Note: The translation for this entry is currently under quality review. Some content is temporarily displayed in English only.
multicollinearity
This term is a technical specialization used almost exclusively in statistics, econometrics, and data science. It describes a specific failure of independence between predictors, which inflates the variance of coefficient estimates and renders p-values unreliable. In a professional research context, it is treated as a diagnostic problem to be solved rather than a characteristic of the data itself.
As an abstract statistical property, the word is uncountable. It refers to a state or phenomenon of a model's structure. You would not refer to "three multicollinearities" in a single model; instead, you would discuss the "degree of multicollinearity" or identify "multicollinear relationships" between specific pairs of variables.
Meanings
A statistical phenomenon in which two or more independent variables in a multiple regression model are highly correlated, making it difficult to determine the individual effect of each variable on the dependent variable.
The researcher had to remove one of the predictors to address the issue of multicollinearity in the model.