House values, rents, and wages vary drastically across metropolitan areas throughout the U.S., and geographic price differentials persist over time. Much of this variation stems from fixed differences (amenities) across locations. Using individual-level data from the 1980, 1990, and 2000 Census five percent PUMS, I determine the extent to which amenities drive geographic differences in house values, rents, and wages. I account for the endogeneity of housing prices in the determination of wages and propose an alternate method, based on metropolitan area-level residuals from wage and housing price estimations, to proxy for the collective amenities and the quality of life associated with a particular location. I rank 250 metropolitan areas based on this quality of life measure. The results indicate that quality of life differences across metropolitan areas do persist over time; in each cross section, the locations commanding a housing premium are nearly identical for renters and homeowners. However, locations where individuals accept lower wages than one would predict given their demographic and human capital characteristics do not correlate to the locations dubbed attractive in housing price regressions. These findings hold for the total wage-earning population as well as each of four permanent income groups. I assess whether these results support an amenity-based (equilibrium) theory of geographic price differentials, and find support for an equilibrium interpretation during the 1980s but not the 1990s. Comparing my findings to data on specific amenity measures yields inconsistent results, making it difficult to conclude which amenities individuals are paying for in desirable locations.

This thesis studies the influence of ethnicity, nationality, and occupation upon military selection of the residents of Ohio during the First World War. This is a quantitative study, based on a data set constructed from samples of the 1910 and 1920 censuses and _The Official Roster of Ohio Soldiers, Sailors, and Marines in the World War, 1917-1918_. Chapter I introduces the sources and the methodology. Chapter II examines the ethnicity of conscripts, and whether or not ethnic identities affected draft registration or military selection. Chapter III examines the numerical significance of resident aliens in the military population. Chapter IV examines the influence of social class on conscription: determining whether persons of wealth avoided military service and the influence of occupational deferments on the population at risk. Chapter V concludes the thesis by summarizing the results.

This paper attempts to understand the forces that have lead to persistent racial wage inequality by developing a dynamic model of statistical discrimination that accounts for the transmission of earnings across generations. The parameters of this model are then estimated using data from the 1970 and 1990 U.S. Census. The results indicate that racial disparities in the quality of information that firms receive about worker productivity are the primary cause of racial wage inequality in 1990. The results also indicate that neither the persistence of income across generations nor the presence of coordination failures explains a sizable fraction of ongoing inequality.

Since the 1960s, scholars have increasingly emphasized the ways in which routinely generated sources such as the census should be understood as creations of a quantitative mentality or statistical mind. However, this emphasis overlooks the fact that it was a particular type of thinking, namely, linearity, that in modem times characterized the features of those sources. Census forms were certainly designed by statistical minds, but, more important, they were crafted within a world view determined to impose linearity on understandings of change and the construction of social hierarchies.

Researchers consistently find that youths who work longer hours during high school tend to have higher rates of crime and substance use. On the basis of this and other research showing the negative developmental impact of an intensive work commitment during high school, the National Research Council (1998) recommended that federal lawmakers place limits on the maximum number of hours per week that teenagers are allowed to work during the school year. However, recent empirical research demonstrates the possibility of severe bias due to failure to control for unobserved sources of heterogeneity.I take advantage of two unique characteristics of the National Longitudinal Survey of Youth 1997 to assess the veracity of the claim that longer work hours are causally related to elevated involvement in crime and substance use. First, since the same respondents are followed over a period of five years, I use individual fixed effects to adjust for the omission of relevant time-stable covariates. Second, I exploit state-to-statevariation in the restrictiveness of child labor laws governing the number of hours per week allowed during the school year, and the fact that these restrictions are relaxed (and eventually expire) with increasing age. In this modelbased on a fixed-effects instrumental variables (FEIV) estimatoridentification of the work intensity effect on problem behavior is predicated on exogenous within-individual variation in school-year work hours attributable to the easing of child labor restrictions as youths age out of their legal status as minors. The attractiveness of the FEIV estimator is its ability to eliminate bias in the estimated work intensity effect due to omitted stable and dynamic variables. The model thus provides an especially powerful test of the thesis that intensive employment during the school year causally aggravates involvement in problem behavior.The empirical results demonstrate that longer work hours are associated with a significant decrease in adolescent crime, contrary to virtually all prior research. The results for adolescent substance use are mixed, suggesting the possibility that longer work hours either increase or have no effect on substance use, depending on whether a fixed-effects or first-differences procedure is implemented.

The 1990s has been characterized by a significant geographic dispersion of Mexican migrants from traditional U.S. gateways to new regions such as the South Atlantic and Northeast. Using data samples from the Mexican Migration Project and the U.S. Census, we find significant differences in the economic experiences of Mexican migrants across U.S. regions, including educational attainment, wages and remittances. In particular, the migrants region of residence explains a significant portion of the differences in these economic characteristics, as do several other individual characteristics, including occupation, year of entry, age, gender and migratory experience.

This paper consists of two parts. The first reviews the historical role that space and place has played in the discipline of demography in the United States. We argue that until approximately the middle of the 20th century nearly all of demographic analysis could be labeled “spatial demography.” Beginning in the 1940s this pattern changed, as an increasing number of microdata files from large sample surveys began to provide attitudinal and behavioral data for individuals and families. The trend was further accelerated by release in the early 1970s of census data in the form of public use microdata sample files (PUMS). We argue that in addition to data availability, the drift away from analysis of aggregated census data was prompted by a conscious desire on the part of researchers to avoid the troublesome issues of aggregation bias and what came to be known in sociology as the ecological fallacy. Sometime around the 1950s to 1960s most population analysis in the U.S. shifted away from macro- (spatial) to micro-level research, although we acknowledge and document that in two small subfields of demography (rural demography and applied demography) fascination with aggregate (spatial) data analysis persisted. In the second part of the paper, we argue that there has emerged in very recent years a renewal of interest in aggregate demographic data. Part of this re-emergence of interest in spatial demography is driven by awareness of developments in the fields of spatial econometrics and regional science that bring fresh approaches to the specification of traditional regression models and new software tools for estimating parameters in the presence of spatial externalities. By way of illustration, we provide a brief data analysis as various aspects of these new developments are discussed. Finally, we close the paper with an overview of how the earlier split between macro and micro approaches to data analysis are now being bridged with multilevel models that simultaneously consider both individual-level variables and aggregated contextual level variables for those areas where the individual lives or works.

This report presents the results of a fifteen-month project undertaken jointly by American Institutes for Research (AIR) and Management Analysis and Planning, Inc. (MAP) to answer the question posed above. The following discussion summarizes the major components of this “costing out” study. “Costing out” is a term regularly applied to this type of analysis of adequacy in education. In the course of this endeavor, AIR/MAP obtained input from professional educators and convened a full-day meeting with representatives of taxpayers, school board members, parents, legislators, and other constituencies.

We utilize an experimental Speed Dating service to examine racialpreferences in mate selection. Our data allow for the direct observation ofindividual decisions of randomly paired individuals; we may thereforedirectly infer racial preferences, which was not possible in prior studies.We observe stronger same race preferences for blacks and Asians than forHispanics and whites, with insignificant overall level of racial preferencesfor female Hispanics and males of all races. Females exhibit strongerracial preferences than males. Differences in self-reported shared interestslargely mediate the observed racial preferences. Collectively, our resultsimply strong but very heterogeneous racial preferences. Finally, wecompare our experimental results with the levels of marital segregation inthe United States.

I use data on part-American Indian children in the 1990 Census 5% PUMS to assess my hypotheses that thick racial ties within the family constrain racial identification, and that structural aspects of the community (group size, inequality, and racial heterogeneity) affect racial identification when racial ties are thin within the family. American Indians present an interesting case study because their high levels of intermarriage and complex patterns of assimilation/identity retention for generations provide a varied group of people who could potentially identify their race as American Indian. Several hypotheses are supported by the data, signifying that racial identification among people with mixed-heritage is affected by the social world beyond individual psychology and racial ties within the family.

Originally, astronomers dealt with stars. Later, with galaxies. Today, large scale cosmological structures are so complex, they must be first reduced into more succinct representations. For example, a universe simulation containing millions of objects is characterized by its halo occupation distribution. This progression is typical of many disciplines of science, and even resonates in our daily lives. The easier it is for us to collect new data, store it and manage it, the harder it becomes to keep up with what it all means. For that we need to develop tools capable of mining big data sets. This new generation of data analysis tools must meet the following requirements. They have to be fast and scale well to big data. Their output has to be straightforward to understand and easy to visualize. They need to only ask for the minimum of user input - ideally they would run completely autonomously once given the data. I focus on clustering. Its main advantage is its generality. Separating data into groups of similar objects reduces the perception problem significantly. In this context, I propose new algorithms and tools to meet the challenges: an extremely fast spatial clustering algorithm, which can also estimate the number of clusters; a novel and highly comprehensible mixture model; a sub-linear learner for dependency trees; and an active learning framework to minimize the burden on a human expert hunting for rare anomalies. I implemented the algorithms and used them with very large data sets in a wide variety of applications, including astrophysics.

This paper shows that it is important to distinguish between many and weak instrument problems in applications. We find that using Bekker (1994) standard errors that account for many instruments, fixes the problems with Angrist and Krueger (1991). We also find that many applications are in a range where this fix is sufficient. To widen the applicability of these standard errors we give theoretical results for non Gaussian models with many and many weak instruments.

This paper investigates the impact of attending college in a state on the probability of working in the state. I use information on the set of colleges students applied to as a way to account for selection in college-attendance patterns. For two samples of U.S. undergraduate students, I find a modest link between attending college in a state and working in the state. The magnitude of the effect raises doubts that location-choice considerations alone can justify state merit-scholarship programs, an increasingly popular form of student financial aid.

We propose a general framework to formalize the problem of capturing the intensity of implication for association rules through statistical metrics. In this framework we present properties that influence the interestingness of a rule, analyze the conditions that lead a measure to perform a perfect prune at a time, and define a final proper order . . .

Age at first marriage increased dramatically between 1960 and 1990, especially among blacks. The median age at first marriage among black men went from 22.5 to 28.6, and among women from 20.4 to 27.8. These increases are among the most rapid ever recoded for any historical population. Census 2000 reveals a dramatic and unexpected surprise: although median age at first marriage continued to rise for black women, it declined sharply for black men (see Figure 1). The census data indicate that black men are now marrying a year younger than black women. This paper will explore the reasons behind this unusual pattern. In particular, we will assess the effects of changing age intervals between spouses among black married couples; changing patterns of intermarriage between blacks and other racial groups; underenumeration of single black men; and changes in the income distribution of young black men.