Multicollinearity problems and ridge regression in sociological models |
| |
Authors: | Robert Mason William G Brown |
| |
Institution: | Oregon State University, Corvallis, Orgeon, USA |
| |
Abstract: | The consequences of omitting relevant variables in regression analysis in order to cope with even moderate levels of multicollinearity can produce severe difficulties in subsequent theoretical interpretation of social science data. Use of ridge regression, a form of biased linear estimation, is explored. Application of ridge regression to a substantive sociological model is given to illustrate ridge regression's strengths as well as its limitations. |
| |
Keywords: | |
本文献已被 ScienceDirect 等数据库收录! |
|