Much of the research on the demography of social stratification has involved highly detailed descriptive analysis. Although much has been learned from this work in the past few decades, we believe that, at this stage of its development, the field would benefit from increasing the scope and breadth of its analytical models. An important task in this regard is to integrate more varied elements of demography and stratification in order to develop richer (albeit increasingly more complex) models of the processes that generate patterns of mobility and inequality. Although valuable insights and results have been obtained from the usual approach, which predicts a dependent variable outcome as a function of a set of exogenous variables, the field would currently benefit from additional work that attempts to incorporate additional factors that may sometimes be
35 This hypothesis is reminiscent of the so-called "Easterlin effect,'' according to which the average labor market outcomes for larger cohorts are worsened due to the greater level of competition. Pampel and Peters
(1995) provide a review of studies that investigate the Easterlin effect.
36 Winegarden's (1987) analysis of household income inequality during the early 1970s does not include data on the earnings of husbands and wives and does not consider the effect of assortative mating. However, his econometric model and results do suggest the possibility that, for developing countries with very low levels of female labor force participation (i.e., even lower than the levels for developed nations during the 1970's), small increases in the labor force participation of women may in itself increase household income inequality.
37 In one relevant study, Boulier (1982) argues that there is no causal effect of income equality on fertility decline.
endogenous. This broader scope would yield results that are more realistic and hence more relevant to the concerns of both explanatory social science and informed public policy deliberations.
In this regard, an exemplary study is Mare's (1997) investigation of the impact of differential fertility (by race and education) on the level of educational attainment in the American population and on the racial differential in educational attainment. Mare (1997) develops a multigroup projection model which is then used in simulations based on a variety of possible assumptions about patterns of intergenerational social mobility, mortality, and the level and timing of fertility. Among his results, Mare finds that differential fertility by educational attainment has not been substantial enough to have had much impact on the overall level of educational attainment, particularly given the high level of intergenerational educational mobility that characterized much of the 20th century. Similarly, fertility differences both within and between the races have not been substantial enough to have had much of an effect on educational inequality between whites and African Americans. These important conclusions are evident in his simulations, which become feasible only after specifying a model that incorporates a variety of variables pertaining to demography and social stratification.
Another interesting study is Lerman's (1996) analysis of the effects of family structure on poverty and income inequality. The other studies on these topics reviewed above were based largely on decomposition methods that do not explicitly consider interaction or endogenous effects. By contrast, Lerman provides a richer analysis by incorporating the effects of changes in family structure on fertility, the composition of extended families, and the earnings of men and women (rather than treating the latter as being given by their observed values), which in turn have consequences for poverty and household income inequality.38 After taking into account these endogenous effects, Lerman (1996) argues that the total impact of family structure on poverty and household income inequality is significantly greater than that typically found in earlier studies based on decomposition methods. While we do not necessarily espouse this conclusion or agree with all of the details of his analysis, Lerman's (1996) more enriched approach nevertheless illustrates an important avenue for future research.
However, increasing the scope and breadth of the demography of social stratification should extend beyond simply applying more complex methods. An additional and complementary concern should be with understanding and developing cumulative knowledge about the commonalities in cross-national patterns of mobility and inequality. If increasing the complexity of analytical models means paying more attention to the details of a social process in a particular setting, the goal of understanding common patterns in cross-national research implies assessing the degree to which generic similarities may characterize the process under varied societal contexts. This latter objective of developing cumulative knowledge about generic processes is common to any scientific enterprise.
An important example of this type of research is that of Erikson and Goldthorpe (1992), who, as mentioned above, clarified the pattern underlying intergenerational class mobility throughout industrial societies. This pattern derives from particular features of the class structure, including its socioeconomic hierarchy, sectoral divisions,
38 Earlier, we also raised the issue of how increases in female labor force participation may also affect men's earnings and thus generate an additional feedback mechanism affecting the level of household income inequality.
cleavages of inheritance, and social affinities (Erikson and Goldthorpe 1987: 64-69). The finding of their research—that these features of the class structure generate a similar pattern of circulation mobility throughout the industrialized world—"ranks as a major discovery'' (Treiman and Ganzeboom 2002: 194). It also suggests important substantive and theoretical issues about the linkages between inequality and mobility that merit further investigation (Goldthorpe 2000: 232-258).
Prior to the publication of Erikson and Goldthorpe (1992), numerous influential studies of various substantive and methodological aspects of social mobility were available (some of which were reviewed above). In fact, in reaching their conclusions, Erikson and Goldthorpe build on dozens of prior studies and data collection projects that had been conducted by colleagues in various industrialized countries.39 Thus, Erikson and Goldthorpe's conclusions were made possible only because of a general concern in the research community for developing cumulative knowledge about social mobility.
Shavit and Blossfeld (1993) provide another major achievement in building cumulative knowledge about cross-national commonalities in the demography of social stratification. As was discussed earlier, these investigators detail the differentials in educational attainment by class origins in 13 industrialized nations and find that they share a great deal of stability and commonality. Inspired by the seminal analysis of Mare (1981), Shavit and Blossfeld's conclusions are another example of a collective research endeavor, and they have already led to new theoretical work that has important implications for understanding cross-national commonalities in educational inequality (Breen and Goldthorpe 1997; Goldthorpe 2000).
In closing, we have suggested that the demography of social stratification should increasingly venture beyond conventional descriptive approaches and attempt to broaden the scope of its analytical methods. Previous work in the field has been highly informative, but it has reached a sufficiently mature stage of development that would benefit from developing more complex models that integrate related processes and explicitly include endogenous variables. At the same time, however, the field should continue to pursue cross-national research in order to promote the complementary objective of building cumulative knowledge about general commonalities in the demographic processes pertaining to mobility and inequality.
Was this article helpful?