What does collinearity mean?
Collinearity is the property of lying on one straight line. In geometry, points are said to be collinear when they all fall along a single line — a simple but fundamental condition used in proofs, coordinate geometry, and computer graphics. The term has taken on an equally important second life in statistics, where collinearity describes predictor variables in a regression model that are so strongly linearly related that they carry overlapping information. When predictors are highly collinear, a model's coefficient estimates become unstable and hard to interpret, which is why analysts routinely run collinearity diagnostics such as variance inflation factors before trusting a regression. The word derives from the Latin root linea, meaning line, combined with the prefix co-, meaning together. Though technical in register, collinearity is essential vocabulary for anyone working with data, geometry, or spatial reasoning, capturing in a single term the idea of things sharing a common straight path.
nounThe condition of lying on the same straight line; in statistics, the situation in which two or more predictor variables in a regression model are highly linearly related.
- Geometry: the property of three or more points lying on a single straight line.
- Statistics: a high degree of linear relationship among predictor variables in a regression model, which can make coefficient estimates unstable.
"The three survey markers showed perfect collinearity, all falling along a single straight line."
"Because the data points exhibited near-perfect collinearity, the regression coefficients could not be estimated reliably."
"A quick determinant test confirmed the collinearity of the four vertices before we attempted the proof."
Rarely used; appears mainly in technical writing when referring to multiple distinct instances or types of collinearity.
"The paper catalogued several collinearities among the sampled landmarks."
In statistics, two variables that are too perfectly aligned can be worse than useless — collinearity means your data may be telling you the same story twice, and your model can't tell which one to believe.
Reviewed by Deb Chak, Editor. AI-assisted content curated by RJS Tech Solutions LLP.
Etymology of collinearity
Collinearity combines the Latin-derived prefix co- ('together') with 'linear', from Latin linea ('line, thread'), which itself likely traces back to linum ('flax'). The adjective 'collinear' entered English mathematical usage by the 19th century, and 'collinearity' developed as its noun form as analytic geometry matured. Its statistical sense arose in the 20th century alongside regression analysis, later spawning the derivative term 'multicollinearity'. Cognates sharing the linea root include 'linear', 'lineage', and 'collineation'.
Related word forms
How collinearity is actually used
Primarily a technical term used in geometry and statistics; rare in everyday conversation. In statistics it is often discussed via its plural-variable extension 'multicollinearity'. Register is formal and academic.
Easily confused with collinearity
Multicollinearity refers specifically to correlation among three or more predictor variables in a regression model, whereas collinearity is the broader term covering any exact or near-exact linear relationship, including between just two variables.
Coplanarity means lying in the same plane, while collinearity is the stricter condition of lying on the same straight line.