multicollinearity
1
9.9
- noun
- /ˌmʌltiˌkɒləˈnɛərɪti/
- Specialized
- Generally, a VIF exceeding 4 is seen as a sign that problems of multicollinearity exist.
- multicollinearity problem
- multicollinearity in regression
- avoid multicollinearity
Examples
-
In multiple regression analysis, high levels of multicollinearity can obscure the individual contributions of each predictor variable.
-
An examination of the variance inflation factors revealed that multicollinearity did not pose a problem for the analysis.
Academic text (2008) -
The last year of the study, 2008, is excluded in order to avoid perfect multicollinearity.
Academic text (2017) -
Of course, one county has to be excluded to avoid perfect multicollinearity.
Academic text (2017) -
Consequently, it was no surprise to find a similar multicollinearity among the short form variables.
Academic text (2002) -
Therefore, there is no problem with multicollinearity in this regression.
Academic text (2017) -
With the number of variables used, multicollinearity may be a problem that warrants further investigation.
Academic text (2017) -
Diagnostic statistics were first run to test for multicollinearity among various constructs included in this study.
Academic text (2008) -
All of the correlations were below 15%, indicating a low risk for multicollinearity.
Academic text (2012) -
Finally, to assure that there was no multicollinearity, correlations were examined.
Academic text (2015)
Synonyms
1
9.5
Two or more variables are so closely linked that their separate effects cannot be seen
How Collinear
- Specialized
1
9.9
- Specialized
1
9.5
- Specialized
14
4.9
441
3.0
Surface Forms
Morphology
multicollinearity = collinearity (opaque) = multi + co + line + ar + ity
Adding multi- to collinearity is structurally regular, but the resulting statistical term is not readily inferable for a B1 learner.
Etymology
Multicollinearity combines multi- ('many') and collinearity ('on the same line'). So, imagine many predictor variables 'lining up' or overlapping on the same line, and that's why multicollinearity makes it hard to tell each variable's separate effect.