images