Our understanding of gender differentiation in contemporary Chinese society is hampered by the lack of comprehensive studies with a consistent framework. In consequence, findings of the few well-designed surveys in China’s largest cities are routinely generalized to all of urban China, and the still fewer rural studies in scattered localities are often the basis for even more egregious overgeneralization. Quantitative analyses of gender using aggregate data have for the most part been presented as dichotomies (urban vs. rural, coastal vs. interior) or limited to provinces.

The papers in this panel move us toward a more comprehensive and discriminating picture of gender in China by placing relevant disaggregated data in a fine-grained spatial framework that reflects the actual structure of China’s regional economies/societies. We apply a model of Hierarchical Regional Space (HRS) that views Chinese society as a nested hierarchy of nodal local and regional systems, each centered on a city or town at one of eight levels in the urban hierarchy. The conceptualization and methodology are briefly characterized in the first two papers, but the focus in all papers is on substantive findings. The analyses of this panel demonstrate systematic and sharp spatial differentiation: (1) in the gender mix of particular occupations; (2) in the gap between male and female educational attainment; (3) in the degree of sex bias in infant mortality; and (4) in the extent of parental strategizing to shape the gender configuration of offspring. Thus, we see that individual opportunities and family choices vary with position in the spatial economy.

Mark Henderson, University of California, Berkeley; Michele Ladenson, University of California, Davis

This paper analyses spatial patterns in the gender composition of particular occupations in contemporary China. We present findings here for the Lower Yangzi macroregional system, centered on Shanghai, with a 1990 population of 140 million. We apply a model of Hierarchical Regional Space (HRS) that highlights variation in socioeconomic patterns at different spatial scales. Two levels of this model apply to county-level data from the 1990 census: Core-Periphery Zoning, capturing structure at the macroregional scale, and an Urban-Rural Continuum index, capturing the relative urbanness of settlements.

Family choices about economic diversification and migration play out in these spatially differentiated structures. We find three distinct spatial patterns: (1) female representation in high-status professions (echoing educational attainment) is highest in the most urban units of the inner core, declining toward the periphery and down the systems hierarchy to lowest in villages of the far periphery; (2) in the case of lower-status non-agricultural occupations, female representation is highest in urban centers of the far periphery, declining toward the regional core and down the systems hierarchy to lowest in villages of the inner core; and (3) the third pattern, which holds for workers in agriculture, displays a pattern that is just the reverse of the second, namely: female representation is highest in inner-core villages, declining toward the periphery and up the systems hierarchy to lowest in urban centers of the far periphery. Thus, the widely noted feminization of agriculture is a phenomenon of the rural core, most pronounced in the inner core. We argue that the second and the third patterns result from sex-selective migration responding to spatially differentiated opportunity structures.

Wei Wang, University of California, Davis

This paper compares the educational attainment of men and women in contemporary China, using individual data from a 1 percent household sample of the 1990 census. We present findings here for the two regional systems in Northwest China: the Wei-Fen system, whose metropolises are Xi’an and Taiyuan, and the Upper Yellow River system, whose metropolises are Lanzhou and Baotou. In 1990 the two regions together had a population of 78 million.

We begin with an analysis of the educational attainment of prime-age adults (20–54) by position in hierarchical regional space (HRS). For those completing each level of schooling, from primary to university, the proportion of females is highest in the most urban units of the inner core declining through the core-periphery structure and down the systems hierarchy to lowest in villages of the far periphery. For instance, women comprise 35.4 percent of those with advanced education in the urbanized innermost core but only 10.9 percent of the highly educated in villages of the farthermost periphery. In a further analysis of educational progression by birth year and sex, we find that gender equality in educational progression ratios is more closely approached at higher levels of educational attainment and in regional cores. Gender bias is most pronounced in access to primary education in villages of the far periphery.

William M. Mason, University of California, Los Angeles

Infant mortality is a shadowy area of Chinese demography. Sex differentials and spatial variability are, needless to say, murkier still. There are clear indications that in some regions, female infant survival is greatly inferior to male, but we have only vague notions about the factors that underlie these variations. Broadly speaking, hypotheses appeal to socioeconomic variation, subcultural variation, and variation in the administration of birth planning policies. This paper attempts to assess the components of spatial variability in a major Chinese subregion. It makes use of individual level data on infant survival reflecting mortality conditions in 1989–90, derived from a 1 percent clustered sample of the 1990 Chinese census. We will view infant mortality (and their sex differentials) as influenced by a nested hierarchy of contexts, including the position of the birth within the sibset, mother’s education, household characteristics, and village, county, and prefectural characteristics. We will attempt to distinguish organizational causes from those arising from individual choice. We suspect, for example, that behavior is highly influenced by policies adopted at the village, county, and prefectural levels. It will be possible to map variability in infant mortality net of a host of individual-level characteristics. If there is a left-over spatial variability component that clearly maps into higher-level administrative subdivisions, this would be evidence of the influence of administrative policies at those levels.

Jianhua Yuan, Beijing Institute of Information and Control

This paper treats the changing reproductive strategies of Chinese families during the period 1966–90, using household data from a 1 percent sample of the 1990 census. We present findings here for a contiguous group of regional systems in East and Central China whose combined population in 1990 was 625 million.