This paper examines the productivity (and wage) gains from locating in dense, urban environments. We distinguish between three potential explanations of why firms are willing to pay urban workers more: (1) the urban wage premium is spurious and is the result of omitted ability measures, (2) the urban wage premium works because cities enhance productivity and (3) the urban wage premium is the result of faster skill accumulation in cities. Using a combination of standard regressions, individual fixed effects estimation (using migrants) and instrumental variables methods, we find that the urban wage premium does not represent omitted ability bias and it is only in part a level effect to productivity. The bulk of the urban wage premium accrues over time as a result of greater skill accumulation in cities.

This article examines ethnic marriage patterns for both men and women in New York City from public use samples of United States census schedules for years with roughly comparable data: 1900, 1910, 1960, and 1980. Log-linear models which take into account both the married and the unmarried as well as differences by gender reveal the structures of ethnic pairings at each census and illuminate important transitions....The intermarriage transition in New York City suggests that, in the first stages of immigration, unbalanced ethnic sex-ratios were powerful forces precipitating intermarriage. The imbalance extended to second-generation ethnics, who, faced with heightened competition from the continued influx of ethnic compatriots, responded by marrying out instead of not marrying at all.

The author discusses problems involved in modeling the effects of demographic factors on historical kinship patterns, with a focus on microsimulation models. "Since microsimulations of kinship ignore the correlations in demographic behavior within kin groups, they ordinarily understate the variance of kinship distributions; for many kin types, they also underestimate the expected number of kin." The author concludes that "those who design demographic models of kinship should be sensitive to the potential for systematic error."

This study posits a juncture of historical context and the incidence of interracial marriage. Demographic opportunity structures are certainly important. But individuals also assess options and risks--both real and perceived--in making decisions. This cultural context changes historically. Interracial marriage provides a vantage point for analyzing shifts in the boundaries between groups. Interracial marriage is about race and about gender. And it is about conflicting ideals within American identity--community versus individual freedom of choice. Historical contexts in which individuals assess options--both real and perceived--are studied. The nature of the proscription against interracial marriage is described, using legal evidence and public opinion polls. Legal variations by time and region define cultural regions for interracial marriage, which affect both real opportunity and perceptions of risk. A measure of the level and type of popular debate is constructed through an analysis of citations in the Readers' Guide. These mass circulation magazines explicate the cultural debate over shifting options, expectations and social mores. Although there is debate over the full diversity within American society, most of the debate focuses on intermarriage between African-Americans and Caucasian-Americans. Finally, films are used to illustrate the power of the proscription in defining even fantasy and art. Public Use Microdata Samples of the United States Census from 1900 through 1980 are used to statistically analyze interracial marriage. National estimates of interracial marriage over time are established. These are further broken down by race, gender and region. Variables are constructed to compare actual marriage rates to the snapshot view of extant families measured by the census. Log-linear analysis examines the likelihood that different types of individuals will choose interracial marriage over remaining single, taking into account the demographic opportunities available. Perceptions are clearly critical. Popular opinion proves to be acutely attuned to differences in behavior and changes over time, even when those differences are difficult to measure because the numbers are extremely small.

Because US Census Bureau reports summarize population counts through cross-tabulation of individual or family characteristics, the criteria for which changes over time, the information is not ideal for analytic research, especially in history. Through the release of census data in "public use sample" form, microdata census samples covering the last one hundred years will soon be available (except for the years 1920, 1930, and 1980). Important differences exist between these important samples, however, in the following categories: census format and the consequent sample format, units of enumeration and sampling, strategies of constructing the samples, methods of reporting household composition, geographic position, ethnicity, economic status, education, and literacy.