korreltype

Korreltype is a term used in statistics and data analysis to categorize the qualitative nature of the relationship between two variables. The word combines korrel, meaning correlation, with type, signaling a classification rather than a single numeric measure. It is not a universally standardized term, but appears in regional glossaries, teaching materials, and some software documentation to describe how variables relate in practice.

Typically, korreltype distinguishes between several broad classes: positive correlations, where variables move in the same direction; negative correlations, where they move in opposite directions; and negligible or zero correlation, where no linear association is detected. Many analyses also recognize nonlinear or non-monotonic relationships that a simple correlation coefficient may underestimate. Some usages further differentiate strong versus weak associations or distinguish linear from nonlinear monotonic patterns.

Statistical methods used to infer a korreltype include Pearson correlation for linear relationships, Spearman's rho and Kendall's tau for monotonic associations, and cross-correlation for time-lagged effects. Visual inspection, scatter plots, and domain knowledge are commonly used alongside coefficients to determine the korreltype of a variable pair.

Example: a scatter plot with a clear upward trend shows a positive korreltype; a U-shaped curve may indicate a nonlinear but nonzero korreltype; a cloud with no pattern suggests negligible korreltype.

In practice, identifying the korreltype aids feature selection, model choice, and interpretation. It also warns against overreliance on a single coefficient and highlights the need to consider confounding variables and the possibility of spurious relationships.

See also: correlation, regression, monotonicity, nonlinearity, mutual information.