That students’ social and economic characteristics shape their cognitive and behavioral outcomes is well established, yet policymakers typically resist accepting that non-school disadvantages necessarily depress outcomes. Rather, they look to better schools and teachers to close achievement gaps, and consistently come up short.
This report describes how social class characteristics plausibly depress achievement and suggests policies to address them. It focuses on five characteristics for purposes of illustration:
- parenting practices that impede children’s intellectual and behavioral development
- single parenthood
- parents’ irregular work schedules
- inadequate access to primary and preventive health care
- exposure to and absorption of lead in the blood.
These are not the only characteristics that depress outcomes, nor are they necessarily the most important. This report makes no judgment about the relative importance of the many adverse influences on child and youth development. Parental unemployment and low wages, housing instability, concentration of disadvantage in segregated neighborhoods, stress, malnutrition, and health problems like asthma are among other harmful characteristics.
Certainly, some children with severe socioeconomic disadvantages achieve at higher levels than typical children without them; a range of outcomes is associated with every characteristic, and descriptions of the impacts of social class characteristics only describe averages, not the performance of any particular child. Likewise, this report does not imply that all lower-social-class families have each of these characteristics. But all have many of them.
Because characteristics of lower-class status overlap and may be interdependent, available data do not permit the isolation of any one. Econometric studies that identify the effect of a particular characteristic by holding others constant are valuable, but no study controls for all, and few control for very many.
For each characteristic reviewed here, this report describes its average incidence by race (black versus white) and socioeconomic status. Data limitations preclude similar descriptions of Hispanics’ characteristics. Where research is available, we then review what is known about the characteristic’s prediction of cognitive (academic performance or IQ, for example) and non-cognitive (behavioral) outcomes. We next review the “plausible pathways” by which the characteristic influences youths’ outcomes—i.e., how these predictions might reflect causality. We conclude by recommending policies to reduce the intensity of these specific disadvantages.
This report’s key findings are as follows:
- Parenting practices that impede children’s intellectual and behavioral development: Lower-social-class parents engage in fewer educationally supportive activities with young children, such as reading aloud or playing cognitively stimulating games. Lower-social-class parents also exert more direct authority and offer children fewer choices in their daily interactions, leaving them less prepared for “critical thinking” when school curricula expect it. Parents’ failure to engage in educationally supportive activities is associated with children’s poorer academic and behavioral outcomes. There are well-validated programs that can offset these effects. High-quality early childhood care and education centers provide intellectually stimulating environments that disadvantaged children may miss at home. Nurse home-visiting services assist disadvantaged mothers with health problems and teach developmentally appropriate parenting skills. High-quality after-school and summer programs that offer cultural and organizational activities are typically attended by middle-class youth, not students from lower-social-class backgrounds.
- Single parenthood: Mothers raising children alone are more likely to be low-income, African American, and less educated. Their children typically have lower test scores, are more likely to drop out of school, and have greater emotional and behavioral difficulties (more delinquency and violence, more school dropout, more suicide). Sex education and school-based health centers that provide long-lasting contraception to teenage girls are important, but they will not be as effective as they have to be if African American men remain poor marriage partners—unable to help support families because of excessive unemployment and discriminatory arrest and incarceration. Full employment as well as labor market and criminal justice reforms that enable fathers to earn middle-class incomes are needed to improve children’s outcomes.
- Parents’ irregular work schedules: Computerized scheduling and the weakening of norms governing employers’ responsibility for employee welfare have combined to produce irregular work schedules for many hourly paid low-wage workers, disproportionately African Americans and the less educated. Unpredictable schedules make it difficult, if not impossible, to place children in high-quality child care centers and to establish regular home routines in which children can thrive. Children of mothers with non-standard schedules have worse verbal and other cognitive skills, mental health, and behavior. New regulatory policies—for example, requiring call-in pay for workers sent home before shifts end—could create incentives for employers to reduce use of “just-in-time” employee scheduling.
- Inadequate access to primary and preventive health care: Minority children and those whose parents are less educated or who live in low-income neighborhoods are less likely to have personal physicians or nurse practitioners, or receive necessary referrals to specialists. No research directly associates physician access with children’s cognitive or non-cognitive outcomes, but a relationship is easy to intuit. Children with limited access are more likely to have routine and preventable illnesses, causing more frequent absences from school. Regulatory changes that support school-based health centers and Medicaid reimbursement changes to create incentives for primary care physicians to locate in low-access neighborhoods could address this.
- Exposure to and absorption of lead in the blood: Children with high blood lead levels are disproportionately low income and African American. Lead reduces cognitive ability (IQ) and causes adverse behavioral outcomes, such as increased violence and other criminal behavior in adolescents and young adults. Although lead was removed from gasoline in the 1970s and 1980s, lead remains on the ground and is frequently stirred up into breathable air. Lead also remains in windows, window frames, the walls of older buildings, and pipes carrying water to residences. Lead cleanup is expensive, but it would result in substantial overall savings in reduced special education placements, reduced criminal behavior, and greater worker productivity from adults with greater cognitive ability.
Policymakers are perplexed about addressing the impact of racial and socioeconomic class differences on student outcomes. While they generally understand that family and community characteristics affect performance, they also fear that acknowledging this fact means we should tolerate lower standards for disadvantaged children, something they consider morally and politically unacceptable. As a result, contemporary education reform efforts focus disproportionately on school and teacher incentives and do little to narrow achievement gaps.
Economists, sociologists, and developmental psychologists have consistently concluded that background characteristics strongly shape cognitive and behavioral outcomes.1 When school improvement is not complemented by policies to narrow social class differences, students’ chances of success are greatly diminished.
Thus, educators should, as educators, be vocal advocates for policies that reduce poverty and address other characteristics of lower-class status.aBut educators can do more, seizing opportunities to coordinate school improvement with community services that ameliorate socioeconomic disadvantage—services such as early childhood care and education, nurse home-visiting programs, after-school and summer opportunities, school-based health centers, and sponsorship of community lead cleanup. Such services cannot substitute for macroeconomic policies like full employment, higher wages, and stable work schedules, all of which help parents nurture and support their children, but the limitations of school-based social and economic improvement programs should not discourage educators from pursuing them.
Highlighting the socioeconomic impediments to student achievement does not make “excuses” for the achievement gap, as some advocates glibly, and defensively, charge.2 Rather, it provides explanations. Although some educators may use student poverty as an excuse for inadequate performance, the conscientious understand that without good explanations for low achievement, policies to address it are unlikely.
Of the many social class characteristics known to depress outcomes, this report deals with five: challenged home intellectual environments, single parenthood, irregular parental work schedules, inadequate health care access, and exposure to environmental lead. These factors were chosen because recent research has offered important new insights regarding each. But these are not the only important characteristics depressing outcomes, nor is there a research basis for determining with any certainty whether they are necessarily the most important. Therefore, this report can make no judgment about the relative importance of such adverse influences on children, although we can assert with confidence, based on available research, that each is quite important. Other characteristics, all associated with poverty—including parental unemployment and low wages, housing instability, concentrated disadvantage in segregated neighborhoods, stress, malnutrition, and health problems like asthma—deserve similar treatment. We will consider these in future work.
Such characteristics interact, so precise estimates are not possible for the shares of achievement gaps attributable to specific social class differences, just as they are not possible for the shares attributable to teacher performance or other school qualities. For example, this report describes how poor parental literacy and irregular work schedules each varies by race and social class and affects children’s outcomes. But they co-vary—simply adding effects together exaggerates estimates of harm because poor literacy and irregular work are themselves correlated; research has not established the additional impact of children’s suffering from the results of both, rather than one factor.
We also underestimate harm by considering disadvantage only for individual children. When disadvantaged children are concentrated in classrooms and schools, their difficulties are exacerbated. Children in racially isolated and predominantly low-income schools have few peers who legitimize higher achievement standards. Their teachers must focus more on remediation and discipline, without time to devote to children whose problems they could address only if those needing special attention were few.
Because characteristics of lower-class status overlap and may well be interdependent, available data do not permit the isolation of any one. Econometric studies that identify the effect of a particular characteristic by holding others constant are valuable. But no study controls for all, and few control for very many. Correlations between specific socioeconomic barriers and child outcomes may, in many cases, seem small. But while we cannot estimate the precise contribution of each disadvantage to achievement gaps, influences not attributable to schools are so numerous that policy should consider how to address them.
No influence is fully determinative. Some children with less literate parents excel beyond the typical performance of college graduates’ children. Each adverse socioeconomic influence has a wide range of outcomes but, on average, exposed children will perform less adequately.
For the characteristics reviewed below, where recent research provides reliable information, we discuss:
- average race and social class differences in incidence b
- prediction of cognitive outcomes
- prediction of non-cognitive outcomes
- the plausible pathways by which these associations may reflect causation.c
We conclude by reviewing practical reforms to narrow differences in these characteristics that would, in turn, likely help narrow outcome gaps. Ultimately, to make substantial progress in narrowing gaps, school improvement should be complemented by and coordinated with amelioration of socioeconomic disadvantages.
Parenting practices that impede children’s intellectual and behavioral development
How parents shape home environments affects children’s outcomes.3
Race and social class differences
The Department of Education’s Early Childhood Longitudinal Study, Kindergarten Class of 2010–2011 (ECLS-K: 2011) asked a nationally representative sample of entering kindergartners’ parents about the number of books in their homes, a measure considered a reliable indicator of home intellectual environment.4 On average, white parents reported 112 books, black parents 44.5
ECLS-K: 2011 also surveyed parents about literacy activities—reading aloud, telling stories, doing art—conducted with their entering kindergartners. White parents consistently reported greater frequency than black parents. However, survey questions were poorly framed, encouraging social acceptability bias (the tendency of interview subjects to exaggerate their responses in accordance with what they understand is considered desirable). A more reliable measure than the ECLS-K on this score is the Bureau of Labor Statistics’ American Time Use Survey (ATUS), which also asks adults about educationally supportive activities.dTable 1 compares white to black adult reports.
White adults spend 36 percent more time than black adults reading to young children, and three times more time talking with and listening to them. Other analyses find that black mothers are about two-thirds as likely as white mothers to read to toddlers daily.7
ECLS-K reports responses by socioeconomic status (SES), using an SES definition including parents’ income, education, and occupational prestige. Such definitions are useful but limited, missing other important social class characteristics. For example, in 2010, the ratio of black to white median family income was 56 percent, while the ratio of black to white median family wealth was 5 percent, owing largely to 20th century housing policy that barred African Americans from purchasing suburban homes that later appreciated in value, a primary source of wealth for households.8
The ECLS-K definition of SES also fails to capture other important social class characteristics that are important for understanding differences in child outcomes—characteristics such as single parenthood, the overall economic circumstances of one’s neighborhood, or other factors. Patrick Sharkey, for example, has shown that the quality of the neighborhood where a child’s mother was raised has a bigger influence on the child’s achievement than the quality of neighborhood where the child was raised.9 Overlooking differences like these leads to reports that black children underperform seemingly similar white children who are actually more advantaged. Thus, we must report data using such definitions with great caution. Mindful of this qualification, Table 2 describes parent reports of books at home by SES quintile.
For each quintile, parents report more books than the next lower quintile, and whites report twice as many books as blacks.
ATUS finds that middle- and upper-class adults (those in the top three weekly earnings quintiles) report 66 percent more time reading books to young children than lower-class adults (those in the bottom two quintiles). Other research finds that parents on public assistance, unemployed, or with less than a high school education typically provide less cognitive stimulation to children.11 When reading aloud, lower-class parents provide less guidance and are less strategic in building on children’s prior knowledge to expand it.12
By age 6, white children have typically spent 1,300 more hours engaged in conversations with adults than black children. Six-year-olds from affluent families have spent 1,300 more hours in indoor and outdoor recreation, churches, businesses, and other non-school, non-home, and non-caretaker settings than children from low-income families. Differences are greater still (1,800 hours) between children of parents with less than a high school education and children of college graduates.13 This gives children of high-income and highly educated families more background knowledge, the most important predictor of later academic achievement.14
How parents shape children’s choice-making, self-direction, and stances toward authority varies by social class. Middle-class parents typically give fewer direct orders, instead providing controlled choices. Lower-class parents expect more deference to authority. Lower-class children typically have more unstructured leisure time where they need not follow adult rules, while middle-class children typically have more structured schedules.15
Middle-class children who are used to operating in controlled environments, similar to classrooms, may be more likely to thrive in school compared with children who are used to unstructured time on the one hand and disciplinary action on the other.
Association with cognitive outcomes
Math and reading skills of entering kindergartners in the top and bottom SES quintiles differ by about 40 percentile points in normal distributions.16 Cognitive gaps do not change much from kindergarten to middle school. This does not mean that schools are ineffective with lower SES children; rather, lower SES gains resemble those of higher quintiles, and initial gaps are left mostly unchanged.17
Low-income parents of children in Head Start who spend more time reading to their children, visit the library more often, keep more children’s books in the home, and begin reading to their children at an earlier age have children with higher literacy skills. These children are more ready to read when they reach school age, have better vocabularies, are better able to identify words and letters, and know more story and print concepts—the title of a book, the author, reading from left to right, understanding characters’ feelings.18 Toddlers of low-income mothers who read to them daily have better vocabulary and comprehension at 24 months.19 Five-year-olds have poorer language and math skills if, when they were two years old, their parents were less educationally supportive—engaging in less cognitive stimulation, being less sensitive to children’s perspectives, and demonstrating less love, respect, and admiration toward their children—when doing activities like puzzles.20
Parents who teach about expectations for schooling have children with better school performance.21
Association with non-cognitive outcomes
Non-cognitive skills of entering kindergartners in the top and bottom SES quintiles differ by about 10 to 23 percentile points in normal behavioral distributions.22 These gaps do not disappear, forecasting adult differences. Children from the lowest quintile have adult arrest rates 15 percentage points higher and high school completion rates 31 points lower than highest-quintile children.23
Parents who are more involved in their children’s educations by volunteering outside the classroom, helping their children with homework, and checking their children’s homework have children with fewer behavioral problems in the classroom.24
The availability of children’s books and whether mothers read aloud, share meals with their children, use non-harsh discipline, expect their children to help keep their homes clean, are affectionate with their children, and encourage children to contribute to conversation all predict better social skills and fewer teacher-reported behavioral problems.25
Parents with more education have greater educational expectations and can convey enjoyment of learning. Children internalize these, leading to higher achievement.26
Parents with less education have fewer educational interactions with their children at home. Among low-income African American mothers, those who are less educated tend to provide less assistance and be less supportive and encouraging of their children during home-based teachable tasks like puzzle-making.27 Because African American mothers have lower average education levels than white mothers, this finding can likely explain an overall racial difference in these parenting behaviors.
Middle-class children with more choice-making opportunities in daily interactions with parents, as well as with more participation in structured leisure activities (e.g., music lessons, organized sports), may be more comfortable navigating environments similar to classrooms than lower-class children more accustomed to following direct instructions and having more unstructured time.28 This difference may be especially advantageous for middle-class children after elementary school, when making intellectual choices (“critical thinking”) becomes more important in school curricula.
Children raised by single parents have lower average outcomes than children raised by two parents.
Race and social class differences
The share of children living with a single mother varies by children’s race and social class.
Table 3 shows shares of children, by race, living with single mothers. Some may have lived with mother-alone only briefly (because of divorce, separation, death of spouse, or absence of spouse, for example), so also shown are shares of children living with never-married mothers.
Although black children are more likely to be living with the mother alone than white children, the share of both white and black children in single-parent homes has grown, partly because falling real wages have made it more challenging for women to find marriage partners who earn sufficient incomes to support families.30 The greater rates of unemployment and incarceration and the lower wages for young black than for young white men help explain racial differences in single parenthood. Table 3 shows that black children have never-married mothers at nearly five times the rate of white children. Although about one in four children lived with a mother alone in 2013, twice as many had lived with a mother alone at some time during childhood.31
Table 4 shows average parental education by family structure.
A child whose parent has only a high school education is twice as likely to be living with a never-married mother than with two parents (35 versus 18 percent), whereas a child with a college-educated parent is more than five times as likely to be living with two parents than with a never-married mother (48 versus 9 percent).
Table 5 compares the economic circumstances of various family structures.
Children living with never-married mothers are four times as likely to be in low-income families as children living with both parents. The poverty rate (not shown in the table) for children living with a never-married mother only is 53 percent, with a mother only 45 percent, and with both parents 13 percent. A child living with a never-married mother alone is therefore four times as likely to be in poverty as a child living with both parents.34
In 2003, average single-mother household income was 37 percent of married household (with children) income; even after public assistance of various kinds was included, the ratio was still only 55 percent. The disparities are likely greater now, after the weak recovery for lower-income families from the recession, than they were in 2003. Another estimate found that in the year following a divorce, a single-mother household suffers a 40 percent income loss, with little change in subsequent years. And single-parent households have, on average, substantially less wealth than two-parent households.35
In 1999, 10 percent of all births were births to unmarried teenagers, many of whom then had additional children as unmarried adults. These children are still of school age today, the oldest in high school, with their siblings in the lower grades. By 2014, the share of births that were births to unmarried teenagers had dropped to 6 percent, and it is reasonable to assume that these teenagers too will have some subsequent children as unmarried adults. These children will be of school age in the years to come. Although fewer teens of all social classes now become mothers, those who do are more likely to be school dropouts, and their children are more likely to suffer from abuse and neglect.36
Association with cognitive outcomes
Children of single parents, especially teenage single parents, are less likely to graduate from high school or college.37 Thirty percent of adolescents in single-parent households repeat a grade, compared with 19 percent of other children.38 Children of single parents have lower test scores.39
Association with non-cognitive outcomes
Children of single parents also have worse behavioral outcomes.40Table 6 summarizes results from a nationally representative 1995 survey.
On each measure, adolescents in single-parent households reported greater emotional and behavioral difficulty, including higher rates of delinquency and illicit drug use. Girls raised in single-parent households are more likely to give birth themselves as single mothers.42
If a teenage mother has a high school diploma, the average annual incomes of her children when they reach young adulthood will be more than 10 percent greater, on average, than those of children of a teenage mother without a diploma. For teenagers, delaying a first birth improves their children’s circumstances as adults in terms of annual incomes, likelihood of depression, and chances of single parenthood.43
Table 7 shows outcome differences at five life stages for children of never-married, divorced, and married parents.
At each stage, children of never-married parents have worse outcomes than children whose parents were married during some of their childhoods and those whose parents were married throughout their childhoods.
Single parents are less able to provide resources for children—high-quality child care, books, computers, and consumer goods (such as cell phones, shoes, and clothes) that give children status with peers—and less able to offer structure, conversation, and time.46 The importance of lower income in explaining the adverse effects of single parenthood is suggested by data showing that children of single mothers do better when fathers pay child support than when fathers do not—although fathers who pay child support may also be involved with their children in other ways, also contributing to better outcomes.47
Single parents’ time is also scarcer. Teenagers, especially boys, in single-parent households are more likely to be unsupervised after school.48
Single mothers are also more likely to smoke cigarettes and use illicit drugs during pregnancy, receive inadequate prenatal care, and have low-birthweight babies, a well-established predictor of poor cognitive and behavioral outcomes.49 A national survey, the Fragile Families and Child Wellbeing Study (FFCWS) of the late 1990s, found nearly 8 percent of unwed mothers drinking heavily, four times the married mothers’ rate.50
Single parents’ lower incomes also entail greater likelihood of living in poor neighborhoods with more disorder and crime that cause children stress.51 Single parents themselves are more stressed.52 Mothers with greater stress are less emotionally supportive of children and employ less consistent and harsher discipline. Single mothers are more depressed and more likely to abuse children, causing worse outcomes for children themselves.53
Frequent home moves also cause childhood stress, from loss of friends and from instructional discontinuity. Single-parent families move more, partly because mothers may form new relationships that entail moving in with new partners.54 The adverse consequences of moving are especially severe for boys because moving often results in reduced (or loss of) contact with fathers.55
Family instability also predicts poor outcomes for children who may not form healthy child-parent relationships with their mothers’ subsequent partners. During the five years of the FFCWS, one-quarter of unmarried mothers lived with a new partner, and one-fifth had a child with a new partner. Half of children’s biological fathers saw their children less than once a month, making effective parenting much less likely.56
The FFCWS also found that children in single-parent families are in consistently worse health—more asthma, obesity, accidents, or injuries—likely contributing to greater school absenteeism, which leads to lower academic performance.57
Parent’s work schedules
Employers in the retail and service sectors now have computer technology that predicts customer and supplier traffic levels.58 This technology, in combination with a weakening of the implicit contract governing relationship norms between employers and employees, has encouraged supervisors to create “just in time” work schedules, in which workers can be called to work or sent home on short notice, based on predictions of customer demand or supply delivery times.59 If a delivery truck’s arrival time is expedited or delayed, workers may be called in on short notice to unload it, without regard to these workers’ previously established schedules. Or if customer traffic patterns at 4 p.m. have previously predicted retail or restaurant patterns at 6 p.m., employers have the ability on short notice to send employees home, call them in, or hold them beyond the end of their scheduled shifts, without regard to previously posted schedules.
Race and social class differences
Table 8 shows racial differences in non-standard work, i.e., work in which most hours do not fall between 8 a.m. and 4 p.m., shifts rotate, or schedules vary weekly or otherwise. The table describes workers at age 39, a typical childrearing year. White parents are only two-thirds as likely to be assigned non-daytime shifts as minority parents.
Table 9 shows that less-educated workers are more likely to have non-standard schedules. College-educated parents have non-daytime shifts one-third as often as parents with high school degrees or less.
Mothers with non-standard schedules are more likely to be low-income and younger and to have spent more years as single parents.62 But not all parents with non-standard schedules are low-income: Some work non-standard schedules by choice—for example, to ensure that one of two parents is always available to care for children. Such families tend to be more affluent, older, and married, and tend to include mothers who are better educated, than families where parents (and especially single mothers) work nights or evenings.63 Lower-income parents who work such shifts may be required to do so by their employers, or these parents may find it necessary to assemble multiple part-time jobs to earn full-time incomes.
Table 10 shows shares of hourly paid workers, particularly working parents, who receive little advance notice of weekly hours, making earnings unpredictable. Of hourly workers, about half of African Americans, half of low-wage workers, and one-third of mothers with pre-teen children get one week or less notice of weekly schedules. Most have substantial fluctuation in weekly hours. Sixty-nine percent of hourly paid mothers of pre-teens report weekly fluctuations in work hours.
Association with cognitive outcomes
Children whose parents work non-standard schedules demonstrate poorer cognitive performance. Comparing 15-month-olds of otherwise observably similar mothers (similar in age, cognitive capacity, educational level, marital status, family size, level of depression, and both recent and long-term poverty status), the toddlers whose mothers have non-standard work demonstrate worse perception, memory, learning, problem solving, and verbal communication.66 At 36 months, they have worse verbal comprehension and have a harder time naming objects.67 Patterns established this early in life are difficult to reverse, and have a strong influence on adolescent and adult outcomes.
Association with non-cognitive outcomes
Children and adolescents whose parents work non-standard schedules have worse mental health and behavior. Preschoolers whose mothers work non-standard schedules lose from 10 to 12 percentile points in a normal distribution of preschoolers’ negative behavior (e.g., depression, anxiety, withdrawal, aggression).68
Each additional nighttime hour that low-income African American mothers work is associated with a decrease in their preschoolers’ positive behavior (e.g., being “playful”); the preschoolers of mothers working full eight-hour night shifts lose 15 percentile points in a normal behavioral distribution.69
Teachers of schoolchildren whose parents work variable schedules rate these children as less engaged, more aggressive, and impulsive.70
Children age 13 and 14 whose mothers and/or fathers work night shifts are more likely to engage in risky behavior (e.g., smoking, consuming alcohol, delinquency, sex), and are more likely to be depressed. The negative outcomes were apparently set in motion when parents had worked non-standard schedules earlier in these adolescents’ lives. The number of years fathers work nights before children’s fifth birthdays predicts increased sexual activity for the children when they reach adolescence. The 13- and 14-year-olds are more likely to drink alcohol if, when they were between 5 and 10 years old, their mothers worked night shifts.71
Children and adolescents whose parents work non-standard hours have more physical health problems that, in turn, are associated with poorer academic outcomes.
Children with parents who work non-standard hours are heavier than those whose parents work regular schedules. Lower-middle-class children (those whose families are in the second income quartile) whose mothers have worked non-standard shifts for from one to four years have close to twice the odds of being overweight at age 13 or 14 as children from economically similar families whose mothers do not work non-standard shifts.72
It is plausible that parents’ non-standard working hours, independent of other characteristics, would inhibit children’s cognitive and behavioral outcomes.
Mothers with non-standard schedules must make last-minute child care arrangements with friends or relatives; many cannot enroll children in high-quality centers that require predictable drop-off and pick-up times. Many states scale child care subsidies to the number of hours parents work, so parents working irregular and variable schedules are at heightened risk of losing eligibility for subsidies and, when they do, can no longer afford to place their children in formal centers.73
Parents with non-standard schedules find it more difficult to spend time with children and engage in cognitively stimulating activities with them.74 For example, for low-income African American mothers of preschool children, each additional nighttime hour of work is associated with a decrease in cognitively stimulating mother-child activities of about 1.5 percentile points in a normal distribution of mothers’ engagement in such activities. Thus, mothers who work a full eight-hour night shift decrease their engagement in cognitively stimulating mother-child activities by about 21 percentile points in such a distribution.75 Parents who work non-standard schedules are less able to spend time with their children, take care of their homes, have meals with their children, and, particularly in the case of fathers, be close to their children. Parents working nights can’t supervise children’s critical after-school time.76 Parents with non-standard hours are more tired, anxious, irritable, and stressed, making children’s delinquency, aggression, and other negative behavior more likely.77
Parents’ variable schedules require irregular family mealtimes and bedtimes for children that interfere with their healthy development.78
Because parents, especially single parents, with variable schedules cannot easily schedule doctor appointments, their children likely receive less non-emergency, routine, and preventive care.79 They may then have marginally worse health and more school absenteeism, also harming their development.
Parents with variable part-time schedules cannot easily secure second jobs to support their children’s welfare. Variable schedules prevent parents from enrolling in school themselves, an activity that might enable them to provide better intellectual environments and models of educational aspiration for their children.80
In general, the harmful effects of working non-standard schedules seem to be more pronounced for children of parents who work night shifts (starting after 9 p.m.) than for children of parents who work other non-standard schedules (evening shifts beginning after 2 p.m., rotating or variable shifts). At first glance this empirical result seems to be counterintuitive, because supervision of adolescents is most necessary in the after-school and evening hours. There are two plausible explanations for the excess harm attributable to night shifts. One is that night work has more of an effect on parents’ moods, making them more anxious and irritable than parents who work evening shifts, and this reduces the quality of parent-child time, even when a parent is present. The other seems to be that data showing a greater harmful effect from night than from evening or rotating shift work results from the experience of two-parent families where parents organize their schedules so that when one is working, the other is available to supervise children. This option is more available to parents who are married, are older, have higher maternal education, and have higher family income. It is in single-mother families that shift work is most strongly associated with delinquent behavior.81
Access to physicians
Health differences exacerbate social class outcome gaps. Differences in access to primary care physicians are a factor.82
Race and social class differences
Table 11 describes differences by race, maternal education, neighborhood safety, and family income in whether children have personal physicians or nurse practitioners, and whether children can easily get necessary referrals to specialists.
Although public programs such as Medicaid and the Children’s Health Insurance Program (CHIP) have now extended coverage to most low-income children, they have not fully equalized access to primary and preventive care. Restricted access impedes timely treatment of conditions like skin allergies, asthma, and dental problems, more prevalent (partly from less access) for black than white children.84 For each race or social class category surveyed, disadvantaged children have less access to quality health care. For each comparison, differences in having personal physicians or nurses are small, but the differences are so consistent across all comparisons that real problems seem evident.
A 2006 survey found that in higher-income communities there were fewer than 1,000 children per physician, and 86 percent of doctors were board certified. But in more middle-class neighborhoods there were over 3,000 children per physician, and only 76 percent were board certified.85 Physician supply in low-income neighborhoods was likely even more restricted.86
Parents seeking appointments with specialists for their children are refused two-thirds of the time with public insurance but only 11 percent of the time with private insurance.87
Available data do not directly associate physician access with children’s educational outcomes. But it seems apparent that children with limited access are more likely to be sick and absent from school. A 2011 survey of public preschools in Chicago found that African American 4-year-olds missed 7 percent of school days from illness, while whites missed 3 percent. Although the white and black children came from economically similar families, the black children came from poorer neighborhoods.88
Differences in treatable illnesses may contribute to differences in performance even when children are present in school. For example, asthmatic children who do not have inhalers are more likely to be awake at night, and come to school more drowsy and inattentive.
Public health authorities measure children’s lead absorption by micrograms of lead per blood deciliter. In 2000, approximately 4 percent of children tested for lead had very dangerous levels of 10 micrograms or higher, but by 2013 less than 1 percent had levels that high.89
Previously, a major source was exhaust from automobiles, aircraft fumes, and industrial emissions.90 In 1973, the Clean Air Act required phase-out of leaded gasoline. By 1990 lead was almost entirely eliminated from gasoline, and industrial and aircraft emissions continued to decline, although at a slower rate than in the 1975–1990 period. By 2010, lead-in-air levels had declined to only 3 percent of their 1980 level.91 Average blood levels dropped from 16 micrograms per deciliter in 1976 to 3 in 1991. About half of the decline in lead-in-blood levels in this period was probably attributable to banning leaded gasoline. Removal of solder from food cans also likely contributed.92
But lead once used in gasoline remains on the ground and is kicked into the air when ground dirt is disturbed. Food grown in contaminated soil continues to carry lead. However, the most important remaining source of lead today is lead-based paints in homes built before 1978, particularly those built before 1950, and lead pipes that carry drinking water from municipal water supplies.93
In recent decades, the lead-in-blood level considered dangerous was reduced from an earlier standard of 10 micrograms per deciliter, and in 2003 the Centers for Disease Control and Prevention (CDC) determined that there is no safe blood lead level—any exposure is harmful.94 The danger is greatest for children younger than seven, and effects of this early exposure persist throughout life.95 Of the many social, economic, and environmental conditions that influence youth performance, the relationship between lead and negative outcomes is one of the most firmly established, partly because so many studies have consistent findings, and partly because the rates at which states required the elimination of leaded gasoline differed from 1975 to 1985, creating a natural experiment that reasonably well isolated the role of lead in causing cognitive and behavioral changes.96
Race and social class differences
Blacks remain about twice as likely as whites to have levels greater than a dangerous 5 micrograms of lead per blood deciliter.97 Most American children have levels of 1 or 2, but about half a million, mostly living in urban neighborhoods, have levels above 5.98
Table 12 reports results of a CDC survey measuring lead-in-blood levels of children from 1 to 5 years of age.
Home intellectual environment, households with children 5 years of age and younger, 2013
|Adult (parent) activity the day prior to survey||Ratio of white adult minutes to black adult minutes in the activity|
|Read books to child(ren)||1.36|
|Played with child(ren), not sports||3.20|
|Did arts and crafts with child(ren)||*|
|Talked with/listened to child(ren)||3.25|
* Ratio is not calculable because there were reported minutes for white adults but not for black adults.
Source: BLS (n.d.)
Books at home, entering kindergartners, fall 2010, by race and socioeconomic status
Number of books at home reported by parents
|Low SES (quintile 1)||Low-middle SES (quintile 2)||Middle SES (quintile 3)||High-middle SES (quintile 4)||High SES (quintile 5)||Average|
|Ratio, white to black||2.1||2.2||2.0||2.3||2.1||2.5|
Percent of children living with mother alone, by race, 2013
|Children under 18 living with mother only||24%||51%||18%|
|Children under 18 living with never-married mother only||11||34||7|
Source: Census (2013)
Percent of children living in various family structures, by parent’s education, 2013
|Children under 18||Less than H.S. graduate||H.S. graduate only||Some college or associate degree||Bachelor’s degree or more|
|Living with both parents+||7||18||26||48|
|Living with mother only||18||30||37||16|
|Living with never-married mother only||22||35||34||9|
* Does not include children in households where no parent lives.
+ Shows attainment of more highly educated parent.
Source: Census (2013)
Percent of children living in various family structures, by family income, 2013
|Children under 18||Lower (below $30,000)||Lower-middle (from $30,000 to $50,000)||Middle (from $50,000 to $75,000)||Upper-middle (from $75,000 to $100,000)||Upper (over $100,000)|
|Living with both parents||16||15||18||16||36|
|Living with mother only||57||21||12||5||5|
|Living with never-married mother only||65||19||9||4||4|
* Census does not report data corresponding to actual income quintiles. In 2013, the actual upper limits of the bottom four income quintiles were approximately $28,600, $51,400, $79,300, and $121,700.
** Does not include children in households where no parent lives.
Source: Census (2013)
Adolescent reports of problem behavior, by family type, 1995
|Adolescents living in households with:|
|One parent||Two parents|
|Suspended from school||40%||21%|
Source: Amato (2005), Table 1 (p. 86)
Outcomes at five life stages for children of never-married, divorced, and married parents (percent successful)
|Early childhood||Middle childhood||Adolescence||Young adulthood||Adulthood|
|Children of parents who are:|
Source: Grannis and Sawhill (2013);Howard and Reeves (2014)
Definitions of success:
Early childhood: acceptable pre-literacy, math, and behavioral skills
Middle childhood: acceptable reading, math, and social-emotional skills
Adolescence: high school graduation with GPA ≥ 2.5, no criminal conviction, no parenthood
Young adulthood: lives independently and has college degree or income ≥ 250% of poverty
Adulthood: family income ≥ 300% of poverty
Percentage of workers (age 39) who work non-standard schedules, by race
|Working variable, non-daytime, or other non-standard schedule||24%||20%|
|Working non-daytime schedule||17||11|
* Mostly non-Hispanic whites, but also includes Asians, and others
Source: Presser and Ward (2011),Chart 3.
Share of workers (age 39) who work non-standard schedules, by educational attainment (percent of attainment group)
|Less than H.S.||H.S.||Some college||College or more|
|Working variable, non-daytime, or other non-standard schedule||23%||24%||18%||19%|
|Working non-daytime schedule||17||15||10||5|
Source: Presser and Ward (2011), charts 4, 5
Advance notice given to hourly employees, childrearing years (age 26–32)
|Notice of schedule (percent receiving notice of):|
|≤ 1 week||≤ 2 weeks||≤ 4 weeks||Share of hourly workers whose weekly hours vary (percent)|
|All hourly employees||41%||55%||61%||74%|
|Mothers of children < 13 years of age||32||44||51||69|
|Women in part-time jobs||41||58||68||81|
|Low-wage (<$15/hr) part-time workers||49||67||76||83|
|Janitors and housekeepers||40||54||60||66|
|Food service workers||64||81||84||90|
|Home care workers||55||65||67||71|
Source: Lambert, Fugiel, and Henly (2014), tables 3, 5, 12, 13; Fugiel (2015)
Access to physicians, children under age 18 (2007)
|Have a personal physician or nurse||Can, without difficulty, get referral to specialists when needed|
|Difference||6 ppt.||7 ppt.|
|Difference||5 ppt.||15 ppt.|
|More than high school||95%||84%|
|Less than high school||84%||73%|
|Difference (more than h.s. – less than h.s.)||11 ppt.||11% ppt.|
|= or > 400%||97%||88%|
|200% – 399%||94%||83%|
|100% – 199%||89%||79%|
|Difference (= or > 400% – < 100%)||12 ppt.||16 ppt.|
Source: Strickland et al. (2011), Table 2
Percent of children age 1–5 with lead-in-blood levels ≥ 5 micrograms per deciliter, 1999–2002 and 2007–2010
For Hispanics in the United States, the educational experience is one of accumulated disadvantage. Many Hispanic students begin formalized schooling without the economic and social resources that many other students receive, and schools are often ill equipped to compensate for these initial disparities. For Hispanics, initial disadvantages often stem from parents' immigrant and socioeconomic status and their lack of knowledge about the U.S. education system. As Hispanic students proceed through the schooling system, inadequate school resources and their weak relationships with their teachers continue to undermine their academic success. Initial disadvantages continue to accumulate, resulting in Hispanics having the lowest rates of high school and college degree attainment, which hinders their chances for stable employment. The situation of Hispanic educational attainment is cause for national concern.
Today, most parents and their children believe that a college degree is necessary for obtaining stable and meaningful work (Schneider and Stevenson, 1999). This attitude is reflected in the educational expectations parents hold for their children and in the expectations that young people have for themselves (U.S. Department of Education, 1995b, p. 88). High educational expectations can be found among all racial and ethnic groups regardless of their economic and social resources (p. 73). Although parents and children share high educational aims, their aspirations do not necessarily translate into postsecondary matriculation. This is especially the case for Hispanic high school students, particularly those whose parents have not attended college (Nuñez, Cuccaro-Alamin, and Carroll, 1998).
Despite high educational expectations, Hispanics are among the least educated group in the United States: 11 percent of those over age 25 have earned a bachelor's degree or higher compared with 17 percent of blacks, 30 percent of whites, and 49 percent of Asian Americans in the same age group (U.S. Census Bureau, 2003).1 Even more troubling, more than one-fourth of Hispanic adults have less than a ninth-grade education (U.S. Census Bureau, 2002b). These numbers represent all Hispanic groups and include recent immigrants. When examined by country of origin, educational attainment for Hispanics varies. As shown in Figure 6-1, Mexican Americans, who are the largest and fastest growing Hispanic subgroup in the United States, have the lowest rates of educational attainment compared with other groups. Cuban Americans report the highest levels of high school completion, and “other Hispanics” report the highest levels of bachelor's degree attainment. Most data sets do not distinguish among Hispanic subgroups, disregarding important cultural and economic differences among them. Whenever possible, analyses in this chapter attend to such differences.
Given the growth of the Hispanic population in the United States, most notably in the past decade (U.S. Census Bureau, 2001a), and the increasing importance of a college degree even for entry-level jobs (Carnoy, 2000), the barriers Hispanics face in realizing their educational ambitions is a major policy concern (see Chapter 4). This chapter presents the current state of educational opportunities available to the majority of Hispanic students in elementary, secondary, and postsecondary schools. Similar to other chapters in this volume, this chapter moves beyond the descriptive and explores some of the institutional and student-level factors that appear to be hindering Hispanic educational success. The goal is to identify some of the barriers to educational advancement experienced by Hispanic students in the United States, including entering school at a disadvantage because of a lack of exposure to literacy activities at home and in early formalized school settings, teacher assessments of students' language proficiency unduly influencing instructional practices, how the relationship between Hispanic students and their predominantly non-Hispanic teachers encourages disengagement from academic work, and how the lack of academic guidance pertaining to course selections and college choice impedes Hispanics from attending four-year colleges.
TAKING THE FIRST STEPS: ACTIVITIES AT HOME
One of the most important factors in school success is the extent to which parents actively participate in their children's education prior to their entry into formal preschool or kindergarten programs (U.S. Department of Education, 2003d). Specific activities, such as reading to children, have been shown to enhance children's language acquisition, early reading performance, social development, and later success in school (Loeb, Fuller, Kagan, and Carrol, 2004; National Research Council, 1998). National trend data from the National Household and Education Survey (NHES) from 1993 to 1999 indicate that Hispanic children age 3 to 5 are less likely to be read to compared with non-Hispanic children. Families in which parents' primary language at home is Spanish have especially low rates of participation in literacy activities. With respect to reading to children three or more times per week, Hispanic families in which both parents speak only Spanish at home had participation rates that were nearly 50 percentage points lower than white families in 1999. By contrast, for Hispanic families in which both parents speak English at home, participation rates were only 15 percentage points lower than white families. Hispanic households are also less likely than white households to participate in other prekindergarten literacy activities, such as telling their child a story or visiting a library, again with a pronounced difference between Hispanic families who speak English in the home and those who do not.
Families with limited economic, educational, and social resources are often less likely to participate in literacy activities than those with greater resources. Using data from the NHES, families were categorized by income level to determine whether literacy activities still differ by race/ethnicity when resources are taken into account.2Figure 6-2 suggests a statistically significant association between literacy activities and family resources across racial/ethnic groups.3 However, at all income levels except the highest, Hispanic families are less likely than other groups to participate in literacy activities (see Figure 6-2), indicating that lower participation in literacy activities can be partially explained by lack of financial resources. An additional mechanism explaining different rates of participation is language: within each income bracket except the highest, Hispanic families in which neither parent speaks English were less likely to read to their children, tell a story, or visit a library than Hispanic families in which both parents speak English in the home.4 The rates of literacy participation for Hispanic families who speak English at home more closely resemble those of white and black families, suggesting that bilingual families may be more assimilated into American culture, and specifically into practices that increase school performance.
Average rates of participation for 3- to 5-year-olds not yet enrolled in kindergarten in being read to by a family member, by race/ethnicity according to income. NOTES: Differences between racial groups and language status are statistically significant (more...)
It is difficult to draw causal conclusions regarding the effects of language spoken at home across racial/ethnic groups and within the Hispanic population due to methodological shortcomings of existing data sets: the small numbers of non-English speakers in the existing samples, some surveys not being administered bilingually, and questions regarding literacy activities not differentiating between reading to a child in Spanish and doing so in English. However, multivariate analyses based on these NHES data show that, regardless of mother's educational attainment and household income, Hispanic parents who speak only Spanish at home are less likely to read to their children than other Hispanic parents (both bilingual parents and those who speak only English).5 However, NHES data indicate that parents who are bilingual are more likely to engage their child in literacy activities than parents who speak only Spanish, but their children are still at a disadvantage in reading compared with children whose parents speak only English. While participating in literacy activities in English is the optimal preparation for schooling, being read to in Spanish also exposes children to literacy strategies that will be beneficial as they start school. Students who are successful readers in their native language employ the same strategies to help them read in English (Jiminez, Garcia, and Pearson, 1996; Saville-Troike, 1984). However, parents who speak only Spanish in the home are more likely to be recent immigrants, live in disadvantaged communities, be unfamiliar with American cultural and educational practices, and have lower levels of education and less income. Taken together, this confluence of language, nativity, and environment creates obstacles for young children as they prepare to enter school.
Most young children will attend some type of preschool program before entering kindergarten. Increasingly, scholars have pointed to the importance of having children attend preschool, arguing that it produces persistent gains on achievement tests and reduces the likelihood of grade retention and placement in remedial programs, especially for low-income children (Barnett and Camilli, 2002). Quality preschool and kindergarten experiences provide the basic foundation for children's later cognitive and social development (Elkind, 1981; Wadsworth, 1989). Specifically, for Hispanic children, preschool can serve as a mediator between home and school. By exposing children to English and by socializing them into academic and cultural norms, even early schooling can reinforce the importance of education for future job success (Currie and Thomas, 1996). Despite evidence showing the benefits of preschool attendance, Hispanic children are the least likely to be enrolled in preschool. In 1999, 60 percent of white children who were 3 years old attended preschool, whereas only 26 percent of Hispanic children had started their education at this age (U.S. Department of Education, 2003d, p. 23). Among Hispanic 4- and 5-year-olds, enrollment rates were slightly higher and more closely resemble those of white and black children: 64 percent of Hispanic 4-year-olds attended preschool, compared with 69 percent of white and 81 percent of black 4-year-olds; among 5-year-olds, 89 percent of Hispanic, 93 percent of white, and 99 percent of black children attended preschool. Black children, however, are significantly more likely to attend preschool than Hispanic children in all age groups.
Some positive changes in Hispanic attendance in preschool programs can be seen by looking at participation in Head Start, which is specifically designed to serve disadvantaged children and uses federal poverty guidelines as a key factor for assessing eligibility. In 1998, black children age 5 and under had an attendance rate that was almost 10 percent higher than eligible Hispanics. By 2003, however, black children had an attendance rate that was only about 1 percent higher than Hispanic children (see Table 6-1). The higher attendance rates of Hispanic children may be the result of more parents taking advantage of Head Start, or it may merely reflect increases in the numbers of Hispanic children eligible for the program.
Head Start Enrollment Trends for Children Age 5 and Under by Race/Ethnicity, 1998–2003 .
Attending Head Start appears to be a positive experience for most Hispanic children. Currie and Thomas (1996) have shown that Hispanic children who are enrolled in Head Start perform slightly better on a series of cognitive tests than those who do not attend any preschool program. However, the effects of participating in Head Start differ across Hispanic subgroups. The advantages of attending Head Start are the greatest among children of Mexican origin; Puerto Rican children appear to reap fewer benefits, although they do perform better than those who attend other types of preschool programs. One explanation for this difference may be the poor quality of other available preschool programs (Currie and Thomas, 1996).6 While attending Head Start programs appears to provide some benefits, lack of available quality preschool programs remains an obstacle for some Hispanic children. Currently, programs such as universal preschool are being implemented in several states, including California. However, critics of such programs argue that while state-funded preschool allows access to preschool to more children, it detracts from creating quality preschools (Olsen, 1999).
Risk Factors for Kindergartners
Limited success in early schooling can be traced to several family background characteristics. Specific factors, such as having a mother who did not complete high school (Bianchi and McArthur, 1993), living in a single-parent home (McLanahan and Sandefur, 1994), living in a low-income or welfare-dependent household (U.S. Department of Education, 1995a), and having parents who speak a language other than English in the home (Kao, 1999; Rumberger and Larson, 1998) place children at risk of not succeeding academically (Pallas, Natriello, and McDill, 1989). These broad indicators, several of which are interrelated, do not necessarily predict that a student is destined for school failure. However, students whose families have combinations of these factors are more likely to have difficulty in school. Hispanic and black children entering kindergarten are disproportionately from families with one or more of these risk factors (see Figure 6-3). The proportion of children with two or more risk factors is five times larger among Hispanics (33 percent) and four times larger among blacks (27 percent) than among whites (6 percent) (U.S. Department of Education, 2001a).
Percentage distribution of kindergartners by number of risk factors and race/ethnicity: Fall 1998. NOTE: Percentage may not add to 100 due to rounding.
To examine the risk factors for first-time kindergartners of different racial/ethnic backgrounds, several analytic models were constructed distinguishing among whites, blacks, Asians, and Hispanics (categorized by the language parents report is primarily spoken at home).7 The three models, in Appendix Table A6-1, show that race/ethnicity is differentially associated with each risk factor, and that Hispanics who speak English at home face different risks than those who speak Spanish at home. Hispanics, especially those who speak Spanish at home, are much less likely than blacks to be in a single-parent family relative to whites. There is a strong sense of family among Hispanics that is reinforced by religion, perhaps making single parenthood less likely to occur. For example, only 47 percent of Hispanics who primarily speak Spanish find divorce acceptable, compared with 72 percent of the U.S. population as a whole (Pew Hispanic Center Survey Brief, 2004). However, as shown in Chapter 5, single parenting is now rising among Hispanic families; if this trend continues, it may place more Hispanic students at risk.
The picture changes, however, when examining the likelihood of having a mother who does not have a high school diploma or being raised in a low-income family. Hispanics are between two (those who speak English at home) and three times (those who speak Spanish at home) more likely to have a mother with low educational attainment compared with whites, even when other risk factors and socioeconomic status have been taken into account. In addition, Hispanic families in which the parents speak Spanish at home are more than twice as likely to be below the poverty threshold as non-Hispanic whites.8 The risk factors seem to interact or be predictive of one another as well. Parents in Spanish-dominant families tend to be both less well educated and more likely to be poor. In general, these findings suggest that, although there are large numbers of Hispanics with two or more risk factors, the pattern of risk differs considerably for Hispanics who speak English at home and those who speak Spanish at home.9
As with literacy activities that occur prior to formal schooling, parental education and limited English proficiency play an important role in academic success when examining risk factors contributing to school performance. A parent's primary language has implications for how involved he or she can be in their child's education. Even a bilingual parent may have trouble with reading comprehension if he or she has not completed high school (Huerta-Macias, 2003; Zulmara and Necochea, 2003). Visiting the library or enrolling one's child in a preschool program requires knowledge of what is available, where it is located, and how to get there. The most economically advantaged parent still needs logistical and organizational support to enroll and transport their young child to a preschool program. Furthermore, with respect to formal schooling, if kindergarten is not required, then parents may not even receive information about available programs.10
Most existing data do not indicate whether Hispanic children in preschool or formal school are taught in English or Spanish. However, because of the monolinguistic nature of the U.S. school system, encouraging English proficiency in students and parents at the earliest possible stage is likely to lead to a stronger foundation for school learning and later academic success. Parents with young children, especially those who are first-generation immigrants, are likely to benefit if their schools and communities worked together to provide parent literacy programs, translators at school-related activities, advice on how to assist children in homework or engage them in academic activities, before- and after-school child care, and community outreach programs.
ACADEMIC PERFORMANCE IN THE PRIMARY GRADES, MIDDLE SCHOOL, AND HIGH SCHOOL
By the time they enter kindergarten, Hispanic students for the most part already trail their classmates in reading and mathematics achievement. Results from a recent national study of kindergartners, the Early Childhood Longitudinal Study Kindergarten Class of 1998–1999 (ECLS–K), point to a problematic academic future for Hispanic children. Non-Hispanic white children were more likely to score in the highest quartile in reading, mathematics, and general knowledge than black or Hispanic children (U.S. Department of Education, 2000a). Examining early literacy skills, Asian and non-Hispanic white children were more likely to recognize letters, beginning sounds, ending sounds, and sight words than blacks or Hispanics (see Table 6-2). With the exception of American Indians, Hispanic children whose parents do not speak English at home were the least likely to have passing reading proficiency scores across all tasks.
Percentage of First-Time Kindergartners Passing Each Reading Proficiency Level, by Child's Race/Ethnicity, Fall 1998 .
Results for mathematics proficiency were similar to those for language proficiency (see Table 6-3). Hispanic students whose parents primarily speak Spanish at home were the least likely to have passing scores for number and shape recognition, relative size, ordinal sequence, and addition and subtraction. In this instance, passing rates were lower than those for American Indians in all categories. The academic achievement gap between Hispanics and other groups at the onset of schooling continues through the primary grades, suggesting that the effects of family background characteristics, including language, create an initial barrier that is difficult to overcome.
Percentage of First-Time Kindergartners Passing Each Mathematics Proficiency Level, by Child's Race/Ethnicity, Fall 1998 .
Using ECLS data from kindergarten and first grade, Reardon and Galindo (2003) conducted a series of multivariate analyses that show substantial variation in mathematics achievement scores among Hispanic subgroups (see Figure 6-4). Cuban Americans are the most similar to non-Hispanic whites, with Mexican Americans and Central Americans scoring nearly one standard deviation below their white classmates. This trend persists over time: by the end of first grade, Cuban Americans catch up to whites while Puerto Ricans, Mexican Americans, and Central Americans fall further behind. The achievement gap in mathematics is especially troubling because both instruction and performance in mathematics tend not to be dependent on language, in contrast to reading. One might expect that academic performance would improve as English language proficiency increases. However, this does not appear to be the case.
Hispanic mathematics achievement gap by country of origin, kindergarten and first grade. SOURCE: U.S. Department of Education (1998).
In the next set of analyses, Reardon and Galindo (2003) examined mathematics performance from kindergarten to first grade among first-, second-, and third-generation Mexican Americans and non-Hispanic whites. First- and third-generation Mexican immigrant students started kindergarten with lower levels of math skills than second-generation students, and that pattern did not change over time (see Figure 6-5).
Hispanic mathematics achievement gap by generational status, kindergarten and first grade. SOURCE: U.S. Department of Education (1998).
The language barrier seems to place first-generation immigrant students at a decided disadvantage compared with second-generation Mexican Americans and whites. What is surprising is that this performance gain by second-generation immigrants does not hold for those who are third generation. One explanation may be that parents of second-generation students are motivated to succeed and instill those values in their children. The parents of third-generation students, particularly those who continue to speak only Spanish and who live in barrio communities with limited economic resources and poor schools, may become disillusioned with education as a path to social mobility and transmit these attitudes to their children.
Another problem may be teachers' perceptions of their students' abilities. Reardon and Galindo (2003) found that Hispanic students entering kindergarten were rated lower than white students by their teachers, regardless of their academic ability. In the ECLS kindergarten survey, teachers were asked to rate the math and literacy readiness and proficiency of each of the students in the sample in math and literacy skills. Each student was also tested in mathematics and reading by a trained ECLS assessor. Reardon and Galindo (2003) conducted a series of multivariate analyses to look for evidence of teacher bias. Given students of equal ability from the same classroom, they estimated the extent to which the teachers rated Hispanic students lower than non-Hispanic whites. Table 6-4 shows, that, in the fall of the kindergarten year, the Hispanic students were rated, on average, more than one-eighth of a standard deviation below the non-Hispanic white students. This rating gap was reduced by one-half to two-thirds by the spring of the kindergarten year and completely disappeared by the spring of first grade.
Average Difference in Standardized Teachers' Ratings of Hispanic and Non-Hispanic White Students in the Same Classroom, by Subject and Grade .
It appears that some teachers base their initial ratings of students, in part, on the student's ethnicity. This teacher bias is reduced as teachers come to know students better over time, explaining the substantial initial rating gap and its decline over kindergarten and first grade. Furthermore, it is possible that this bias acts as a self-fulfilling prophecy, so that students' test scores come to more closely match their teachers' ratings over time. This could be the result of subtle or overt differences in instructional practices directed toward Hispanic students who are rated lower than their white classmates. Another possible explanation is that the teacher ratings are unbiased measures of some aspect of mathematics and reading skills that is not measured by the tests, but on which non-Hispanic white students rate higher than Hispanic students.11 Whatever the explanation for the closer match between student performance and teacher ratings over time, the initial gap in teacher assessment between white and Hispanic students does point to teacher bias. Such bias at the onset of formal schooling sets the stage for lower expectations and underperformance by Hispanic students.
The initial achievement gap between Hispanic and white students persists throughout middle school and high school. Tracing the academic performance of Hispanic students over the past 20 years using trend data from the National Assessment of Educational Progress (NAEP) shows that Hispanic students continue to lag behind non-Hispanic whites (U.S. Department of Education, 2003d). Figure 6-6 indicates that for fourth, eighth, and twelfth graders, differences in average reading scores between Hispanic and non-Hispanic white students are evident, and this pattern is consistent across Hispanic subgroups.
Differences between white, non-Hispanic, and minority students' average NAEP reading scale scores, by age: Selected years, 1992–2002. NOTES: Year 1 = 1992, Year 2 = 1994, Year 3 = 1998, Year 4 = 2002. The scale scores reported are plausible value (more...)
In 2002, Hispanic fourth graders scored close to 30 points lower than their white classmates in reading (see Figure 6-6). Although there have been some fluctuations in the scores of Hispanics and non-Hispanic whites, a 30-point differential after several decades of school reform is clearly problematic. The situation for eighth-grade students closely resembles that for fourth graders. By twelfth grade, Hispanic students have closed the gap somewhat, with Hispanics scoring an average of 18 points lower than whites. However, since Hispanics have a higher dropout rate, these averages are probably inflated because they reflect only the scores of more promising students who have stayed in school through twelfth grade. Compared with black students, Hispanic students are doing slightly better at all three grade levels, although the differences are small. The lower test scores of blacks and Hispanics suggest that low socioeconomic status may play a role in creating this achievement gap.12
When Hispanic reading scores are examined by subgroup, Mexican Americans and Puerto Ricans tend to score the lowest. This is cause for concern, because Mexican American immigrants are the largest and fastest growing minority group among young elementary school students. Assuming there is no immediate and effective intervention strategy to improve their reading skills, it is reasonable to expect that by eighth grade, these students' levels of achievement will continue to be low, which may contribute to their higher dropout rates.13
The mathematics achievement gap between Hispanics and non-Hispanic whites is similar to that for reading scores (see Figure 6-7). Across all grades, Hispanic students scored higher than blacks, but lower than whites, in mathematics. For twelfth graders, the gap between Hispanic and non-Hispanic white students is smaller than in earlier grades, most likely because Hispanic students with poor academic records and low test scores tend to leave school before twelfth grade.
Differences between white, non-Hispanic, and minority students' average NAEP mathematics scale scores by age: selected years 1990 to 2000. NOTES: Year 1=1990, Year 2=1992, Year 3=1996, Year 4=2000. Data for Cuban respondents were not available until (more...)
In looking at NAEP data over time, it appears that Hispanic children have been making achievement gains, but so have other groups, including whites; thus the achievement gap is not narrowing (Pew Hispanic Center Fact Sheet, 2004b). In the 1990s, even though Hispanic scores in reading and mathematics increased overall, the achievement gap actually increased, suggesting that this gap will widen by the time this cohort of students reaches twelfth grade. Overall, achievement results from kindergarten through twelfth grade show differences in test scores among Hispanic subgroups and across generations. However, not all data sets include sufficient information on generational status or Hispanic subgroups to identify interventions that would be most effective for different groups. Better national longitudinal data must be collected that distinguishes between Hispanic subgroups, English as a second language (ESL) versus non-ESL curricula, and immigrant status. Data should also be collected on the types of students who drop out between eighth and twelfth grade. Nonetheless, even these broad indicators point to the need for immediate academic interventions at the primary and middle school levels.
Elementary and Middle School Contexts
As illustrated by growing gaps in achievement from fourth to eighth grade, school characteristics can play a role in student achievement, especially during middle school, when students form attachments to their teachers and schools.14 Fostering attachment and a sense of belonging is particularly problematic in large school environments. Hispanics are the most likely to be enrolled in large schools with large class sizes, and these schools are also more likely to be underfunded and deficient in resources.15 The majority of Hispanic students at both the elementary and secondary levels attend urban schools that are above average in size (U.S. Department of Education, 1996b). Hispanics comprise one-quarter of the student population in central-city schools (ERIC Clearinghouse on Urban Education, 2001).16 Compared with other groups, Hispanic students disproportionately attend schools with the highest levels of poverty, as measured by the proportion of students who qualify for a free or reduced price lunch, and are enrolled in the most highly segregated schools (Orfield and Yun, 1999). Approximately 75 percent of Hispanic students attend schools with over 50 percent minority student populations, and a little over 35 percent of Hispanic students attend schools with over 90 percent minority student populations (Orfield and Yun, 1999).
Hispanics are also more likely to be in schools with inexperienced or noncertified teachers (U.S. Department of Education, 2003a; Valencia, 2002). Public and private schools with the highest percentages of minority and limited-English proficient students are more likely to employ beginning teachers than schools with lower percentages of minority limited-English proficient students, thus virtually ensuring that a high proportion of Hispanic youth, who most need experienced teachers, are taught by less-qualified instructors. Furthermore, these schools often have too few bilingual teachers certified in ESL (Hacsi, 2002; U.S. Department of Education, 1996b). The quality of bilingual programs also varies across schools and districts and may in some instances interfere with, rather than enhance, students' ability to master both Spanish and English.
Finally, many urban schools have very few Hispanic teachers compared with the number of Hispanic students they instruct: only 4 percent of public school teachers are Hispanic, whereas Hispanic students at the elementary school level constitute about 15 percent of the student body nationally (U.S. Department of Education, 1997). This sometimes makes it difficult for Hispanic students to identify with teachers and view them as role models. By sharing a cultural identity with their teachers, Hispanic students might benefit by seeing someone from their own cultural background succeed. Having a teacher who is sensitive to cultural differences may also help students feel more engaged and less alienated (Graham, 1987; Valencia and Aburto, 1991). Therefore, it is important for schools serving Hispanic students to recruit more Hispanic principals and teachers to act as role models. Some strategies for recruiting minority students into the teaching profession have been recommended, for example, a forgiveness loan program in which minority students who pursue teaching need not repay student loans, credentialing of experienced minority teacher aides, and early identification and recruitment of minority students (Valencia and Aburto, 1991).
One of the most important findings of the 1980s was the recognition of the significance of the middle school experience and its lasting effects on students' schooling careers (Carnegie Council on Adolescent Development, 1989). The experiences Hispanics have in middle school often follow them through high school, creating obstacles for future schooling success. Specifically, it appears that teacher interactions and the less than optimal school contexts that Hispanic students encounter in middle school contribute to their academic and social difficulties in later years.
Finn (1989) claims that school success depends on students' sense of a close connection with their schools. Students who identify with their schools have an internalized sense of belonging; that is, they feel they are a part of the school community and that school constitutes an important aspect of their own experience. Students who feel this way are more likely to value and pursue academic or school-relevant goals and thus are more likely to participate in the classroom (Finn, 1989). In studying student–teacher relationships, Payne (1994) and Valenzuela (1999) have found that negative attitudes or teacher stereotypes of minority students may weaken bonds necessary for learning. These findings indicate that success in the classroom depends on students' ability to accept their teacher as a credible source of information. Students have to believe that the teacher respects and cares about their well-being. When this bond is not established or fully developed, students resist teachers both personally and academically, become detached from school, and consequently are less likely to succeed in school. The ability to form these types of bonds with minority students is particularly difficult for white middle-class teachers working in urban schools (Buriel, 1983; Katz, 1999; Rosenbloom and Way, 2004). Payne (1994) found that when these teachers avoid or reject negative attitudes and stereotypes, they are able to offer minority students the respect and high expectations that facilitate academic success. Prior research indicates that when minority students are aware of negative stereotypes regarding their academic ability, “stereotype threat” is activated. This heightened awareness of negative stereotypes may cause Hispanics to underperform, particularly on aptitude or cognitive ability tests, and score lower than white students (McKown and Weinstein, 2003; Steele and Aronson, 1995).
Teachers and administrators who lack an understanding of cultural differences can also hinder the academic success of Hispanics by misassigning bilingual students to remedial programs. This inaccurate assessment of student abilities has more recently been found even among kindergarten students, whose teachers tend to underestimate the literacy skills among Hispanic kindergarteners (Reardon and Galindo, 2003). When teachers or administrators use lack of English proficiency to signal special needs, language-minority students are overrepresented in special education classes, in which academic performance is underemphasized in favor of social adjustment (Schmid, 2001).
Bryk and Schneider (2002) found that many teachers, particularly those working in urban schools, do not know their students well and lack an empathetic understanding of their situations or the interpersonal skills to engage them—conditions that are necessary for a trusting relationship to evolve and be sustained. Martinez (2003) found that Mexican Americans more than other students feel better when they are not with their teachers. Data from 625 students who participated in the Alfred P. Sloan Study of Youth and Social Development, a longitudinal study of how young people form ideas about postsecondary school and work, show that when Mexican American students are not with their teachers, they are happier and more excited, feel better about themselves, and believe that they are living up to their own expectations (Csikszentmihalyi and Schneider, 2000). In addition, when in the company of a teacher, Mexican American students also are more likely to believe that teachers have more unfavorable thoughts about them than about other racial/ethnic groups. In contrast, when white students are with their teachers, they report feeling that they are meeting their own expectations, are relaxed and challenged, and indicate that what they are doing is important to their future goals. Like Mexican American students, black students feel happier and more relaxed when not with their teachers. Black students, however, experience higher levels of challenge when they are with teachers, much like white students.
Weak relational ties between Hispanic students and their teachers may diminish motivation or engagement in academic work, which in turn can undermine academic achievement. When weak relational ties exist between students and teachers, students may feel that teachers have low expectations of them or do not care about them, which can be highly discouraging and cause Hispanic students to disengage from classroom activities or ultimately withdraw. When disengaged in the classroom, Hispanic students are less likely to see the relevance of what is being taught to their future schooling or careers. It seems particularly important for Hispanic students to have teachers who have high expectations for their academic performance, strengthen personal ties between themselves and their students, and point out the relevance of schoolwork to future opportunities in both school and the labor market.
Transitioning into High School
Moving from middle school to high school is a challenging and uncertain process for many students, even under optimal circumstances (Schiller, 1995). This transition is especially problematic for Hispanics and blacks living in urban areas. These students are more likely than Asians and whites to be uncertain about what high school they will attend and seem to have the most difficulty adjusting to a new school. Based on these findings, Schiller (1995) concludes that Hispanics and blacks require assistance in making the transition from middle school to high school. Not only do they require social and psychological support, but they especially need early guidance about the consequences of taking specific courses for postsecondary school options.
One school organizational factor that is strongly related to academic performance is curricular differentiation, that is, how students are sorted into different ability groups and courses (Hallinan, 1994; Smith, 1995; Stevenson, Schiller, and Schneider, 1994). The course selection process, especially in the eighth grade, affects standardized test scores and college attendance and completion (Schneider and Stevenson, 1999; Stevenson et al., 1994). For example, unlike English and social studies, the mathematics curriculum becomes sharply differentiated beginning in middle school (Usiskin, 1987). Students given instruction in algebra rather than general mathematics in eighth grade are at an advantage as they can take more advanced courses in high school and move through the high school mathematics curriculum more quickly. Smith (1995) also finds that students who took algebra in eighth grade had higher mathematics achievement scores and expressed higher educational aspirations in the tenth grade. That Hispanic students are less likely than Asians, whites, and blacks to take algebra in the eighth grade greatly limits their future curricular options (U.S. Department of Education, 1990).
Before eighth graders enter high school, they are given the opportunity to select a specific curricular program from several different options, commonly labeled college preparatory, general or comprehensive, and vocational. These programmatic choices are not benign with respect to students' schooling careers and academic achievement. Data from High School and Beyond and the National Education Longitudinal Study of 1988–2000 (NELS: 88-2000) have linked high school curricular placement to achievement, educational expectations, and occupational aspirations (Gamoran and Mare, 1989; Oakes, Gamoran, and Page, 1992; Schneider and Stevenson, 1999). By spring of eighth grade, only 23 percent of Hispanic students plan to enroll in a college preparatory curriculum, compared with 25 percent of blacks, 31 percent of whites, and 37 percent of Asian Americans (U.S. Department of Education, 1990). However, Hispanics also comprise 29 percent of eighth graders who are unsure about their high school curriculum program.
Part of the uncertainty Hispanic students and their parents feel about educational practices can be traced to parents' limited experiences with the U.S. education system and the trust they place in the authority and knowledge of teachers. Mexican American immigrant parents are particularly vulnerable and more likely to defer to teachers and administrators, rarely questioning their decisions (Bryk and Schneider, 2002). Curricular counseling for college, especially for recent immigrants who may be unfamiliar with the complexities of the U.S. education system, must begin before high school. For this reason, providing Hispanic eighth graders early and more detailed information about which curricular programs lead to college admission would greatly assist many students in making choices that promote higher levels of educational attainment.
High School Course Selection
Schools play a critical role in influencing what courses students will take by deciding what courses will be offered, establishing procedures for admission to particular courses, and creating a climate whereby teachers and counselors are encouraged to adopt a selective or universal approach to student counseling and academic planning. Taking specific course sequences has certain educational advantages. For example, students who successfully complete courses in algebra, geometry, and trigonometry are much more likely to take the next advanced level of mathematics than their classmates who take other course sequences. In certain subjects, such as mathematics, students are typically not allowed to take advanced courses out of sequence. This makes it difficult for students to take a high-level course if they lack the necessary prerequisites. Courses taken in high school better predict who attends college than family background, school characteristics, or educational expectations. Course selection decisions are in turn more influenced by student academic ability and prior achievement than by family background characteristics, such as parents' educational attainment (Stevenson et al., 1994).
Hispanic students are less likely than white students to complete advanced mathematics; they are also less likely than both white and black students to take certain advanced science courses. Table 6-5 shows that Hispanics are about 20 percent less likely than whites to take advanced course work in mathematics. The low numbers of Hispanics taking advanced mathematics courses are of serious concern because advanced mathematics course-taking, more than any other subject, appears to have the strongest relationship to whether a student matriculates to a two-year or a four-year college (Riegle-Crumb, 2003).
Percentage Distribution of High School Graduates, by Highest Levels of Mathematics Courses Completed and Race/Ethnicity, 1998 .
Several subjects, including mathematics, are often offered as advanced placement (AP) courses. Students who enroll in AP classes may elect to take an examination in that subject, and if they score above a designated threshold, can earn college credit for that course. Minority students, with the exception of Asian Americans, are less likely than white students to take AP examinations (see Table 6-6). While the proportion of Hispanic test takers increased after 1997, only 9 percent of AP test takers in 2002 were Hispanic, half of whom were Mexican American (see Table 6-6). The increase among Hispanic twelfth graders taking AP exams appears to be driven mainly by the population growth of Hispanics, especially Mexican Americans, in the United States who are now staying in high school through twelfth grade. In 1997, 10 percent of all twelfth graders were Hispanic; by 2001, 12 percent of the twelfth-grade population was Hispanic (Common Core of Data, the Department of Education's database on public elementary and secondary U.S. schools, which contains basic information and descriptive statistics on schools, school districts, students, staff, and fiscal data).
Proportion of Twelfth-Grade Students Who Took Advanced Placement Examinations by Race/Ethnicity, 1997, 2001, 2002 .
Perhaps one of the most significant indicators of preparedness for college is one's score on college entrance examinations such as the ACT and SAT, which measure students' verbal, mathematical, and analytic skills. Over the past decade the number of minority students taking the SAT has risen dramatically. Hispanics accounted for 9 percent of the SAT-taking population in 2001 (U.S. Department of Education, 2003d, p. 62; see Table 6-7); however, they constituted 14 percent of the U.S. high school population enrolled as juniors or seniors (U.S. Census Bureau, 2001b).
Percentage Distribution of Students Who Took the Scholastic Assessment Test (SAT), by Race/Ethnicity, 1991 and 2001 .
With respect to SAT performance, Hispanics, although scoring higher than blacks, continue to lag behind whites and Asians on the SAT in both the math and verbal components of the exam (see Figure 6-8). Low test scores coupled with fewer college preparatory courses decrease the chance that Hispanics will be accepted into highly selective colleges.
Average scholastic assessment test (SAT) scores (verbal and math) for college-bound seniors, by race/ethnicity, 2003. SOURCE: College Board (2003).
Hispanic students are the least likely group to take college entrance examinations and to apply to college (Fry, 2004). To further explore the relationship between race/ethnicity and academic preparation, particularly among Hispanics, a series of analyses predicting the likelihood of taking advanced course sequences and college admission tests was conducted with data from the second follow-up of NELS: 88-2000. Figure 6-9 shows the probabilities that a student will take an advanced math course (e.g., trigonometry or above), an advanced science course (e.g., physics, advanced biology, or chemistry), or the SAT by the end of high school. These probabilities are estimated taking into account racial/ethnic background and language spoken in the home. When background characteristics are not accounted for, there are pronounced differences between groups with respect to academic preparation, with Hispanics among the least likely to be engaged in college preparatory activities.17 Once background characteristics are taken into account, Hispanics who are bilingual are more likely than whites to take advanced courses and the SAT (see Figure 6-9). Bilingual students have the advantage of having parents who are proficient in both English and Spanish and who are thus able to bridge cultural and language barriers to secure educational opportunities for their children (see Kim and Schneider, 2004).
Probability of taking advanced math courses, advanced science courses, and the SAT by the end of high school by race/ethnicity, controlling for various family and student characteristics. NOTES: Probabilities are based on equations from logistic regression (more...)
The next set of analyses shows the probability of attending a four-year college for all racial/ethnic groups, taking into account academic course preparation and having taken the SAT. Figure 6-10 shows that the probability of matriculating to a four-year college is higher among students who have taken the SAT and advanced coursework in mathematics and science compared with those who are not so prepared. The second panel predicts the probability of attending a four-year versus a two-year college by academic preparation and race/ethnicity. The likelihood that Hispanics and whites will attend a four-year college increases by about 30 percentage points when academic preparation is taken into account. Hispanics from Spanish-speaking families are nearly as likely as blacks to attend four-year colleges when they have high levels of academic preparation. Highly prepared Hispanics are even more likely than comparable whites to attend a four-year versus a two-year college. These analyses suggest that one reason why Hispanics are more likely to matriculate to two-year rather than four-year colleges is poor academic preparation. Therefore, for Hispanics, especially those who are first-generation college-goers, it is imperative that schools offer programs explaining the importance of college preparatory curricula, SAT preparation, and advanced course-taking.
Predicted probabilities (in percentages) of going to a four-year college versus a two-year college or not going to college by academic preparation (advanced math course, advanced science course, and taking the SAT) and race/ethnicity. NOTES: Probabilities (more...)
High School Noncompleters
Despite high educational expectations, Hispanics have the highest high school dropout rate (28 percent in 2000) compared with blacks and whites (U.S. Department of Education, 2000b). While the percentage of 16- to 24-year-old Hispanics without a high school diploma has decreased over the past 30 years, the status dropout rate of Hispanics is still more than double the rate of both whites and blacks (see Figure 6-11).18
Status dropout rates of 16- through 24-year-olds, by race/ethnicity: October 1972 through October 2001. NOTES: Due to small sample size, American Indians/Alaska natives and Asians/Pacific Islanders are included in the totals but are not shown separately. (more...)
However, this status dropout rate is inflated by recent increases in teenage Hispanic immigrants who never enroll in U.S. schools (Fry, 2003). Hirschman (2001) estimates that almost half of Mexican 15- to 17-year-olds who arrived in the United States between 1987 and 1990 did not enroll in school. These numbers are considerable, especially when compared with the dropout rates of Mexican Americans born in the United States. In 2001, 43.1 percent of foreign-born Hispanics did not complete high school compared with only 15 percent of U.S.-born Hispanic students (U.S. Department of Education, 2004a).
Figure 6-12 displays the differences between the dropout rates of foreign- and U.S.-born Hispanics. Rates for foreign-born Hispanics are double those of their U.S.-born counterparts, with the exception of South Americans. The dropout rate among U.S.-born Hispanics decreased from 1990 to 2000; however, the dropout rate among this group (14 percent) is still higher than that of blacks or whites (12 and 8 percent, respectively; Fry, 2003). One positive finding is that immigrant children who do enroll in high school are not more likely to drop out than U.S.-born students (Pew Hispanic Center Fact Sheet, 2004a).
Dropout rates according to Hispanic subgroups by immigrant and U.S.-born status. SOURCE: Census 2000 Supplementary Survey.
Some students may temporarily “stop out” of high school and later return to receive their degree through alternative programs or by earning general educational development (GED) certification. In 2001, the national high school completion rate for Hispanics was 64 percent, compared with 92 percent for whites. Such low completion rates are typical of urban schools that serve large numbers of minority students, many of whom come from low-income families (U.S. Department of Education, 2004b).19 Hispanic students remain concentrated in large urban school systems, such as Los Angeles, Chicago, and New York, where overall graduation rates are less than 60 percent. Overall, almost 40 percent of Hispanic students attend high schools in which the graduation rate is less than 60 percent (Balfanz and Letgers, 2004). Educators and policy makers remain acutely aware of the difficulties facing Hispanic students, and a variety of retention programs have been instituted in high schools with large numbers of non-completers. The most effective intervention programs for high school completion are those that link graduation to college matriculation by including college-based or college-level courses or programs, after-school academic preparation, SAT test preparation, and tutoring. While there is a documented relationship between these programs and academic success, most have not been thoroughly evaluated and have been criticized for not accounting for selection bias—that is, the most talented students are those most likely to seek out these programs. Evidence suggests a causal relationship between program participation and college matriculation; however, systematic evaluations including clinical randomized trials pertaining to these interventions are limited (U.S. Department of Education, 2001b). These types of scientifically rigorous evaluations need to be designed and conducted.20
Pathways After High School
Although the high school completion rates of Hispanic students have risen over the past decade, their job prospects remain weak because the standard requirement for stable employment in many fields is a baccalaureate degree.21 Many Hispanic students will enter the labor force immediately after high school, a pathway that economists have estimated will eventually lead to unstable employment and low wages (Levy, 1995). Some Hispanics will also enter the military, a pathway that few high school graduates are taking. It appears that Hispanic entry into the military has increased significantly; Hispanics constituted just 4 percent of military personnel in 1985, but that number rose to 11 percent in 1999 (U.S. Department of Defense, 2000). However, data are not available as to whether Hispanics are taking advantage of the educational benefits offered to military personnel.
Even though a large percentage of Hispanics choose to work after high school, over half of Hispanic high school seniors plan to attend a four-year college. College expectations of Hispanic students doubled from 24 percent in 1972 to 50 percent in 1992; actual college enrollment for Hispanics has increased, as it has for other racial/ethnic groups (U.S. Department of Education, 1995b). In 1972, over 14 percent of Hispanic high school graduates matriculated to four-year colleges (Olivas, 1979); by 2000, 36 percent of Hispanic graduates were enrolled in four-year colleges (U.S. Department of Education, 2003d). Although the number of high school graduates attending college has risen, Hispanics constitute a disproportionately small portion of those attending four-year colleges: 12.5 percent of the U.S. population in 2000 identified themselves as Hispanic, and only 7 percent of four-year college students were Hispanic (U.S. Census Bureau, 2002b; U.S. Department of Education, 2003b). The burgeoning number of first-generation Mexican Americans may account in part for the low rates of college attendance among Hispanics. First-generation immigrant parents may be unfamiliar with the complex policies and practices of the U.S. education system, which require a high level of parent knowledge and involvement, particularly with respect to academic preparation for college.
However, the enrollment rates for Hispanics are misleading, since they are more likely to enroll in two-year rather than four-year institutions, especially first-generation college-goers. In 2000, Hispanics accounted for 14 percent of students enrolled in two-year colleges and only 7 percent of those enrolled in four-year institutions (see Table 6-8).
Percentage Distribution of Enrollment in Two-Year Versus Four-Year Colleges, by Race .
One of the primary missions of the two-year community college is to provide low-cost local access to postsecondary education. Many students choose to attend two-year colleges because of financial limitations or inadequate preparation and with the intention of transferring to a four-year college. However, the majority of high school graduates who begin their postsecondary education at a two-year institution do not transfer to a four-year institution (Rendon and Garza, 1996; Schneider and Stevenson, 1999). This problem is not unique to Hispanics; for example, only 36 percent of white students who attend community college transfer to four-year colleges or complete a bachelor's degree. However, the transfer or completion rates for Hispanic students at two-year colleges are even lower; only 25 percent will go on to a four-year institution or eventually complete a bachelors degree (Fry, 2004). Schneider and Stevenson (1999) refer to this discrepancy as an ambition paradox—students with high ambitions choosing an educational route with low odds of success. Given the low transfer rates and length of time young adults spend in community college without receiving a degree, transition programs are needed to assist Hispanic students considering transferring to four-year institutions. These types of programs may also be subject to selection bias, in that students who seek out assistance may be more motivated to transfer regardless of the presence of extra help. Key features of these programs, including academic counseling and guidance about the transfer process and requirements, have been shown to help students who are unsure about the college process and may be one strategy for increasing Hispanic students' access to information (Schneider and Stevenson, 1999).
Compared with white students with similar abilities and levels of preparation, fewer Hispanic students enter highly selective colleges, attending less rigorous postsecondary institutions instead (Fry, 2004). Even more problematic, Hispanics have the lowest degree completion rate of any racial group four years after high school (U.S. Department of Education, 2003c). In the 1999–2000 academic year, Hispanics earned only 9 percent of all associate degrees, 6 percent of bachelor's degrees, 4 percent of master's degrees, 3 percent of doctoral degrees, and 5 percent of professional degrees (see Table 6-9). Overall, Hispanics tend to earn relatively more associate degrees and fewer advanced degrees than Asians, blacks, or whites.
Percentage Distribution of Degrees Conferred by Colleges and Universities, by Race/Ethnicity and Degree Level, 1999–2000 .
As Hispanic enrollment in postsecondary institutions increases, Hispanic-serving institutions (HSIs) play an important role in providing Hispanics with access to college education. HSIs are public or private degree-granting institutions in which Hispanics comprise 25 percent or more of the undergraduate full-time-equivalent enrollment. According to the U.S. Department of Education (2004), at least 50 percent of Hispanic students who are enrolled in these institutions have low family income. There are approximately 242 HSIs located in 14 different states and Puerto Rico (U.S. Department of Education, 2004b). Nearly 46 percent of all HSIs are located in Texas and California. About half of the HSIs are four-year institutions. In 1999, nearly one-half of the total Hispanic undergraduate enrollment in colleges and universities was in HSIs (U.S. Department of Education 2003d, p. 96). Much like historically black colleges and universities, HSIs also enroll a considerable population of first-generation college-goers. There are limited evaluations of the effectiveness of HSIs in terms of their matriculation rates, graduation rates, and job placement. Further research should be conducted to understand how successful these institutions are in serving Hispanic college students.