Relative distribution methods in the social sciences (Q4700563)

In an increasing number of social science applications, the comparison of an attribute across groups requires consideration of more than the usual summary measures of location and variation. Survey and census data on attributes, such as earnings, test scores, birth weights, and survival times, all contain a wealth of distributional information. Traditional methods for the analysis of such data rely on measures that capture only differences in averages between groups or rough measures of dispersion over time. Such summary measures leave much of the information inherent in a distribution untapped. More recent exploratory data analysis techniques have provided important complementary tools for traditional methods, and have helped to change the way we look at data, check the assumptions of our models, and evaluate their performance. The motivation of authors for developing the relative distribution approach is based on this gap in existing statistical methodology. This book presents the techniques of relative distribution analysis in alternating chapters of statistical development and application. Interested readers with minimal statistical training should be able to work through the applications chapters independently to gain an understanding of the methods and their potential uses. Those interested in the statistical theory will find the chapters on measure development, estimation, and inference contain all that is required to understand and apply these techniques. Exercises are provided at the end of these chapters to reinforce key theoretical points and provide an introduction to data analysis using relative distribution methods. Computer programs and data extracts used in the book are available via the Internet. Chapter 2 provides a technical introduction to the relative distribution. It begins with a review of basic distributional concepts: probability mass functions for discrete populations, probability density functions (PDF) for continuous functions, cumulative distribution functions (CDF), quantiles (including percentiles and deciles), the quantile function, and transformation. These concepts are then used to define the relative distribution and its associated graphical representation. The chapter consludes with a review of the history and literature that contributes to the development and understanding of these methods. Chapter 3 develops the technical basis for the decomposition of the relative distribution into location, scale, and shape shifts. Chapter 4 applies the basic relative distribution methods to an analysis of the changes in the annual earnings distribution for full-time, full-year white male workers from 1967 to 1997. A decomposition of these changes into location and shape shifts shows both a decline in median real earnings, and a dramatic polarization in earnings over this period. This polarization is a key concept in the debates over growing inequality, and motivates the summary measures developed in the next chapter. Chapter 5 discusses summary measures for the divergence between two distributions. It develops a decomposition of these measures into their location and shape components. It also develops a set of summary measures for capturing polarization in the distribution: the median relative polarization index, and its component upper and lower polarization indices. Chapter 6 applies the divergence and polarization indices in an analysis of the changes in annual earnings for full-time, full-year workers by race and sex from 1967 to 1997. The analysis here focuses on the shape and location shifts that have taken place in the earnings distributions within each group. The relative distribution graphs, entropy summaries, and polarization indices provide a detailed picture of the earnings trends by group. Chapter 7 extends the methods to the situation where covariates are measured on the individuals within the groups, and the comparisons need to be adjusted to take into account any differences in the distributions of these covariates. Chapter 8 presents an application of the covariate adjustment procedures to the study of wage mobility, using data from two longitudinal panels of the National Longitudinal Survey (NLS). The location/shape decompositon is used to identify how wage growth has changed for the two cohorts. Covariate decomposition is then used to isolate the impact of differences in educational attainment between the two groups, and to contrast the trends in mobility between more and less educated workers. Chapter 9 develops the estimation and inference for the relative distribution, with emphasis on the relative CDF and PDF. The development begins with the case where the reference distribution is known, and then generalizes to the case where this distribution is estimated from the data. Chapter 10 consideres inference for summary measures based on the relative distribution. In addition to measures developed in Chapter 5, measures are motivated by considering alternative hypotheses in testing situations. Chapter 11 defines the relative distribution for discrete distributions and connects its properties to those of the relative distribution in the continuous situation. Estimation based on grouped data is discussed. Chapter 12 applies the discrete data methods to an analysis of the changes in weekly hours worked for white male workers from 1980 to 1997. A significant polarization in work schedules is observed in the data. The covariate adjustment techniques are then applied to identify over the period. Chapter 13 describes quantile estimation, focusing on quantile regression techniques. The most common model assumes the quantiles are a linear regression function of the covariates. The nonparametric quantile regression model is also considered.