Search

Supplemental material, sj-docx-1-wes-10.1177_09500170221133714 for Education-Occupation Linkage in the Highly-Educated Workforce: Patterns and Sources of Difference by Race/Ethnicity by Yao Lu, Xiaoguang Li and Benjamin Elbers in Work, Employment and Society

The aim of this project is to produce daily sub-national population estimates for Ukraine disaggregated by age and sex based on social media user activity. The report describes the methodology developed for this purpose and the supplementary materials include daily population estimates for 5-year age-sex groups in each Oblast, figures highlighting daily demographic changes, and GIS files for mapping these results.

A recent study by UC Berkeley's Othering & Belonging Institute (Menendian, Gailes, and Gambhir 2021) came to an astonishing conclusion: Of large metropolitan areas in the U.S., 81% have become more segregated over the period 1990-2019. This finding contradicts the recent sociological literature on changes in segregation in the U.S., which has generally found that racial residential segregation has slowly declined since the 1970s, especially between Blacks and Whites. The major question then is: What accounts for this difference? This paper answers this question in two parts. First, it shows that the preferred segregation measure of the Berkeley study, the “Divergence Index” (Roberto 2015), is identical to the Mutual Information Index M (Theil and Finizza 1971; Mora and Ruiz-Castillo 2009; Mora and Ruiz-Castillo 2011), a measure that is mechanically affected by changes in racial diversity. Given that the U.S. has become more diverse over the period 1990 to 2019, it is not surprising that this index shows increases in segregation. Second, by making use of a decomposition procedure developed in Elbers (2021), the paper shows that once the changes in segregation are decomposed into components that account for the changing racial diversity of the U.S., the findings are in line with the sociological literature. Residential racial segregation as a whole has declined modestly in most metropolitan areas of the U.S., although segregation has increased slightly when focusing on Asian Americans and Hispanics.

Racial residential and school segregation, while having declined in recent decades, are still pervasive in U.S. metropolitan areas. Given the consequences of segregation for individual life outcomes and its role in exacerbating inequalities in the U.S., it is of major importance to better understand the processes that shape segregation. The goal of this dissertation is to develop methods that allow us to better understand which social processes are producing increases and declines in segregation. The dissertation consists of five substantive chapters. In chapters two and four, I develop two decompositions methods that allow the decomposition of changes in segregation. The first decomposition method focuses on providing a mechanical solution to the problem of "margin dependency." Unlike alternative methods, this decomposition does not attempt to "purge'" the index from its margin dependency, but instead quantifies how much of a given change in segregation is due to changes in the margins, compared to structural changes. Arguably, this method provides more information about changes in segregation than a simple trend analysis. The fourth chapter introduces a more flexible method of decomposition, which allows the researcher to specify decompositions that are guided by theoretical considerations. This decomposition method is based on the Shapley value, originally developed in game theory. This chapter also shows that the Shapley value decomposition has many applications outside of segregation studies. The remaining substantive chapters are applications demonstrating the usefulness of these decompositions to understand changes in segregation. The third chapter applies the marginal-structural decomposition to a topic usually not considered in segregation analysis: the study of school-to-work linkages. This coauthored paper compares the skill-formation systems of France and Germany. Stratification research has often made a distinction between two ideal-types: "qualificational spaces," exemplified by Germany with a focus on vocational education, and "organizational spaces," exemplified by France with a focus on general education. Most studies that investigated this distinction did so by focusing only on the size of the vocational sector, not on whether graduates with a vocational degree actually link strongly to the labor market. Moreover, these studies often studied male workers only, ignoring potential gender differences in how school-to-work linkages are established. Our approach is instead to map the change in education-occupation linkage in France and Germany between 1970 and 2010, using the marginal-structural decomposition to distinguish between changes in rates (marginal changes) and changes in the structure of school-to-work linkages (structural changes). Surprisingly, we find that the German vocational system in 1970 was not, on average, substantially more efficient in allocating graduates to specific occupations than the French system. This finding is a major departure from earlier results, and it shows that the differences between 1970’s France and Germany, on which the qualificational-organizational distinction is based, are smaller than previously assumed. Partly, this is due to the fact that the female labor force was omitted from earlier analyses. We thus show that ignoring the female workforce has consequences for today’s conception of skill formation systems, particularly because a large share of educational expansion is caused by an increase in female enrollment in (higher) education. In the remaining two chapters, I apply the Shapley decomposition strategy to two long-standing interests of U.S. sociology: racial residential and racial school segregation. The fifth chapter, on racial residential segregation, studies changes in segregation from 1990-2010. This paper engages with a prominent concept in segregation studies, the idea of micro and macro segregation. Micro segregation refers to the small-scale neighborhood segregation within cities and suburbs, while macro segregation refers to segregation between larger geographical areas, such as cities, suburbs, and school districts. The paper first shows that, contrary to other results in literature, while micro segregation decreased, macro segregation remained at similar levels. Second, the paper shows that declines in segregation are almost exclusively caused by the Black and Hispanic populations, which have increasingly moved to majority-White areas. The sixth chapter studies changes in between-district school segregation from 2009 to 2016, studying both the relationships between school district racial composition and school district performance, as quantified by average test scores. Also in this later period, declines in segregation are mostly driven by the Black and Hispanic populations. Additionally, the decomposition by school district performance shows that families of all racial groups move from badly-performing school districts to better-performing districts.

Supplemental material, sj-docx-1-srd-10.1177_23780231211053982 for Trends in U.S. Residential Racial Segregation, 1990 to 2020 by Benjamin Elbers in Socius

Supplemental Material, sj-docx-1-smr-10.1177_0049124121986204 for A Method for Studying Differences in Segregation Across Time and Space by Benjamin Elbers in Sociological Methods & Research

Many regression modeling problems involve ordinal predictors with equally spaced values, such as age or income groups. Orthogonal polynomial contrasts provide a useful way to incorporate the information that the groups are ordered, unlike dummy or effects coding. This technique has been useful to separate the linear and non-linear effects in age-period-cohort (APC) models. This paper addresses three practical issues when implementing orthogonal polynomial contrasts: First, if the linear effect is to be interpreted, it should be identical to the linear effect that we obtain if we treat the predictor as a metric variable. Second, if the data is unbalanced, weighted orthogonal contrasts need to be used to obtain unbiased estimates. Third, by dropping-higher order terms, orthogonal polynomials can easily be used to smooth non-linear estimates. Two examples from the APC literature are discussed, and it is found that the use of unweighted polynomials has led to small biases in the results. A R function is provided that implements weighted polynomials with an interpretable linear term.

An important topic in the study of segregation are comparisons across space and time. This article extends current approaches in segregation measurement by presenting a five-term decomposition procedure that can be used to understand more clearly why segregation has changed or differs between two comparison points. Two of the five terms account for differences in segregation that are due to the differing marginal distributions (e.g., the gender and occupational distributions), while one term accounts for differences in segregation due the different structure of segregation (what might be termed “pure” segregation). The decomposition thus presents a solution to the problem of margin dependency, frequently discussed in the segregation literature. Finally, two terms account for the appearance or disappearance of units when analyzing change over time. The method can be further extended to attribute structural changes to individual units, which makes it possible, for instance, to quantify the effect of each occupation on changing gender* segregation. The practical advantages of the decomposition are illustrated by two examples: a study of changing occupational gender segregation in the United States and a study of changing residential segregation in Brooklyn, New York.