Коэффициент корреляции

The correlation coefficient r measures the strength and direction of a linear relationship between two variables, with values always between -1 and +1. Values close to -1 or +1 indicate stronger linear relationships, while r = 0 suggests no linear correlation. The sign of r indicates direction: positive values mean variables move together, negative values mean they move oppositely.

A positive correlation coefficient indicates that as one variable increases, the other tends to increase as well, like height and weight in athletes. A negative correlation means that as one variable increases, the other tends to decrease. The sign of the correlation coefficient matches the slope of the best-fit line through the data.

The correlation coefficient r remains unaffected if the x or y values are swapped or if one variable is converted to a different scale. This property makes r a robust measure for comparing linear relationships across datasets measured in different units or orders.

The correlation coefficient is strongly affected by outliers, which can distort the measured strength of the linear relationship. If outliers are known to be errors, removing them improves the accuracy of the r value and provides a more reliable measure of the true linear association between variables.

When r = 1, there is a perfect positive correlation, meaning all data points lie exactly on a straight line with positive slope. When r = -1, there is a perfect negative correlation with all points on a line with negative slope. In real-world datasets, these perfect values rarely occur.

A scatter plot should always accompany the correlation coefficient calculation to visually confirm the linear relationship. Data exhibiting curved or horizontal patterns may have a correlation of zero despite appearing related, so viewing the scatterplot prevents misinterpretation of the numerical r value.

Karl Pearson developed the correlation coefficient in the early 1900s as a numerical measure of linear association strength and direction. It provides a standardized, scale-independent way to quantify relationships between variables, making it essential for comparing correlations across different datasets and research contexts.