The Truly Advantaged: Examining the Effects of Privileged Places on Educational Attainment

ABSTRACT

Inspired by William J. Wilson’s The Truly Disadvantaged, hundreds of studies have focused on the detrimental effects of disadvantaged neighborhoods. Consequently, far less is known about the contextual effects of advantaged neighborhoods, and what is known does not take into consideration long-term exposure. The present study extends research on advantaged neighborhoods by examining how respondents' neighborhood contexts across their entire childhoods influence adult educational attainment. Findings indicate that structural effects in advantaged neighborhoods influence residents’ educational attainment—especially for White residents. Results suggest that addressing the issues associated with the truly disadvantaged requires examining the compounding privilege of the truly advantaged.

KEYWORDS:

• Neighborhood effects
• disadvantaged neighborhoods
• neighborhood inequality educational mobility

Acknowledgments

Special thanks to Michael Emerson, James Elliott, Elizabeth Korver-Glenn, Kevin Smiley, Ruth Lopez Turley, and Rachel Kimbro, who gave feedback on this work at various stages of its development, as well as the editors and anonymous reviewers for their helpful comments.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. In the late 1990s, the PSID added immigrant families to reflect the increase in immigration since 1968. However, these added families do not yet have enough data to meet the conditions of this study. Given the findings of Howell and Emerson (2017), I operationalize all multiracial individuals as the racial group lower on the ethnoracial hierarchy.

2. After 1997, PSID conducted surveys biannually. Thus, depending on the respondents’ birth year, they might be interviewed at age 25 and age 27 but not age 26. For these individuals, we use their educational obtainment at age 27.

3. Additional disadvantage indexes were considered with various neighborhood characteristics such as proportion with at least a high school diploma, mobility rate, unemployment rate, and average room per capita. Results were comparable to those presented in the article and one additional index, the neighborhood privilege index, is used as a robustness check. I chose the three variables used in the study because of their theoretical relevance to the literature’s primary arguments.

4. Cronbach’s alphas measure the internal consistency between variables. Over time the Black population has desegregated and become less concentrated in poor tracts with high proportions of single-parent families. Hence, the disadvantaged index’s Cronbach's alphas have decreased.

5. Given that not all places were assigned census tracts or BNAs (rural tracts) in the 1970 census, the neighborhood disadvantage index was calculated starting in 1980. Because the sample was born between 1975 and 1985, this means that for some individuals the neighborhoods they lived in during the first few years of their childhood were not included in the averages.

6. Scholars often combine “disadvantaged” (i.e., poverty rate) and “privileged” (i.e., proportion with bachelor’s degrees) neighborhood characteristics into one factor (e.g., Chetty, Hendren, and Katz 2016; Turley 2003; Vartanian and Buck 2005; Vartanian and Houser 2010; Wodtke, Harding, and Elwert 2011). I explore such combinations starting with seven different variables common in the literature, including Black proportion, poverty, single-parent families, median income, bachelor’s proportion, unemployment rate, and percentage in professional or managerial occupations. However, internal consistency is higher when these variables are divided into two indexes. Moreover, the correlation between the disadvantaged and privileged index is only −0.4457 in 1980, −0.4826 in 1990, −0.5925 in 2000, and −0.5722 in 2010. These findings echo Small, Manduca, and Johnston's (2018) research, which suggests that poverty and wealth coexist within neighborhoods. Using two indexes enables me to capture the existing neighborhood heterogeneity.

7. PSID to date has only allowed respondents to choose from binary gender categories.

8. All models were also run with regional controls. However, region was not substantively or statistically significant and thus was not included in the final models.

9. Though the exact dollar amounts are displayed in the descriptive tables, for the models, income was divided by 10,000 so that coefficients were easier to interpret.

10. Alternative approaches include parental educational attainment at one point in time or the highest educational attainment in the entire time period. These approaches, however, lose some of the complexity available in the data. For example, consider a hypothetical family whose father is college educated but whose mother has a high school diploma. The father dies when the child is 5. Though this child will benefit from having a father who was college educated, he or she was primarily raised by his or her high school–educated mother. Thus, the child's educational attainment will likely reflect this fact. The temporal approach enables us to capture this complexity.

11. Previous scholars have used sibling fixed effects models to examine within-family differences net of unobserved family-level confounders. Although most siblings have slightly different neighborhood conditions across their childhoods, the vast majority of families remain in socioeconomic and racially similar communities (Sampson 2012; Sharkey 2013; Sharkey and Elwert 2011). In fact, familial race persist as one of the largest determinants of neighborhood location (Besbris 2016; Howell and Emerson 2018; Howell and Korver-Glenn 2018; Korver-Glenn 2018). Thus, fixed effects models are unable to capture how neighborhood effects differ across the most and least advantaged neighborhoods.

12. To be clear, neighborhoods can have differing values on individual variables and the same neighborhood disadvantaged factor because they are composite scores. Thus, the proportions provided here are simply the mean values across all neighborhoods with a particular factor score.

13. This is the average proportion for neighborhoods with a neighborhood disadvantage score greater than 4.5.

14. In addition, supplemental models using a hybrid approach found the both within-family (fixed effects) coefficients and between-family (random effects) coefficients were comparable in magnitude, direction, and statistical significance to the results presented in Table 2.

15. Before I introduced the quadratic term, I shifted the neighborhood disadvantage index to the right such that the minimum value was zero instead of negative one. This ensures that the quadratic term is distinct for the positive and negative values.

16. Although the two are correlated (r = 0.53), multicollinearity is not an issue in the model. See endnote 6 for further explanation.

Additional information

Funding

This work was supported by Rice University’s Social Science Research Institute (U50176 869000).

Notes on contributors

Junia Howell

Junia Howell, Ph.D., is an Assistant Professor of Sociology at the University of Pittsburgh. Howell’s work investigates how local and national policies perpetuate racial and socioeconomic inequality within urban context. Specifically, she applies critical race theory to urban sociological investigations of neighborhood effects, racial residential segregation, home appraisals, and natural hazards. Her mixed‐methods and cross‐national research illuminates how policies and practices enable opportunity hoarding among White middle‐class communities, exacerbating urban inequities.