Über die Autorin bzw. den Autor

Stephen L. Morgan is Associate Professor of Sociology at Cornell University.

Von der hinteren Coverseite

Aus dem Klappentext

Auszug. © Genehmigter Nachdruck. Alle Rechte vorbehalten.

ON THE EDGE OF COMMITMENT

STANFORD UNIVERSITY PRESS

Contents

List of Tables and Figures.............................................................................................................xAcknowledgments........................................................................................................................xiiiINTRODUCTIONChapter One Why Do We Need a New Model of Educational Attainment?.....................................................................3COMING TO TERMS WITH THE SOCIOLOGICAL LITERATURE ON EDUCATIONAL ATTAINMENTChapter Two Expectation Formation and the Status Socialization Theory of Educational Attainment.......................................35Chapter Three Do Beliefs Matter? A Reanalysis of the Relationship Between Educational Expectations and Attainment.....................57A COMMITMENT-BASED MODEL OF EDUCATIONAL ATTAINMENTChapter Four The Generation of Preparatory Commitment from Forward-Looking Beliefs....................................................99Chapter Five The Evolution of the Beliefs That Determine Commitment...................................................................139Chapter Six Incorporating Imitative and Normative Sources of Commitment...............................................................175Chapter Seven A New Agenda for the Sociology of Educational Attainment................................................................207References.............................................................................................................................219Index..................................................................................................................................233

Chapter One

Explanations for patterns of social inequality are only as strong as the models of educational attainment on which they depend. Educational training creates and then uniquely signals many of the skills and habits that determine styles of life and economic well-being. Educational credentials and the social connections they embody facilitate the allocation of individuals to alternative occupational and labor market positions. No convincing causal model of intergenerational mobility can be fashioned without incorporating these mechanisms.

Shifting from one depiction of the structure of inequality to another does not relieve the burden of having to account for differences in educational attainment. Class schemas, prestige hierarchies, and labor markets are similarly incomprehensible without an explanation of how the distribution of educational attainment emerges. Even explanations for patterns of ascriptive inequality are crucially reliant on foundational models of educational processes. The extent of racial and gender discrimination in the labor market, for example, cannot be assessed without taking a position on how and why individuals accumulate alternative educational credentials and skills.

This justification for rigorous modeling of educational attainment, on which generations of social stratification researchers and sociologists of education have relied, is even more compelling now. Over the past three decades, the evolution of postindustrial society has increased the stakes for comprehensively modeling patterns of educational attainment, as these now more strongly predict economic well-being. Between 1979 and 1999, the real wages of high school graduates decreased by 8.9 percent, whereas the real wages of college graduates and advanced degree graduates increased by 13.0 and 18.9 percent, respectively (Mishel, Bernstein, and Schmitt 2001: 153). By the late 1990s, levels of labor market inequality in the United States reached levels more extreme even than those observed in the 1930s. And therefore, as educational attainment becomes a stronger predictor of lifetime well-being, understanding why some adolescents carry on to postsecondary education and others do not has become increasingly important.

Unfortunately, it is also becoming clear that current models of educational attainment are insufficiently complete. Consider the capacity to determine the consequences of the growth in labor market inequality for patterns of educational attainment. Over the same time period, policies designed to increase college enrollments have changed only modestly. From a variety of theoretical perspectives, it can be argued that increases in labor market inequality should have variable effects on different groups of prospective college students. The increasing incentives for obtaining college degrees should prompt more students to enter college, but the rate of increase in college enrollments should differ by social background. One would expect the increase to be less pronounced for prospective students from relatively disadvantaged social origins, for these students' relative access to liquid funds to finance a college education has declined as inequality has grown.

Explanatory models of educational attainment in the social sciences are unable to determine whether this prediction is accurate, to say nothing of the underlying mechanism that supposedly generates it. None of the established models, I will argue in this book, can be used to determine whether or not students' beliefs about their future prospects are responsive to changes in incentives, and, if so, how changes in beliefs affect the long sequence of commitment decisions that determine a student's final level of educational attainment. The main goal of this book is to develop the foundation for a new and better model of educational attainment, one that can generate sufficiently complete explanations by specifying and modeling the belief-based mechanisms that have been assumed away in the past.

Before proceeding to development of the model, in this introductory chapter I present the basic facts on college entry and college completion in the United States. I then demarcate the limits of current knowledge and develop two sets of reasons why the social sciences need a new and better model of educational attainment-to resolve empirical puzzles and to guide better policy development and evaluation. I then discuss why sociology as a discipline is well prepared to produce and embrace a new model, and I conclude by justifying the mode of theoretical development that I will pursue in subsequent chapters.

THE BASIC FACTS ON EDUCATIONAL ATTAINMENT IN THE UNITED STATES

Who goes to college, and who graduates? Table 1.1 presents the college entry and graduation patterns of the high school classes of 1982 and 1992, as of 1990 and 2000, respectively. The findings are based on parallel analyses of the two most widely used national surveys sponsored by the U.S. Department of Education, the High School & Beyond Survey (HS&B) and the National Education Longitudinal Study (NELS) (U.S. Department of Education 1995, 2002).

In the first column of each panel, the college entry patterns of high school graduates are presented separately for the four largest racial/ethnic groups in the two surveys. For example, for white students of the class of 1982, 53.84 percent of students entered college within one year of graduating from high school, 18.10 percent entered college more than one year after graduating from high school, and the remaining 28.07 percent never entered college (as of 1990, the final year of the measurement window).

In the second column of each panel, bachelor's degree attainment patterns are then tabulated, conditional on the timing of college entry. Again, for white students, 35.79 percent of those who entered college within one year of high school graduation then obtained a bachelor's degree within five years. An additional 16.60 percent obtained a bachelor's degree in more than five years, and the remaining 47.60 percent never obtained a bachelor's degree (again, as of 1990, the final year of the measurement window).

Although there are many patterns within Table 1.1, focus first on the findings common to all groups. Among those students who enter college, many students do not obtain a bachelor's degree. Some of these students enter two-year colleges and never transfer to four-year colleges. Others simply drop out of college after one or more years. Some work part time, obtaining only a few college credits per year, making steady but slow progress toward the credentials they hope to secure at some point in adulthood.

Aside from this pattern, which demonstrates clearly that college entry does not on its own guarantee college graduation, the table shows important racial and ethnic differences in college entry and graduation patterns. Black and Hispanic high school graduates are less likely than white high school graduates to enter college immediately after high school. Asians, in contrast, are much more likely to do so. And, once in college, whites and Asians are similarly likely to receive bachelor's degrees while blacks and Hispanics are much less likely to receive them.

Across all groups, between the classes of 1982 and 1992, college entry and college graduation rates increased notably. This pattern is consistent with other evidence that I will present later, but note the differences in the sampling schemes and measurement instruments of the NELS and HS&B surveys are substantial enough that one should approach with caution the interpretation of the over-time differences revealed in Table 1.1. For example, although the rate of college entry increased for Hispanic students, the college completion rate appeared to (1) decline for those who entered college immediately but (2) increase for those who entered college more than one year after high school graduation. Although possibly true, the sampling scheme for Hispanics changed between the two surveys (as did the target populations because of immigration patterns), and as a result I will refrain from overinterpreting such changes. Indeed, for much of the empirical analysis in this book, I will focus narrowly on whites and blacks, in part because over-time comparisons are more tricky for other groups, and in part because the classic literature in the sociology of education-which I aim to reinterpret and extend-is often built around black-white comparisons of educational outcomes.

To further assess changes in college entry rates, consider the next set of findings, based on an analysis of the 1976 through 2000 October Current Population Surveys, the data source most commonly used to track changes in college entry rates. Figure 1.1 presents estimates of the college entry rate of recent high school graduates between the ages of 18 and 21, separately for white males, white females, black males, and black females. The trends presented in Figure 1.1 are consistent with the college entry rates of whites and blacks from Table 1.1. But, because many more graduating cohorts of high school seniors are analyzed, additional patterns and more recent trends are revealed as well.

Notice, first, the decline in the college entry rate of blacks in the early 1980s. Immediately after this sharp decline, the rate of college entry among blacks rebounded quickly. By 2000, black males and females were nearly as likely as white males to enter college after graduating from high school. But a new trend emerged in the 1990s. White females began to enter college at a record pace. By 2000, nearly 70 percent of white female high school graduates entered college immediately after graduating from high school.

What predicts college entry and college graduation? Table 1.2 presents the results of nine separate models where the three-category breakdown of college entry (from the first column of each panel of Table 1.1) is specified as the outcome variable. Each of the nine variables listed in Table 1.2 is specified in a separate multinomial logit model as a linear predictor of the probability of entering college within one year of graduating from high school and simultaneously of the probability of entering college more than one year after graduating from high school (both in comparison to not entering college by the end of the measurement window, 1990 and 2000 respectively).

For the discrete change in the predictor variable presented in the relevant cells in columns one and two of the table, the last four columns report the corresponding changes in these estimated probabilities. For example, for the class of 1982, the probability that a student would enter college within one year of high school graduation was .274 higher for students whose fathers were college graduates (that is, those whose father's education was equal to 16 years) than for students whose fathers were only high school graduates. Likewise, the probability that a student would enter college more than one year after high school graduation was .051 lower for students whose fathers were college graduates. This implies that the probability of ever entering college was .223 higher for students whose fathers were college graduates (that is, .274 - .051). And, the simple qualitative conclusion to draw from the estimates is that having a father who is a college graduate increases one's likelihood of entering college and also of entering college immediately after graduating from high school.

The next four rows present similar estimated changes in probabilities for mother's education, father's and mother's occupational prestige, and family income. The designated discrete change for mother's education is again 12 versus 16 years of education. The discrete change for occupational prestige is fixed at the SEI values of 35 and 48, which is close to one standard deviation centered on the mean of occupational prestige. For family income, the designated discrete change is also centered around the observed mean of family income, but it is proportionately much narrower at only $6,000 in 1992 and $4,150 in 1982, an equivalent amount after adjusting for inflation. This dollar amount is roughly equivalent to the combined costs of tuition and fees of $2,300 and living expenses of $3,700 at a typical four-year public university in the 1990s (see Kosters 1999).

The resulting predicted probabilities imply the same basic pattern as for father's education. For example, students from the class of 1992 and from families with $6,000 more income were more likely to enter college immediately (.03 in probability), less likely to enter after one year (-.012), and hence in combination more likely to ever enter college (at .018 in probability, which equals .03 - .012).

For these education, occupational prestige, and family income variables, the predicted probability changes are slightly different for the earlier time period, suggesting a greater increase in the probability of entering college immediately, a smaller decline in the probability of entering college after more than one year, and in combination a greater relative increase in the probability of ever entering college for the class of 1982. One could interpret this small cohort difference as an indication that family background was more predictive of college entry patterns for the earlier cohort, but since the two datasets are slightly different, results differing by this magnitude could have emerged because of study design differences. And yet, no matter what interpretation one gives to these small differences, there is surely no evidence of a large increase in the importance of family background for college entry patterns, as one might hypothesize would result from the generalized increase in inequality in the 1980s and 1990s.

What about the characteristics of students themselves? For the models reported in Table 1.2, test scores in the sophomore and senior years were specified analogously in separate models. The specified discrete changes are equal to one standard deviation on the relevant test, centered around the mean test score. The resulting estimates show that for both cohorts, the probability of entering college immediately after graduating high school was higher by at least .2 for those students who scored one-half of a standard deviation above the mean than for those who scored one-half of a standard deviation below the mean.

In the last two rows of the table, two additional predictor variables were deployed, each of which will be discussed in detail throughout this book, especially in Chapters 2 and 3. Significant others' influence is the proportion of a student's parents, teachers, and peers whom the student felt expected him or her to attend college immediately after high school. Educational expectations are answers to the question "How far in school do you expect to get?"-which were then converted from levels of educational attainment to a years-completed metric. The last two rows of the table show that the views of students' significant others and their forecasts of their own future behavior are strongly predictive of the probability of entering college.

Table 1.3 repeats the same basic analysis for the college graduation patterns of Table 1.1, but only for those students who entered college within one year after high school graduation. Virtually all of the patterns are the same, except that there is no estimated movement out of the category "Obtain bachelor's degree in more than five years" in the way that there was movement in the models of Table 1.2 out of the category "Enter college after more than one year after high school graduation." That is to say, empirical support for shifts toward faster bachelor's degree completion (as opposed to slower bachelor's degree completion) was not as strong as evidence for shifts toward immediate college entry (as opposed to delayed college entry).

(Continues...)

Excerpted from ON THE EDGE OF COMMITMENTby Stephen L. Morgan Copyright © 2005 by Board of Trustees of the Leland Stanford Junior University. Excerpted by permission.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.