Non-random Exogenous Variables in Path Analysis

Path analysis requires that exogenous and residual variables not be correlated. An alternative to assuming the two to be independent is to treat the exogenous variables as fixed or non-random. Examination of path models, however, reveals that the usual standardized path coefficients are biased and inconsistent when exogenous variables are considered fixed. At the same time, the path coefficient divided by the residual term is a large-sample unbiased estimate of the true coefficient. Variants of this basic transformation are discussed and an empirical example is included

Econometrics · Path (computing · Path analysis (statistics · Statistics · Advanced Statistical Modeling Techniques · Computer Science · Mathematics · Optimal Experimental Design Methods · Psychology

• On the evaluation of causal models

  Open Access•David A Specht•Social Science Research•1975

  Añadir item

• Non-Random Exogenous Variables in Path Analysis

  Lawrence S Mayer•American Sociological Review•1974

  Añadir item

• Non-Random Exogenous Variables in Path Analysis

  William T Bielby, James R Kluegel•American Sociological Review•1974

  Añadir item

• Priority Position of Communication in the African Development Process

  David O Edeani•African Studies Review•1980

  Añadir item

• The Method of Path Coefficients

  Open Access•Sewall Wright•The Annals of Mathematical…•1934

  Añadir item

• Problems in Path Analysis and Causal Inference

  D R Heise•Sociological Methodology•1969

  Añadir item

• Principles of Path Analysis

  K C Land•Sociological Methodology•1969

  Añadir item

• Causal Inferences, Closed Populations, and Measures of Association

  Open Access•Hubert M Blalock•American Political Science Review•1967

  Añadir item

• Path Analysis

  Otis Dudley Duncan•American Journal of Sociology•1966

  Añadir item