The most disadvantaged children: Trends over time

Social Service Review

Volumn 73, Issue 4, 1999, Pages 560-578

  Foster, E Michael ^a   Furstenberg Jr , Frank F ^a  

^a NONE

Author keywords

[No Author keywords available]

Indexed keywords

EID: 0040408728     PISSN: 00377961     EISSN: None     Source Type: Journal    
DOI: 10.1086/514445     Document Type: Article

References (50)

1. Leatha Lamison-White, Poverty in the United States: 1996, Current Population Reports, U.S. Bureau of the Census (Washington, B.C.: Government Printing Office, 1997).
2. Government statistics show a large racial gap in employment among families with children. In 1995, only 11 percent of white families with children had no family work experience. This is just over half the rate for African Americans (20 percent). The racial gap in family structure is well-known. In 1996, 18 percent of white families with children were headed by women. This is roughly one-third the proportion for African Americans (55 percent). Finally, African-American children are much more likely to live in a family that received welfare. In 1995, 11 percent of white children lived in families that received cash benefits. Over one-third (37 percent) of African-American children lived in such families. See U.S. Census Bureau and Bureau of Labor Statistics, Presence of Related Children under 18 Years Old: All Families, by Total Money Income in 1996, Type of Family Work Experience in 1996, Race and Hispanic Origin of Reference Person (May 1998, http://ferret.bls.census.gov/macro/ 031997/faminc/04_001.htm), FINC-04, and "Table 3: Program Participation Status of Household-Poverty Status of Persons in 1995" (May 1998, http://ferret.bls.census.gov/ macro/031996/pov/3_001.htm).
3. Government statistics show a large racial gap in employment among families with children. In 1995, only 11 percent of white families with children had no family work experience. This is just over half the rate for African Americans (20 percent). The racial gap in family structure is well-known. In 1996, 18 percent of white families with children were headed by women. This is roughly one-third the proportion for African Americans (55 percent). Finally, African-American children are much more likely to live in a family that received welfare. In 1995, 11 percent of white children lived in families that received cash benefits. Over one-third (37 percent) of African-American children lived in such families. See U.S. Census Bureau and Bureau of Labor Statistics, Presence of Related Children under 18 Years Old: All Families, by Total Money Income in 1996, Type of Family Work Experience in 1996, Race and Hispanic Origin of Reference Person (May 1998, http://ferret.bls.census.gov/macro/ 031997/faminc/04_001.htm), FINC-04, and "Table 3: Program Participation Status of Household-Poverty Status of Persons in 1995" (May 1998, http://ferret.bls.census.gov/macro/031996/pov/3_001.htm).
4. E. Michael Foster and Frank F. Furstenberg, Jr., "Most Disadvantaged Children: Who Are They and Where Do They Live?" Journal of Poverty 2, no. 2 (1998): 23-47.
5. Note that we do not assume that these disadvantages cause undesirable outcomes. The issue of the causal impact of each of these disadvantages has been explored in greater detail elsewhere. For a discussion of the causal impact of family structure, see Sara McLanahan and Gary D. Sandefur, Growing Up with a Single Parent: What Hurts, What Helps? (Cambridge, Mass.: Harvard University Press, 1994). Similarly, other studies examine the effect of poverty (Greg J. Duncan and Jeanne Brooks-Gunn, Consequences of Growing Up Poor [New York: Russell Sage Foundation, 1997]); welfare receipt (Jay D. Teachman, Kathleen M. Paasch, Randal D. Day, and Karen P. Carver, "Poverty during Adolescence and Subsequent Educational Achievement," in Consequences of Growing Up Poor, ed. Greg J. Duncan and Jeanne Brooks-Gunn [New York: Russell Sage Foundation, 1997], pp. 382-418); and neighborhood conditions (E. Michael Foster and Sara McLanahan, "An Illustration of the Use of Instrumental Variables: Do Neighborhood Conditions Affect a Young Person's Chance of Finishing High School?" Psychological Methods 3, no. 1 [1996]: 249-60; and E. Michael Foster, "Instrumental Variables for Logistic Regression: An Application of the Generalized Method of Moments," Social Science Research 26, no. 4 [1997]: 487-504).
6. Note that we do not assume that these disadvantages cause undesirable outcomes. The issue of the causal impact of each of these disadvantages has been explored in greater detail elsewhere. For a discussion of the causal impact of family structure, see Sara McLanahan and Gary D. Sandefur, Growing Up with a Single Parent: What Hurts, What Helps? (Cambridge, Mass.: Harvard University Press, 1994). Similarly, other studies examine the effect of poverty (Greg J. Duncan and Jeanne Brooks-Gunn, Consequences of Growing Up Poor [New York: Russell Sage Foundation, 1997]); welfare receipt (Jay D. Teachman, Kathleen M. Paasch, Randal D. Day, and Karen P. Carver, "Poverty during Adolescence and Subsequent Educational Achievement," in Consequences of Growing Up Poor, ed. Greg J. Duncan and Jeanne Brooks-Gunn [New York: Russell Sage Foundation, 1997], pp. 382-418); and neighborhood conditions (E. Michael Foster and Sara McLanahan, "An Illustration of the Use of Instrumental Variables: Do Neighborhood Conditions Affect a Young Person's Chance of Finishing High School?" Psychological Methods 3, no. 1 [1996]: 249-60; and E. Michael Foster, "Instrumental Variables for Logistic Regression: An Application of the Generalized Method of Moments," Social Science Research 26, no. 4 [1997]: 487-504).
7. Note that we do not assume that these disadvantages cause undesirable outcomes. The issue of the causal impact of each of these disadvantages has been explored in greater detail elsewhere. For a discussion of the causal impact of family structure, see Sara McLanahan and Gary D. Sandefur, Growing Up with a Single Parent: What Hurts, What Helps? (Cambridge, Mass.: Harvard University Press, 1994). Similarly, other studies examine the effect of poverty (Greg J. Duncan and Jeanne Brooks-Gunn, Consequences of Growing Up Poor [New York: Russell Sage Foundation, 1997]); welfare receipt (Jay D. Teachman, Kathleen M. Paasch, Randal D. Day, and Karen P. Carver, "Poverty during Adolescence and Subsequent Educational Achievement," in Consequences of Growing Up Poor, ed. Greg J. Duncan and Jeanne Brooks-Gunn [New York: Russell Sage Foundation, 1997], pp. 382-418); and neighborhood conditions (E. Michael Foster and Sara McLanahan, "An Illustration of the Use of Instrumental Variables: Do Neighborhood Conditions Affect a Young Person's Chance of Finishing High School?" Psychological Methods 3, no. 1 [1996]: 249-60; and E. Michael Foster, "Instrumental Variables for Logistic Regression: An Application of the Generalized Method of Moments," Social Science Research 26, no. 4 [1997]: 487-504).
8. Note that we do not assume that these disadvantages cause undesirable outcomes. The issue of the causal impact of each of these disadvantages has been explored in greater detail elsewhere. For a discussion of the causal impact of family structure, see Sara McLanahan and Gary D. Sandefur, Growing Up with a Single Parent: What Hurts, What Helps? (Cambridge, Mass.: Harvard University Press, 1994). Similarly, other studies examine the effect of poverty (Greg J. Duncan and Jeanne Brooks-Gunn, Consequences of Growing Up Poor [New York: Russell Sage Foundation, 1997]); welfare receipt (Jay D. Teachman, Kathleen M. Paasch, Randal D. Day, and Karen P. Carver, "Poverty during Adolescence and Subsequent Educational Achievement," in Consequences of Growing Up Poor, ed. Greg J. Duncan and Jeanne Brooks-Gunn [New York: Russell Sage Foundation, 1997], pp. 382-418); and neighborhood conditions (E. Michael Foster and Sara McLanahan, "An Illustration of the Use of Instrumental Variables: Do Neighborhood Conditions Affect a Young Person's Chance of Finishing High School?" Psychological Methods 3, no. 1 [1996]: 249-60; and E. Michael Foster, "Instrumental Variables for Logistic Regression: An Application of the Generalized Method of Moments," Social Science Research 26, no. 4 [1997]: 487-504).
9. Note that we do not assume that these disadvantages cause undesirable outcomes. The issue of the causal impact of each of these disadvantages has been explored in greater detail elsewhere. For a discussion of the causal impact of family structure, see Sara McLanahan and Gary D. Sandefur, Growing Up with a Single Parent: What Hurts, What Helps? (Cambridge, Mass.: Harvard University Press, 1994). Similarly, other studies examine the effect of poverty (Greg J. Duncan and Jeanne Brooks-Gunn, Consequences of Growing Up Poor [New York: Russell Sage Foundation, 1997]); welfare receipt (Jay D. Teachman, Kathleen M. Paasch, Randal D. Day, and Karen P. Carver, "Poverty during Adolescence and Subsequent Educational Achievement," in Consequences of Growing Up Poor, ed. Greg J. Duncan and Jeanne Brooks-Gunn [New York: Russell Sage Foundation, 1997], pp. 382-418); and neighborhood conditions (E. Michael Foster and Sara McLanahan, "An Illustration of the Use of Instrumental Variables: Do Neighborhood Conditions Affect a Young Person's Chance of Finishing High School?" Psychological Methods 3, no. 1 [1996]: 249-60; and E. Michael Foster, "Instrumental Variables for Logistic Regression: An Application of the Generalized Method of Moments," Social Science Research 26, no. 4 [1997]: 487-504).
10. Foster and Furstenberg.
11. Ibid.
12. Ibid.
13. Duncan and Brooks-Gunn, eds.
14. Arnold J. Sameroff, Ronald Seifer, Ralph Barocas, Melvin Zax, and Stanley Greenspan, "Intelligence Quotient Scores of 4-Year-Old Children: Social and Environmental Risk Factors," Pediatrics 79, no. 3 (1987): 343-50.
15. Ibid., p. 348.
16. Arnold J. Sameroff, Ronald Seifer, Alfred Baldwin, and Clara Baldwin, "Stability of Intelligence from Preschool to Adolescence: The Influence of Social and Family Risk Factors," Child Development 64, no. 1 (1993): 80-97.
17. Fong-ruey Liaw and Jeanne Brooks-Gunn, "Cumulative Familial Risks and Low-Birthweight Children's Cognitive and Behavioral Development," Journal of Clinical Child Psychology 23, no. 4 (1994): 360-72.
18. Frank F. Furstenberg, Jr., Thomas D. Cook, Jacquelynne Eccles, Glen H. Elder, and Arnold Sameroff, Managing to Make It: Urban Families and Adolescent Success (Chicago: University of Chicago Press, 1999).
19. Ibid., p. 177.
20. Michael Rutter, "Protective Factors in Children's Responses to Stress and Disadvantage," in Primary Prevention of Psychopathology: Social Competence in Children, ed. M. W. Kent and J. E. Rolf (Hanover, N.H.: University Press of New England, 1979), pp. 382-418.
21. Arnold J. Sameroff, Ronald Seifer, Melvin Zax, and Ralph Barocas, "Early Indicators of Developmental Risk: The Rochester Longitudinal Study," Schizophrenia Bulletin 13, no. 3 (1987): 383-94.
22. See Furstenberg et al. Williams and colleagues also examine the link between mental health and multiple risks. While they find large differences in the prevalence of "pervasive disorder" between children experiencing multiple and none or only one disadvantage, that gap is no larger than that between individuals who do and do not experience particular disadvantages (such as maternal depression). See Sheila Williams, Jessie Anderson, Rob McGee, and Phil A. Silva, "Risk Factors for Behavioral and Emotional Disorder in Preadolescent Children," Journal of the American Academy of Child and Adolescent Psychiatry 29, no. 3 (1990): 413-19.
23. See Furstenberg et al.; Thomas P. Thornberry, D. Huizinga, and Rolf Loeber, "The Prevention of Serious Delinquency and Violence: Implications from the Program of Research on the Causes and Consequences of Delinquency," in Sourcebook on Serious Violent and Chronic Juvenile Offenders, ed. James C. Howell, Barry Krisberg, J. David Hawkins, and J.J. Wilson (Thousand Oaks, Calif.: Sage Publications, 1995); and Michael D. Newcomb, Ebrahim Maddahian, and Peter M. Bentler, "Risk Factors for Drug Use among Adolescents: Concurrent and Longitudinal Analyses," American Journal of Public Health 76, no. 5 (1986): 525-30.
24. See Furstenberg et al.; Thomas P. Thornberry, D. Huizinga, and Rolf Loeber, "The Prevention of Serious Delinquency and Violence: Implications from the Program of Research on the Causes and Consequences of Delinquency," in Sourcebook on Serious Violent and Chronic Juvenile Offenders, ed. James C. Howell, Barry Krisberg, J. David Hawkins, and J.J. Wilson (Thousand Oaks, Calif.: Sage Publications, 1995); and Michael D. Newcomb, Ebrahim Maddahian, and Peter M. Bentler, "Risk Factors for Drug Use among Adolescents: Concurrent and Longitudinal Analyses," American Journal of Public Health 76, no. 5 (1986): 525-30.
25. Survey Research Center, Panel Study of Income Dynamics: 1968-1992 Cross-Year Individual File (Ann Arbor, Mich.: Institute for Social Research, 1993, computer file); Economic Behavior Program, User Guide to the Panel Study of Income Dynamics (Ann Arbor: University of Michigan, Institute for Social Research, Center for Political Studies, Inter-University Consortium for Political and Social Research, 1984); Martha S. Hill, The Panel Study of Income Dynamics: A User's Guide (Newbury Park, Calif.: Sage Publications, 1992).
26. Survey Research Center, Panel Study of Income Dynamics: 1968-1992 Cross-Year Individual File (Ann Arbor, Mich.: Institute for Social Research, 1993, computer file); Economic Behavior Program, User Guide to the Panel Study of Income Dynamics (Ann Arbor: University of Michigan, Institute for Social Research, Center for Political Studies, Inter-University Consortium for Political and Social Research, 1984); Martha S. Hill, The Panel Study of Income Dynamics: A User's Guide (Newbury Park, Calif.: Sage Publications, 1992).
27. Survey Research Center, Panel Study of Income Dynamics: 1968-1992 Cross-Year Individual File (Ann Arbor, Mich.: Institute for Social Research, 1993, computer file); Economic Behavior Program, User Guide to the Panel Study of Income Dynamics (Ann Arbor: University of Michigan, Institute for Social Research, Center for Political Studies, Inter-University Consortium for Political and Social Research, 1984); Martha S. Hill, The Panel Study of Income Dynamics: A User's Guide (Newbury Park, Calif.: Sage Publications, 1992).
28. Respondents are asked whether they receive Aid to Families with Dependent Children as well as whether they receive "other welfare." Researchers have found that including both as "welfare" produces rates of welfare use comparable to other data. We follow that convention here.
29. Terry K. Adams, Documentation for 1968-1985 PSID-Geocode Match Files (Ann Arbor: University of Michigan, Institute for Social Research, Survey Research Center, Economic Behavior Program, 1991), and Documentation for 1990 Census Extract File (Ann Arbor: University of Michigan, Institute for Social Research, Survey Research Center, Economic Behavior Program, 1997). These geocode files are part of the sensitive data files of the Panel Study of Income Dynamics and were obtained under special contractual arrangements designed to protect the anonymity of respondents. These data are not available from us. Persons interested in obtaining PSID sensitive data files should contact Panel Study of Income Dynamics, Box 1248, Ann Arbor, MI 48106-1248.
30. Terry K. Adams, Documentation for 1968-1985 PSID-Geocode Match Files (Ann Arbor: University of Michigan, Institute for Social Research, Survey Research Center, Economic Behavior Program, 1991), and Documentation for 1990 Census Extract File (Ann Arbor: University of Michigan, Institute for Social Research, Survey Research Center, Economic Behavior Program, 1997). These geocode files are part of the sensitive data files of the Panel Study of Income Dynamics and were obtained under special contractual arrangements designed to protect the anonymity of respondents. These data are not available from us. Persons interested in obtaining PSID sensitive data files should contact Panel Study of Income Dynamics, Box 1248, Ann Arbor, MI 48106-1248.
31. In untracted areas, information on the minor civil division (MCD) was used. (The minor civil division is defined as the "primary political/administrative subdivisions of counties"; most MCDs are called "townships." For details, see Adams, Documentation for 1968-1985 PSID-Geocode, p. 26.)
32. We raised the poverty threshold because it appears that families report more of their income in the PSID than they do in government surveys. Raising the poverty threshold brings the poverty rate for the sample into better agreement with official figures. Poverty status was determined using the 1967 poverty thresholds adjusted for inflation using the Consumer Price Index (Bureau of Labor Statistics, http://stats.bls.gov/cpiovrvw.htm, CPI-U-X1).
33. In the 1960s, the Office of Economic Opportunity and the U.S. Census Bureau defined areas with poverty rates above 20 percent as "low income areas." This cut-off was determined based on an index of family income, the proportion of children not living with their parents, the proportion of adults with less than an eighth-grade education, the proportion of male workers who are unskilled, and the prevalence of substandard housing. More recently, William J. Wilson has defined poverty neighborhoods as those with poverty rates above 30 percent. See William J. Wilson, The Truly Disadvantaged: The, Inner City, the Underclass, and Public Policy (Chicago: University of Chicago Press, 1987).
34. Attrition represents an added disadvantage of using panel data to track changes over time. While the overall level of attrition in the PSID since 1968 has been relatively high, it appears that attrition has not damaged the overall representativeness of the data. See John Fitzgerald, Peter Gottschalk, and Robert Moffitt, "An Analysis of Sample Attrition in Panel Data: The Michigan Panel Study of Income Dynamics," Journal of Human Resources 33, no. 2 (1998): 251-99. There are, however, advantages to using a panel to track changes over time. Principal among these is improved (statistical) efficiency. This benefit reflects the fact that the variance of between-period change is a decreasing function of the between-wave correlation in the outcome of interest. In a series of repeated cross sections, that correlation is zero. In panel data, however, that correlation is presumably positive. For more details on the advantages and disadvantages of panel data as a means of tracking change over time, see Greg J. Duncan and Graham Kalton, "Issues of Design and Analysis of Surveys across Time," International Statistical Review 55, no. 1 (1987): 97-117.
35. Attrition represents an added disadvantage of using panel data to track changes over time. While the overall level of attrition in the PSID since 1968 has been relatively high, it appears that attrition has not damaged the overall representativeness of the data. See John Fitzgerald, Peter Gottschalk, and Robert Moffitt, "An Analysis of Sample Attrition in Panel Data: The Michigan Panel Study of Income Dynamics," Journal of Human Resources 33, no. 2 (1998): 251-99. There are, however, advantages to using a panel to track changes over time. Principal among these is improved (statistical) efficiency. This benefit reflects the fact that the variance of between-period change is a decreasing function of the between-wave correlation in the outcome of interest. In a series of repeated cross sections, that correlation is zero. In panel data, however, that correlation is presumably positive. For more details on the advantages and disadvantages of panel data as a means of tracking change over time, see Greg J. Duncan and Graham Kalton, "Issues of Design and Analysis of Surveys across Time," International Statistical Review 55, no. 1 (1987): 97-117.
36. Furstenberg et al. (n. 13 above).
37. Ibid.
38. The difference between these figures and government tabulations reflect differences in the data sources (involving, e.g., differences in income reporting) as well as in the definitions of each disadvantage. For example, the employment figures cited in n. 1 above refer to the employment of all family members; the employment disadvantage used here refers only to the employment status of the household head.
39. 2 statistic with 1 df, the test statistic is 3286.1, 4404.1, 4042.7, and 5637.5 for joblessness, poverty, welfare receipt, and female headship, respectively.
40. As one reviewer noted, observations contributed by the same individual are not statistically independent. Some of the reported test statistics do not reflect this statistical interdependence. However, this omission does not affect any of our findings due to the large sample sizes; we assessed this by performing key analyses a second time using only one observation per child. This had no practical effect on the significance of our findings.
41. Table 1 also reveals that the four disadvantages examined here overlap to some extent but not completely. If the overlap were complete, as one reviewer suggested, all children experiencing one or more disadvantages would experience all four. This is clearly not the case. For both blacks and whites and for all periods, only a minority of children experiencing any disadvantage experience all four. Another reviewer suggested defining poverty status using a three-year window. This would strengthen the link between poverty status and the other disadvantages and between poverty status and race. I examine these issues from a longitudinal perspective in Foster (1999).
42. The shift over time in the distribution of disadvantages is significant for both whites and African Americans (p < .01).
43. While not apparent in fig. 2, trends in the racial gap in the poverty rate reflect reductions in the poverty rate for both blacks and whites. (The rate for blacks fell less than that for whites, broadening the gap slightly.) Current Population Survey figures, in contrast, show an increase over time in the poverty rate for white children, indicating an apparent convergence between black and white children. See Lamison-White (n. 1 above). There are two explanations for the differences in the time trend in the white poverty rate between data sources. The first is that, as discussed above, the PSID underrepresents Hispanic children. In contrast, the CPS data include these children. Since Hispanic children can be of any race, they are included in the CPS tabulations for both blacks and whites (predominantly the latter). When we approximate what the CPS figures for whites would look like without Hispanics, we find little change in the childhood poverty rate over time. A second explanation involves the use of the Consumer Price Index to adjust poverty thresholds for inflation. The CPS figures are based on poverty thresholds that are updated using the standard index, the Consumer Price Index for Urban Consumers (CPI-U). Our tabulations rely on an alternative price index, the CPI-U-X1. The CPI-U-X1 is generally considered superior because it handles changes in the cost of housing over time more sensibly. The CPI-U is believed to overstate inflation. If this is true, then the poverty thresholds would be overinflated, resulting in poverty rates that are too high and reductions in poverty rates that are too low.
44. These supplemental tabulations are available from E. Michael Foster.
45. Using bootstrapping, we determined that the change in the white-black ratio of the percentage of children living in high-poverty neighborhoods between the first and fifth periods is statistically significant at p < .05.
46. The trend over time in the proportion of black children who are most disadvantaged and live in poor neighborhoods is statistically significant at conventional levels (p < .01). (Statistical significance was determined using a weighted logistic regression that included a linear time trend.)
47. Council of Economic Advisors, Economic Report of the President Transmitted to Congress February 1997 (Washington, D.C.: Government Printing Office, 1997).
48. Council of Economic Advisors, Economic Report to the President Transmitted to Congress February 1999 (Washington, D.C.: Government Printing Office, 1999).
49. We obtained these figures by averaging the March unemployment rate across the years in each period. (Unemployment data were taken from the Current Population Survey [http://stats.bls.gov/top20.htm/]. We selected March because PSID interviews generally occur in the spring.)
50. Council of Economic Advisors, Economic Report to the President Transmitted to Congress February 1999.