Imagen de Portada
Código QR

CORRELATED INDEPENDENT VARIABLES: THE PROBLEM OF MULTICOLLINEARITY.

Whenever two supposedly independent variables are highly correlated, it will be difficult to assess their relative importance in determining some dependent variable. The higher the correlation between independent variables the greater the sampling error of the partials. Numerical examples are given...

Descripción completa

Publicado en:Social Forces Vol. 42; no. 2; pp. 233 - 238
Autores principales: Blalock, H. M., Jr.
Formato: Artículo
Publicado: Oxford University Press / USA Dec63
Materias:
Multicollinearity
Mathematical variables
Statistical correlation
Sociological research
Mathematical statistics
Mathematics
Acceso en línea:Ver este registro en EBSCOhost
Sumario:Whenever two supposedly independent variables are highly correlated, it will be difficult to assess their relative importance in determining some dependent variable. The higher the correlation between independent variables the greater the sampling error of the partials. Numerical examples are given to illustrate the practical implications for social research.