Decomposition of Differences

Abstract

This paper examines methods of decomposing a difference in levels between groups for a dependent variable such as income. Applied to regression equations, this technique estimates the contribution to the difference from divergent characteristics and divergent rates of converting characteristics into the dependent variable. The consequences of an "interaction" component being present in the decomposition is examined. The paper, using data from the 1960 Census, shows how ignoring the interaction term can influence results. n n comparative social research investigators frequently

The Work-Family Role System and Physician Productivity

“…Finally, the sex difference in mean hours worked for all married respondents was decomposed according to techniques described by Winsborough and Dickinson (1971), Althauser and Wigler (1972), Iams and Thornton (1975), and Wolf and Fligstein (1979). This procedure illustrates the extent to which the sex difference in hours is attributable to the following components: (1) the different composition (or distribution) of the male and female samples with respect to the independent variables; (2) the different effects of the independent variables on men and on women; (3) the interaction between composition and effects; and (4) sex differences not attributable to these other components (that is, the difference in intercepts).…”

Lasting Consequences of the Summer Learning Gap

“…The second and third estimates are a form of statistical experimentation, in the nature of regression decomposition as applied to least squares regression analyses (e.g., Althauser and Wigler 1972;Iams and Thornton 1975;Jones and Kelley 1984). The first is derived by evaluating the logistic equation for the focal group at that group's mean regressor values-high SES youth in the case of college track placement and four-year college enrollment; low SES youth in the case of college track placement and permanent dropout.…”

Wealth Inequality: Ethnic Disparities in Israeli Society

“…Because we are mainly interested in the role played by labor market earnings and intergenerational transfers in producing wealth disparities between ethnic groups, we adopt a procedure recommended by Oaxaca (1973) Iams and Thornton (1975) and Jones and Kelly (1982) that enables us to examine three major components. The first component pertains to differential returns (differences in coefficients); the second component pertains to differences in mean characteristics; and the third component captures an interaction between returns and characteristics across groups.…”