Premium
The Use of Regression for Detecting Competition with Multicollinear Data
Author(s) -
Carnes Bruce A.,
Slade Norman A.
Publication year - 1988
Publication title -
ecology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.144
H-Index - 294
eISSN - 1939-9170
pISSN - 0012-9658
DOI - 10.2307/1941282
Subject(s) - collinearity , ordinary least squares , competition (biology) , regression , statistics , mathematics , regression analysis , monte carlo method , stepwise regression , least squares function approximation , econometrics , eigenvalues and eigenvectors , ecology , biology , physics , quantum mechanics , estimator
Monte Carlo simulations were used to demonstrate that regression methods could be successfully used to estimate competition coefficients with collinear data, but only under conditions that may be difficult to meet with ecological data. Ordinary least squares performs well when estimating coefficients associated with noncollinear predictor variables. Stepwise regression and maximum eigenvalue least squares reduce collinearity by deleting information that, even though not statistically significant, may be important in accurately estimating interaction. We propose that apparent competition (Holt 1977) and the failure of regression techniques to detect competition when it is known to exist experimentally may be due to the omission or lack of measurement of critical elements of the community matrix.