The prevailing discourse informing most Canadian training and labor market policy assumes a positive link between individuals' training and their labor market returns in the new knowledge economy. The primary objective of the study is to test the current rhetoric by developing a statistical model of women's job-related training. Training participation is a complex and multi-dimensional social phenomenon. Within the arsenal of existing statistical methods, structural equation modeling is one of the few methods with a capacity to represent complex phenomenon by simultaneously testing cause-and-effect hypotheses. The study uses structural equation modeling to develop, test, and evaluate a model of the determinants and rewards of women's job related training. Empirical studies of job-related training in Canada are few. Until the recent launch of a linked, employer-employee survey, there had been a paucity of data sources that facilitate a national level analysis of job-related training. By using this new data source, the study makes a significant contribution filling the existing gap in our understanding of the determinants of, and returns to women's training in Canada. The study confirms that women are deriving significant economic benefits from their training participation. The conclusion drawn from the national-level patterns in Canada is that training is a crucial element in the reward structure of the labor market for women, as it plays a dual role of being both a reward in itself and a predictor of other labor market rewards. (Contains 2 figures, 1 table, and 5 notes.)