This paper discusses the methodological perspectives of structural change in the study of social mobility. It emphasizes the statistical decomposition of total mobility into structural and non- structural components. Focusing on occupational mobility table, it points out that the conventional statistics is not the best approach to the analysis of social mobility. A Log - Linear Model is introduced to facilitate the decomposition of total mobility. With some specific operations of the Log - LinearModel,a measure ,ofthe formal goodness of fit, G2, is decomposed to test the statistical significances of structural and non - structural components and to compute the relative weights of the two components in total mobility. The approach is comparable to the decomposition of R2 in regression analysis to yield the statistics of F ratio. The paper explores the conclusions of Americanoccupational mobility studies and finds the possibilities of improvement. The paper also provides a critical review of the effect of structural change. It appears that the education structure affects the occupation structure which, in tum, affects the eamings structure. The change in occupational structure is intimately related to the trend in earnings inequality and both of them are the results of the redis - tribution of educational opportunities. A close examination of the Americaneducation, occupation, and earnings distributions indicates that the population heterogeneity has increased and then decreased. I tis concluded that structural change produces a cyclical pattem in social mobility.