Premium
A General Approach to Effect Decomposition
Author(s) -
Hou Feng
Publication year - 2014
Publication title -
social science quarterly
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.482
H-Index - 90
eISSN - 1540-6237
pISSN - 0038-4941
DOI - 10.1111/ssqu.12074
Subject(s) - covariate , mediation , decomposition , path analysis (statistics) , outcome (game theory) , econometrics , simple (philosophy) , psychology , regression analysis , computer science , mathematics , statistics , sociology , social science , epistemology , mathematical economics , ecology , philosophy , biology
Objective This article illustrates the commonality among the Oaxaca decomposition, mediation analysis, and path analysis used in various social science fields to decompose the effect of a predictor on the outcome into constituent components. Methods A general approach is proposed that extends the Oaxaca decomposition to continuous predictors. It also removes one critical restriction on covariates in mediation analysis. Results An empirical example shows that the effect of fathers’ education on children's years of schooling primarily works through children's early skill development and educational aspirations. Conclusion The proposed approach is easy to implement since it requires only two pieces of information: simple correlations and standardized regression coefficients.