Premium
An indirect test of correlation
Author(s) -
Lower W. R.,
Thompson W. A.
Publication year - 1988
Publication title -
environmental toxicology and chemistry
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.1
H-Index - 171
eISSN - 1552-8618
pISSN - 0730-7268
DOI - 10.1002/etc.5620070111
Subject(s) - covariance , partial correlation , covariate , statistics , correlation , mathematics , analysis of covariance , regression analysis , variables , correlation coefficient , variable (mathematics) , regression , mathematical analysis , geometry
Measurements are frequently made of two unpaired dependent variables, Y and Z , at levels of a common independent variable, x . However, the relationship of primary interest may not be between Y and x or Z and x , but between Y and Z . Since Y and Z are not paired it is not possible to directly regress Y on Z or Z on Y , calculate a partial correlation coefficient or perform a covariance analysis with x as a covariate. A method based on normal theory has been developed to indirectly test for correlation from regression analysis of two or more dependent variables when they share a common independent variable. Examples of application of the technique are taken from field studies of the toxic effects of the emissions of a lead smelter on a variety of biological systems.