x . “‘3’; unit!” ’ . I .‘, .- 5 Lnr‘ 1.3"}: v1. " v 1: .w. 53 i;‘ £1. 32 Em‘fim ‘1‘ ‘1 fig); . I ‘ n' a _ . I - 5a A? ‘ " V ' 4 , .7. 3 t, 1 ~ .m’ ‘J § . V . V y‘astgfgga . - . ‘ ‘ .155. .3 1%.: 5., 9 5:55;, - p C '24 2 -_ bl J 53;» f t A 'f «~52 $3. m.” THESiS r’ . "' 9 \i‘x ‘ K Zoe/9M; LIBRARIES MICHIGAN STATE UNIVERSITY EAST LANSING, MICH 48824-1048 1 .1 ’i ( This is to certify that the dissertation entitled The Contribution of Race, Class, Gender, and Family Structure on Adolescents’ Expectations of the Future presented by Judy Dearing Berglund has been accepted towards fulfillment of the requirements for the Ph.D degree in Social Work W ‘0 W Major Professor’s Signature W I 1 5 1° o ‘1 Date MSU is an Affinnative Action/Equal Opportunity Institution ---O-l-O-o_-.o PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 6/01 c:/ClRC/DateDue.p65-p.15 THE CONTRIBUTION OF RACE, CLASS, GENDER, AND FAMILY STRUCTURE ON ADOLESCENTS’ EXPECTATIONS OF THE FUTURE By Judy Dearing Berglund A DISSERTATION Submitted to Michigan State University in partial fiilfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY School of Social Work 2004 ABSTRACT THE CONTRIBUTION OF RACE, CLASS, GENDER, AND FAMILY STRUCTURE ON ADOLESCENTS’ EXPECTATIONS OF THE FUTURE By Judy Dearing Berglund Combining the theories of ecological systems, domination and oppression, social reproduction, and labeling, this author examines the contribution of race/ethnicity, social class or income level, gender, and family structure on the development of an adolescent’s positive or negative expectations of their future. Based on the above theories, the author hypothesizes that adolescent expectations will be higher or lower depending on the adolescent’s specific intersection of race/ethnicity, income, gender, and family structure. Statistical tests supported predicted impact of race/ethnicity, income, and family structure, but not of gender. The research used cross-sectional 1997 data from the National Longitudinal Survey of Youth to test the hypotheses. DEDICATION To my wonderful daughter Hannah Erika Berglund: Always follow your dreams. and To my parents Wayne Dearing and Joyce Thurman Dearing: Thanks for teaching me about oppression and social class. iii ACKNOWLEDGEMENTS As with any major project, a dissertation is not the sole work of one individual. I have had many sources and avenues of support over the last several years while starting and completing this project. My family, friends, colleagues, and mentors gave me tremendous emotional support and encouragement. My constant lament was, “I don’t think I can do this” and their collective response was, “yes you can”. My dissertation committee, including my chair, Dr. Margaret Nielsen, and Dr. Victor Whiteman, from the School of Social Work, Dr. Kevin Kelly, from Sociology, and Dr. Thomas Luster, from Family and Child Ecology, were very supportive, encouraging, and helpful, every step of the way. My Dean at Saginaw Valley State University (SVSU), Dr. Donald Bachand, gave financial support and cracked the whip occasionally with useful deadlines. My colleagues in the social work department at SVSU (Dr. Steve Yanca, Dr. Lucy Mercier, and Vanessa Brooks-Herd, MSW) went out of their way to take up the slack for me while I was constantly working and fretting. The friends that I made in my PhD. cohort are incredible: Carol Burrell-Jackson, MSW and Dr. Noriko Kubota — I couldn’t have made it without your support in statistics, Dr. Peg Whalen and Dr. Noriko Kubota, the best support group anywhere (we should write a book!). My lifelong friends, Jeanette Sublett Neal, Cathy Villanova Ivcich, Devon Soderquist Polly, and their respective husbands and family, Al Way, Leo Dhont, and Ellen Hughes, my sister Kathy and her family (Scott, Natalee, Scotty), and my daughter Hannah, gave all the support that I needed with love and encouragement. A special acknowledgement to my daughter Hannah for her constant computer and word processing help: you saved me iv from having to re-type the whole thing. And, thanks to the Center for Human Research at Ohio State University for making the data set available for research. TABLE OF CONTENTS LIST OF TABLES ......................................................................................... viii CHAPTER 1 .................................................................................................. 1 INTRODUCTION ......................................................................................... 1 Why study adolescents and their expectations? ................................. 4 How expectations motivate behavior ................................................. 5 CHAPTER 2 ................................................................................................. 9 THEORIES AND EMPIRICAL LITERATURE REVIEW .......................... 9 Ecological Systems Theory ................................................................ 9 Theories of Domination and Oppression ........................................... 12 Social Reproduction and Labeling Theories ...................................... 15 EMPIRICAL LITERATURE REVIEW ........................................................ 18 Expectations ....................................................................................... 18 Social Reproduction Theory and the Process of Expectation Development ...................................................................................... 1 8 Specific Influences on Expectation Development ............................. 23 Family Influences ............................................................................... 23 Racial and Ethnic Differences ........................................................... 25 Gender Differences ............................................................................ 27 Environmental Influences‘ ................................................................. 27 The Influence of Family Income and Family Structure on the Socio-emotional Development of Adolescents .................................. 30 Family Structure ................................................................................. 34 Environment and Family Process ...................................................... 35 Research Questions ............................................................................ 37 CHAPTER 3 .................................................................................................. 39 METHODOLOGY ........................................................................................ 39 Study Background and Sample .......................................................... 39 Selected Sample ................................................................................. 42 Positive and Negative Expectations: Dependent Variables ............... 42 Reliability Data .................................................................................. 42 Family Structure, Race, Income Level, and Gender: Independent Variables ....................................................................... 45 Income Level ..................................................................................... 46 Race/ethnicity .................................................................................... 47 Gender ................................................................................................ 48 Mediating Variables ........................................................................... 48 Research Design and Analysis Approach .......................................... 48 Research Hypotheses ......................................................................... 49 vi CHAPTER 4 .................................................................................................. 52 RESULTS ...................................................................................................... 52 Family Structure and Expectations .................................................... 52 Income and Expectations ................................................................... 56 Race/ethnicity and Expectations ........................................................ 57 Gender and Expectations ................................................................... 59 Two-Way ANOVAs to Examine Main and Interaction Effects ........ 61 Family Structure and Income ............................................................. 61 Family Structure and Race/ethnicity .................................................. 63 Family Structure and Gender ............................................................. 64 Race/ethnicity and Income ................................................................. 65 Gender and Income ............................................................................ 67 Mediating Variables ........................................................................... 68 CHAPTER 5 .................................................................................................. 82 DISCUSSION ................................................................................................ 82 Major Findings ................................................................................... 84 Hypotheses 1, 2, 3 .............................................................................. 84 Hypothesis 4 ....................................................................................... 85 Hypothesis 5 ....................................................................................... 85 Hypothesis 6 ....................................................................................... 86 Tests for Main Effects and Interaction Effects .................................. 87 Mediating Variables ........................................................................... 90 Question 1 .......................................................................................... 90 Question 2 .......................................................................................... 91 Question 3 .......................................................................................... 91 Question 4 .......................................................................................... 92 Question 5 .......................................................................................... 92 Discussion of Major Findings ............................................................ 93 Limitations ......................................................................................... 95 Implications for Social Work ............................................................. 96 Implications for future research ......................................................... 98 APPENDICES ............................................................................................... 100 A. Descriptive Statistics for Positive and Negative Expectation Scales ............................................................................................ 101 B. Descriptive Statistics for Individual Expectation Questions ......... 103 C. Demographics of Cases with Missing Income .............................. 106 REFERENCES .............................................................................................. 108 vii 10. 11. 12. 13. 14. 15. 16. 17. LIST OF TABLES NLSY97 Gender, Race/ethnicity, and Sample Assignment .............. 41 NLSY97 Selected Sample: 15 & 16 Year-old Respondents by Family Structure, Race/ethnicity, Income, and Gender ..................... 43 Positive and Negative Expectations ................................................... 46 One-Way AN OVA: Family Structure and Positive Expectations ..... 55 One-Way AN OVA: Family Structure and Negative Expectations... 55 One-Way AN OVA: Income and Positive Expectations .................... 57 One-Way ANOVA: Income and Negative Expectations ................... 57 One-Way AN OVA: Race/ethnicity and Positive Expectations ......... 59 One-Way AN OVA: Race/ethnicity and Negative Expectations ....... 59 One-Way AN OVA: Gender and Positive Expectations .................... 6O One-Way ANOVA: Gender and Negative Expectations ................... 61 Two-Way ANOVA: Family Structure, Income Level, and Positive Expectations ....................................................................................... 62 Two-Way ANOVA: Family Structure, Income Level, and Negative Expectations ....................................................................................... 63 Two-Way ANOVA: Family Structure, Race/ethnicity, Positive Expectations ....................................................................................... 64 Two-Way AN OVA: Family Structure, Race/ethnicity, Negative Expectations ....................................................................................... 64 Two-Way AN OVA: Gender, Family Structure, Positive Expectations ....................................................................................... 65 Two-Way AN OVA: Gender, Family Structure, Negative Expectations ....................................................................................... 65 viii 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. Two-Way ANOVA: Race/ethnicity, Income Level, Positive Expectations ....................................................................................... 66 Two-Way AN OVA: Race/ethnicity, Income Level, Negative Expectations ....................................................................................... 66 Two-Way ANOVA: Gender, Income Level, and Positive Expectations ....................................................................................... 67 Two-Way AN OVA: Gender, Income Level and Negative Expectations ....................................................................................... 68 Multiple Regression: Mediating Effects — Attitude Toward School, Income Level, Positive Expectations ................................................. 71 Multiple Regression: Mediating Effects - Attitude Toward School, Income Level, Negative Expectations ............................................... 71 Multiple Regression: Mediating Effects — Income Level, Race/ ethnicity, Positive Expectations ......................................................... 73 Multiple Regression: Mediating Effects — Income Level, Race/ ethnicity, Negative Expectations ....................................................... 74 Multiple Regression: Mediating Effects -— Peer Characteristics, Income Level, Positive Expectations ................................................. 76 Multiple Regression: Mediating Effects — Peer Characteristics, Income Level, Negative Expectations ............................................... 76 Multiple Regression: Mediating Effects - Income Level, Family Structure, Positive Expectations ........................................................ 77 Multiple Regression: Mediating Effects - Income Level, Family Structure, Negative Expectations ....................................................... 79 Multiple Regression: Mediating Effects — Grades Received in Eighth Grade, Income Level, Positive Expectations ......................... 80 Multiple Regression: Mediating Effects — Grades Received in Eighth Grade, Income Level, Negative Expectations ........................ 81 ix CHAPTER 1 INTRODUCTION Childhood poverty and growing up in a female headed, single parent home are two characteristics of a child’s life that have been widely researched and have been found to affect childhood development in a variety of ways. Axinn, Duncan, and Thornton (1997) researched the effects of income, school completion, and self-esteem. Battistich, Solomon, Kim, Watson, and Schaps (1995) studied poverty levels of students in schools and their attitudes and motives. Bergman and Scott (2001) examined well-being and health-risk behaviors across economic groups. Bolger, Patterson, Thompson, and Kupersmidt (1995) studied psychosocial adjustment and poverty. Brooks-Gunn, Duncan, and Maritato (1997) provide an overview of poor outcomes for children found in low- income families. Astone and McLanahan (1997) studied family structure, parenting, and high school completion. Biblarz and Raftery (1999) also studied family structure and educational attainment. Garis (1998) compared poverty, single-parent households and at- risk behavior of children. Krein (1986) looked at single parent families and their effects on education and earnings of young men. Single, mother-only families are the poorest group of citizens in American society. The Kids Count Data Book (2003) states that the poverty rate for single-female- headed families is 39%, while that of two parent families is only eight percent. Single motherhood is only associated with poverty when the single mothers have little education and little financial support from others (Abramovitz, 1996; McLanahan, & Sandefur, 1994; Rodgers, 1996; Sands & Nuccio, 1989; Sidel, 1996). If single mothers are 1 educated and make adequate incomes, are the effects of being raised in a single parent home the same as being raised in a single parent low-income home? Popular rhetoric that labels single, female-headed households as the cause of poverty would have us believe so (Mullings, 2001). Single, female-headed families are associated with a myriad of social problems, which “provide the foundation for the cultural demonization of welfare mothers” (Hays, 2003, p. 121). The reasons that single, female-headed households are poorer include the often young age and low educational attainment of the mothers. The stigmatization and undeserving status of unwed mothers add to problems experienced by these types of families, as well as, the economic discrimination against women in general (Abramovitz, 1996; Hays, 2003; Gans, 1995; Gordon, 1994; Lord, 1993; Mullings, 2001;Sidel, 1996). Being raised in poverty, whether in a family with two biological parents or in a single parent family can be devastating, not only in the present but also for future ability to transcend that family poverty and develop into a financially independent adult. Transcending that family poverty is becoming increasingly difficult in the United States especially if living in a female- headed family. Educational opportunities are increasingly unequal (Farber, 1989; Gibbs & Bankhead, 2000; Golden, 1997; Hays, 2003; Kozol, 1991; MacLeod, 1995; National Research Council, 1993; Trusty, 2000), jobs programs are inadequate (Gibbs & Bankhead, 2000; Hays, 2003; National Research Council, 1993), and opportunities to make connections to those in power are rare (Farber, 1989; Sanders, Field, & Diego, 2001). With the 1996 changes in the welfare system, the welfare roles have dropped but that does not mean that families who previously relied on AFDC are now successfully working full time and out of poverty. According to Hays (2003), “over one-third of those who left since reform had already returned to welfare at least once by 2002. Even among those employed during that prosperous decade (19905), according to federal statistics their earnings average only $538 a month for the support of themselves and their children” (Hays, 2003, p. 222). Along with these disadvantages, the ability to transcend the poverty status of their childhood is affected by the way in which adolescents perceive their chances for the future (Gibbs & Bankhead, 2000; Mickelson, 1990; Yowell, 2002). If adolescents perceive their chances as limited, these perceptions will afi‘ect their behavior (Gibbs & Bankhead, 2000). Markus and Nurius (1986), Mickelson (1990), and Yowell (2002) specifically studied the development of expectations and the effects of social location on the development of positive or negative expectations of the future. Of specific interest in this research is how much poverty in single, mother—only families affects adolescent perspectives of future opportunities and how the intersections of race, class, gender, and family structure issues affect those perspectives. Why study adolescents: and their expectatigtsz In adolescence, a number of interdependent physical, cognitive, emotional, and social changes take place, making adolescence a pivotal time in the life of a child in western societies; (Bronfenbrenner, 1989; Bush & Simmons, 1989; Erikson, 1963). Adolescents are expected to leave the safety and comfort, if you will, of their status as children, stand on their own as adults, and become productive members of society. Developmental psychologists have examined the task of “identity formation” and describe it as a developmental process “responsive to opportunities that will obviate developmental arrest as well as promote further movement toward maturity” (Kroger, 1989, p. 5). This process is described as pivotal because adolescents are looking back at where they came from, at what their environment provided for them, and looking forward to what they can expect of their future. As adolescents are contemplating their future, their expectations of their possibilities are based on past and present experiences (Armstrong & Crombie, 1999; Gibbs & Bankhead, 2000; Mickelson, 1990; Yowell, 2002). Poverty at this stage in life may be especially detrimental. The development of expectations is dependent upon experiences and is combined with the developmental tasks of deciding on a future (MacLeod, 1987; Mickelson, 1990; Rhea & Otto, 2001; Smith, 1997; Teachman, Paasch, Day, & Carver, 1997; Wilson & Wilson, 1992; Yowell, 2002) Children and adolescents do not develop in a vacuum. At a micro level, their personal lives are affected by the larger environment of their neighborhood, schools, and other societal institutions, as well as by the larger, macro structures of society (Bronfenbrenner, 1979, 1989; Erikson, 1963; Hill- Collins, 1990; Markus & Nurius 1986; Mills, 1959). These environments set the stage for the development of either positive or negative expectations and expectations can motivate behavior (Markus & Nurius, 1986; Mickelson, 1990; Gibbs & Bankhead, 2000; Yowell, 2002). This research rests on the assumption that single, mother-only and minority families in, or near poverty are challenged by social and economic factors. It explores factors that influence the development of either positive or negative expectations of the fixture for adolescents in families. How expectations motivate behavior To aspire to something and to expect a certain outcome are two different thought processes and involve different mental tasks in the course of identity development. Aspirations, or what one aspires to become, are predicated on dreams, hopes, or desires (Markus & Nurius, 1986; Mickelson, 1990; Yowell, 2002). American school children are taught to believe that they can become anything: doctors, lawyers, school teachers, even president of the United States. These beliefs are based on the overriding ideology in capitalist society, and especially in its school systems, of a free and classless society (Bourdieu, in Swartz, 1977; Bowles & Gintes, 1976; MacLeod, 1995; Mickelson, 1990; Kozol, 1996; Zinn, 1995). Expectations, on the other hand, are based on experiences and one’s beliefs about what those experiences say about themselves and one’s possibilities (Yowell, 2002). When children grow into adolescence, their expectations for their future are the probabilities of certain outcomes based on the life they’ve experienced thus far and the lives of those in their social group (Cohen, 1955; Markus & Nurius, 1986; Mickelson, 1990; Rhea & Otto, 2001; Wigfield & Eccles, 2000; Yowell, 2002). Wigfield and Eccles (2000) explain this process from a psychological, as well as a sociological standpoint. They examine achievement motivation and ability beliefs about certain educational tasks. Adolescents’ ability-related beliefs and values become more negative as they grow older. Wigfield and Eccles explain that “children become much better at understanding and interpreting the evaluative feedback they receive and engage in more social comparison with their peers” (p. 77). Markus and Nurius (1986) coined the term “possible selves” to provide a conceptual framework or “link between self-concept and motivation” (p. 954). “Possible selves” are a combination of aspirations, or dreams, and expectations, or what we believe is possible. These “possible selves” are based on impressions or beliefs about the self, determined by past experiences, present realities, and possibilities of the future. They are therefore connected to the social environment and reproduced by the individual’s exposure to popular culture. Marcus and Nurius state that “the pool of possible selves derives from the categories made salient by the individual’s particular sociocultural and historical context and from the models, images, and symbols produced by the media and by the individual’s immediate social experiences” (Markus & Nurius, 1986, p. 954). Erikson (1963) discussed the role of expectations and past experiences that start with the aspirations of young children. It is individual initiative in establishing identity that “sets the direction toward the possible and the tangible which permits the dream of early childhood to be attached to the goals of an active adult life” (p. 258). Yowell (2002) further describes how one’s subjective understanding of the future serves as the link or bridge between one’s current self-concept and one’s behavioral adaptation (p. 63), and Bandura, Barbaranelli, Caprara, & Pastorelli (2001) state that the choices made during “the formative years shape the course of lives” (p. 187). This research examines the link between an adolescents’ environment or social location and the development of expectations along with theories and research based on these ideas. The intersection of race, class, gender, and family structure and their influence on the development of expectations is at the root of this research. The overall conceptual framework for this research is that 1) adolescents interact with multiple ecological systems due to their age and developmental stage; 2) adolescents are socially located in multiple settings based on race, class, and, gender; 3) family structure is a special situation due to certain family structures not considered normal or acceptable in our society; and 4) all of these issues affect the development of positive or negative expectations for the future, which motivate current behavior. Chapter two will discuss underlying theories that contribute to this conceptual model. CHAPTER 2 THEORIES AND EMPIRICAL LITERATURE REVIEW Chapter two will discuss theories that predict the effect of family income, family structure, race and ethnicity, and gender on the development of either positive or negative expectations for the future. This chapter also discusses empirical research that supports these theories. The chapter ends with questions specific to this research. Theoretical Foundations There are several reciprocal underlying theories that guide this research. First, Uri Bronfenbrenner’s ecological systems theory (1979, 1989) shows the link between micro and macro structures that influence socio-emotional development. Second, theories of domination and oppression describe the process of interactions between personal and societal structures which impinge on socio-emotional development. Third, social reproduction theory, along with labeling theory, show how societal institutions are structured to reproduce inequality and how this inequality affects the development of expectations. Ecological Systems Theory Children and adolescents do not develop in a vacuum (Herr, 1996). Nor do families live in a vacuum. Bronfenbrenner (1979, 1989) developed the ecological systems model to describe how ecologies of smaller micro systems are affected by, and interact with, larger macro systems. He specifically described this system as a “set of nested structures, each inside the next, like a set of Russian dolls” (p. 3). This model includes four systems, starting with the small, personal level, and proceeding to the larger, societal level. The four systems are the micro system where the individual 8 interacts with their immediate family, school or place of employment, neighborhoods, church, and peer groups; the mesosystem which includes relationships between single settings, settings in which the individual is involved directly. The relationships between settings are important in that they can support or stunt an individual’s growth. For example, “a child’s ability to learn to read in the primary grades may depend less on how he is taught than on the existence and nature of ties between the school and the home” (Bronfenbrenner, 1979, p. 3). Families that do not have a mutually respectful relationship with the school or their child’s teacher will not have this extra support for educational development. Families in poverty and of the non-dominant culture are at risk of not developing this support due to their marginal status in a predominately white and middle-class culture. The third level is called the exo-system and includes environments that may affect the individual but in which the individual does not personally interact. Examples include parents’ work, school boards, and city government. The final system, the macro-system, includes broad ideological and institutional patterns of culture or subculture that influence the development of cultural norms that dominate the development and interactions of all systems (Bronfenbrenner, 1979, 1989). Ecological theory stresses that individuals and their environments mutually influence each other and that socio-emotional development is an interaction between the individual and their social context or environment. An adolescent is involved in multiple levels in the ecological system due to their age and stage of development. For example, an adolescent’s expectations could be affected by their parents, older siblings, or the peer group they choose to hang with. They could also be affected by teachers or employers, the economic stability of institutions in their community, in essence, anyone or system that the adolescent interacts with. This research is specifically interested in how macro, ideological and institutional patterns of culture situate social location among racial/ethnic, gender, income, and family structure differences and how those situations affect the development of either positive or negative expectations of the future for the adolescents. Eamon (2001) used the ecological framework to analyze the influence of poverty on children’s emotional development. Poverty affects the micro level by producing stress in the family, either through dramatic income loss or daily struggles to meet subsistence needs. Using research from Pearlin (1989), among others, Eamon explained how strained coping skills could lead to depression in parents, which in turn affects children’s emotional development. “Parental depression impairs children’s socioemotional [sic] functioning directly by resulting in low levels of nurturance, uninvolved and inconsistent parenting and harsh discipline, and indirectly by causing conflict in the marital relationship” (p. 25 8). Family income also affects peer group acceptance. Lack of resources to purchase acceptable clothing and to participate in peer group activities serve to isolate and stigmatize poor children. Children in poor families are also more likely to attend poorly funded schools which add to the development of negative expectations for the future (see discussion of Bankhead & Gibbs, 2000, article for example). From Bronfenbrenner’s theory, negative interactions or experiences in an adolescent’s ecological system will directly affect his or her outlook on life, which in turn, will affect their expectations of what their specific future will be. This research applies this theory by examining the 10 intersection of race, class, gender and family structure 011 the deve10pment of either positive or negative expectations. Theories of domination and oppression The intersections approach of studying oppression based on race, class, and gender explains oppression through a “matrix of domination” (Hill-Collins, 1990) or a “politic of domination” (hooks, cited in Hill-Collins, 1990). Systems of oppression form an “overarching structure” (Hill-Collins, 1990) or “foundation” (hooks, cited in Hill- Collins, 1990), designed to dominate and control those in society who are deemed less worthy and considered expendable. Hill-Collins (1990) discusses multiple levels of domination which includes three levels of domination that are interconnected: “the level of personal biography; the group or community level of cultural context created by race, class, and gender; and the systemic level of social institutions” (p. 619). The personal biography level encompasses our personal psychological make-up and our ability to cope. Each biography is unique to the individual and based on individual reactions to experience. Stigma is often internalized at this level and individuals begin to chide themselves for their own shortcomings. The group or community level of domination forms a cultural context of shared experiences and ideas with those similar to the individual. People are oppressed through affinity with certain groups. The third level consists of social institutions such as schools, churches, and the media, which “expose individuals to the specialized thought representing the dominant group’s standpoint and interests” (pg. 620). While societal institutions provide certain opportunities, they require adherence to the dominant “norm” for acceptance. (Hill-Collins, 1990; Young, 1990; Zinn, 1995). 11 Young (1990) further delineates oppression theory by presenting a framework that shows how people continue to be oppressed by well-meaning people in ordinary, everyday situations, that oppression and discrimination are systematic and a part of the social fabric. She discusses five faces of oppression as exploitation, through labor and the division of labor based on those who have power and those who do not; marginalization, where certain groups are considered useless and superfluous for maintaining our society. Examples of marginalized groups include people with mental and physical disabilities, inner-city youth, and the elderly. The third face of oppression is called powerlessness and refers to the division of working people into professional and nonprofessional. The professional groups have considerable power over resources and access to resources that nonprofessionals may need but have to go through the professional group to obtain and, therefore, may experience oppression in the process. Cultural imperialism, the fourth face, is oppressive because some groups are rendered as invisible (their culture not significant or important enough to be recognized) or labeled as deviant or inferior compared to the dominant group. Young (1990) considers violence as the fifth face of oppression because it is pervasive in our society and it is designed to humiliate, destroy, or coerce certain behaviors. Examples of oppressive violence are rape, hate crimes, and lynchings. The death penalty could be considered as an oppressive form of violence because it is indiscriminately applied to certain racial and ethnic groups. Maintaining personal gun laws in our society also perpetuates the notion that a certain amount of violence is acceptable. The effects of discrimination such as racism, sexism, and classism on a person’s psychological development are overwhelming and serve to oppress certain groups while 12 uplifting others. Racism, sexism, and classism have been easily observed in the educational setting and, although illegal for the most part, still exist and are continually reproduced in our society (Young, 1990). Women of any color and anyone in poverty are subject to all these forms of discrimination. When adolescents witness the oppression and discrimination brought to bear on the adults in their life they form expectations of what the possibilities are for people in their social location. (See the anthropological study done by journalist Adrian Nicole LeBlanc, 2003, entitled Random Family, for discussion of teenagers’ expectations of life growing up in the Bronx). Again, theories of domination and oppression are applied in this research. Scml reproduction and labelinLtheories Pierre Bourdieu researched the relationship between higher education and social class structure. His thesis was that the purpose of education was to maintain, rather than reduce, inequality; that education actually “transmits privilege, allocates status, and instills respect for the existing social order” (Swartz, 1977, p 546). Bourdieu coined the term “cultural capital” to explain the processes through which cultural knowledge and cultural habits contribute to social inequality. At the level of the individual, cultural capital refers to a socially inherited “linguistic and cultural competence” that facilitates achievement in school. There is an unequal distribution of cultural capital based on the social location of one’s birth. Therefore, children arrive at school with different levels of cultural capital. Cultural capital is analogous to economic goods that are produced, distributed, and consumed by individuals and groups. Bourdieu believed that whether or not adolescents stayed in school depended upon their “perceptions of the probability that people of their social class would succeed l3 academically” (Swartz, 1977, p. 548). This belief led him to speculate that “there is a close correlation between subjective hopes and objective chances, the latter tending to effectively modify attitudes and behavior by working through the former” (Swartz, 1977, p. 548). Poor adolescents would then self-select out of the pool of possible college enrollments. They would stop aspiring to higher education because they would see no hope. Their expectations about the possibility of transcending their social class would be negative. Essentially one can see the “rebellion” of adolescence to conform to the status quo as cognitive dissonance; not desiring something that they believe they cannot obtain. Bourdieu discussed the personal or micro level characteristics that place an individual socially and mediate their chances at success. Societal or exo and macro level institutions play a role in determining an individual’s future also. Academic institutions, as a part of meso and micro level institutions, play a huge, active role in determining an individual’s educational expectations but because “academic success is perceived in terms of individual talent, effort, and merit, this link to social structures is hidden” (Swartz, 1977, p. 551), allowing society to place the blame at the individual or micro (familial) level. Rist (1977) explains the link between social structure and social reproduction by introducing the contributions of labeling theory in describing the process by which individuals are either encouraged or discouraged to transcend their social status. He states that “the labeling approach allows for an examination of what, in fact, is happening within schools” (Rist, in Karabel and Halsey, eds., 1977, pg. 293, italics in the original). According to Rist, labeling theory provides a framework with which to examine “the various evaluative mechanisms (both formal and informal) operant in schools, the ways in which schools nurture and support such mechanisms, how 14 students react, what outcomes are for interpersonal reaction based on how these mechanisms have evaluated individual students, and how, over time, the consequences of having a certain evaluative tag influence the options available to a student within a school” (p. 293, italics in the original). Rist uses labeling theory to explain the process of the concept of the “self- fulfilling prophecy.” In the classroom, the teacher initiates this process by expecting certain behaviors or abilities based on individual attributes, some of which are personal and some of which are societal. The teacher then treats students in certain ways, which, if consistent over time and the student does not actively resist this treatment, will shape the student’s behavior to conform to the teacher’s expectations. The teacher and the student “move toward a pattern of interaction where expectations are clearly communicated and the behavioral response is consonant with the expected patterns” (p. 301) All three types of theories discussed above are important for providing the framework of understanding the process of expectation development. Ecological systems theory provides the basic framework for interactive development that can change as environments change. Adolescents who have negative expectations while attending a poverty-stricken school system may develop very different expectations if they move to a different school system (see Kozol, 1991, for discussion of effects of different school systems due to funding). Theories of domination and oppression, along with social reproduction and labeling theory provide a multidimensional explanation for how students are affected by larger societal attitudes toward difference. This research essentially applies all three theories by examining the intersection of race, class, gender, and family structure and the effects on expectation development. 15 EMPIRICAL LITERATURE REVIEW The summary of empirical literature will start with a review of the expectations literature and then discuss the effects of poverty on adolescent emotional well-being and expectations and the effects of living in a single parent family on adolescent emotional well-being and expectations. Discussion of the research is further analyzed using the theoretical frameworks discussed above. Expectations The literature on the development of expectations in adolescence can be divided into several broad categories and encompass micro and macro level structures or characteristics. Themes that emerge when reviewing the literature are the traditional categories of race, socioeconomic status, and gender plus family structure, parental relationships, and other environmental factors. The studies can also be divided into two types, those that describe the process of the development of expectations and those that explore factors that influence the development of either positive or negative expectations. 80% reproduction theory and thwocess of expectation develoment Social reproduction theory states that certain structures in society are designed to keep the status quo, to keep those from the working class in the working class. The American dream ideology includes the belief that anyone can transcend their social class if only they work hard enough. We are taught to believe that education is the quick ladder up, that education enhances social mobility (Bourdieu in Swartz, 1977; MacLeod, 1995; Wilson & Wilson, 1992). However, only a handful of people transcend their social class (Bourdieu in Swartz, 1977). The question is, why, if education is so readily available to insure social mobility. The following articles examine this question in light 16 of race and social class. The articles examine the process of social reproduction by examining the process of expectations and aspirations for achievement. Aspirations and expectations were studied to ascertain the reasons for lower achievement among certain groups of adolescents. Exploratory, qualitative methods (MacLeod, 1987; Gibbs & Bankhead, 2000) and descriptive, quantitative empirical methods (Cook, Church, Ajanaku, Shadish, Kim, & Cohen; Mickelson, 1990; Yowell, 2002) examine the process of social reproduction among the adolescents studied. The studies explain the development of expectations as a sociological process within an ecological framework. They point out that expectations are a part of the adolescent’s experience and interaction with their world. Their world includes micro, meso, exo, and macro structures. MacLeod (1995) tested the theory of social reproduction in a qualitative study of two groups of boys, one predominately African American, the other predominately white, in the same housing project in an eastern US. city. His aim was to discover structural and cultural “mechanisms” that contributed to the reproduction of one’s social status. MacLeod asked the adolescents questions about occupational aspirations to uncover a “mediating link between socioeconomic structures (what society offers) and individuals at the cultural level (what one wants)” (p. 22) to research the role aspirations play in social reproduction. He found a racial difference between the levels of aspirations; African Americans had higher occupational aspirations than the white teenagers. The African American teens appeared to accept the ideology of the American dream while the white teenagers seemed to accept the reality of their social class and didn’t have higher aspirations than white, adult males of their social group. 17 In a similar research study, Mickelson (1990) studied the attitude-achievement paradox among African American high school seniors by asking questions about abstract (ideological) and concrete (reality-based) attitudes toward education. While African American teens professed a strong belief in the ideology that education leads to social mobility, their concrete attitudes about what they can expect from education were much lower. The African American students from low-income families had higher discrepancy scores between abstract and concrete attitudes, indicating a high acceptance of the dominant ideology but expectations based in reality. The white students’ scores were less discrepant between the two types of attitudes, leading one to believe that they believed that their aspirations were relatively attainable. Yowell (2002) also tested discrepancies between what one wants and what one expects in order to explain Latino students’ high aspirations for educational and occupational attainment but high dropout rate from school. Yowell believes that students of color experience aspirations and expectations differently due to structural forces such as racism and discrimination and that these different experiences serve to regulate school dropout. Gibbs and Bankhead (2000) discussed the feelings that structural forces such as racism and discrimination create and found that pervasive feelings of hopelessness, anger, and alienation were associated with high rates of unemployment and lack of economic development in south central Los Angeles for African American youth. This qualitative-quantitative study used 17 focus groups and 32 individual interviews to gather data. The qualitative responses indicated that lack of jobs combined with living in poverty and residential segregation provided an apartheid atmosphere that “reinforce(d) 18 their perceptions of lack of opportunity and reduce(d) their motivation to complete high school or pursue higher education” (p. 17). Cook, Church, Ajanaku, Shadish, Kim, and Cohen, (1996) examined occupational expectations and aspirations between two groups of boys in grades 2, 4, 6, & eight. The boys were from four elementary schools and two junior high schools in the city of Memphis, Tennessee. The schools were low-income and predominately African American or middle-income and predominately white. Cook, et al., found that there was a large discrepancy between what jobs boys wanted and what jobs the boys expected among the low- income and largely African American boys. The findings were very similar to Mickelson’s findings in that the gaps between abstract attitudes (ideology) and concrete attitudes (proximal world of family and neighborhoods) were much larger for boys from lower income families. Armstrong and Crombie (1999) studied the process by which adolescents’ compromise or change their occupational aspirations and expectations by looking at aspiration-expectation discrepancies over time. The adolescents were questioned beginning at grade eight and then again in grades nine and ten. The authors used traditional gender roles and socioeconomic status of a given occupation to examine this compromise. The results of the study showed that adolescents do, indeed, make considerable changes between their aspirations and expectations over time. The adolescents in the study who reported discrepancies between aspirations and expectations changed in the direction of their expectations. While this study examined changes between what one wanted and what one saw as possible in their future, limiting the variables to traditional gender roles and socioeconomic status leaves questions 19 unanswered with respect to why adolescents show a change in discrepancy scores over time. Changes in expectations are “assumed to represent changes in adolescents’ assessment of the occupations that are realistic and acceptable firture options” (p. 95) but are these changes towards realistic and acceptable options based on personal issues or structural barriers? Pa and McWhirter (2001) researched perceived influences on adolescents’ career expectations. They found that perceptions influence behavior and that an adolescent’s perception of chances in life are influenced by background, personal, and environmental variables. Gender and ethnicity are key variables that influence the contextual environment of learning experiences and the feedback that individuals receive about their possibilities in life. In Paa and McWhirter’s study, where the students were predominately European-American (88% white, 3% Latino, 3% Asian, 2% African American, 2% other), race or ethnicity was not perceived as a barrier. Specific influences on expectation development Recent studies have focused on specific variables that influence expectation development. The variables are family influences, racial and ethnic differences, gender differences, and other environmental concerns. Family influences Rhea and Otto (2001) researched mother’s influences on adolescent educational outcome beliefs, specifically whether family structure affected educational outcome beliefs. They measured adolescents’ expectations of educational attainment and compared those with family structure, mothers’ educational level, family income, and the “connection” between adolescents and their mothers. “Connection” was measured by 20 agreements and discussion regarding college and occupational careers and life in general. Their findings refute “deficit” theories of family structure. Family composition did not have an effect on adolescent expectations. The results did however, “provide preliminary evidence for the importance of distinguishing between the effects of family structure and family dynamics and taking family processes into account when estimating family effects” (p. 502). It’s not the type of family but the quality of family interaction that counts. Jodl, Michael, Malanchuck, Eccles, and Sameroff (2001) discussed parenting processes and parents’ roles in shaping younger adolescents’ occupational aspirations. This study moved research from structural features of the family to family processes by investigating “the pathways linking parental values, beliefs, and behaviors to adolescents’ occupational aspirations” (p.1247). How much time and effort were invested by mothers and fathers in the area of academics and sports were compared to career desires. Results revealed that parents spend less time at this age on academics but more time on sports, either because parents expect their children to be able to do their homework at this age by themselves or parents know less about how to help them. Parents’ behaviors based on their attitudes and beliefs did not affect occupational aspirations at this age. However, fathers’ behaviors had a partial effect on athletic aspirations. Trusty (1998) also discussed family influences on adolescent educational expectations but focused on adolescents who were two years post high school graduation. His study tested educational expectations against adolescents’ perception of parents’ behavior regarding, first, how much time parents spent talking to their teens about school and careers, second, who decided on career paths for the adolescent, third, how close the 21 family was emotionally, and, fourth, parent-reported school involvement behavior. The adolescents were questioned two years prior to asking about educational expectations. Correlations between adolescent educational expectations and the independent variables were all positive but socioeconomic status was correlated most highly. Trusty assumed a causal relationship between adolescents’ perceptions of parents’ behavior, parents’ self- reported school behavior and educational expectations. Sanders, Field, and Diego (2001) examined adolescent expectations and achievement by researching parental influences and substance use among adolescents. Empirical findings indicated that adolescents who had a close relationship with their mother were found to have a higher interest in school achievement and higher expectations. Adolescents who abused alcohol and other drugs had decreased academic expectations and achievement. Bandura, Barbaranelli, Caprara, and Pastorelli (2001) studied a network of sociocognitive influences on children’s career aspirations. Two of the sociocognitive effects tested were parents’ beliefs about their children’s possibilities and socioeconomic status. The authors found that socioeconomic status had only an indirect effect on children’s perceived occupational efficacy. If parents believed that they could play a part in raising their children’s choices, the higher their expectations and aspirations for their children. The authors tested parents’ perceived efficacy in promoting career paths for their children and how their perceptions influenced the children. The authors discussed possible mediating factors in parents’ efficacy as perceived opportunity structures and social and institutional impediments. They found that the higher the family’s 22 socioeconomic status, the stronger the parents’ belief in their own efficacy to promote their children’s education. Racial and ethnic differences Qian and Blair (1999) studied racial and ethnic differences in educational aspirations using the concepts of human, financial, and social capital. They assumed that educational aspirations were significant predictors of eventual educational attainment and explored the factors which affected educational aspirations and whether those factors were different depending on the racial or ethnic identity of the adolescent. Human capital was defined as the attributes the parents could bring to bear on the adolescent; the most important attribute being the parent’s educational level. Financial capital referred to financial resources available for the adolescents. Social capital, in this study, referred to the “strength of the relationship between parents and children.” (p. 606). The physical presence of adults in the adolescents’ lives were considered as a part of social capital. Qian and Blair found that educational aspirations varied by racial and ethnic group when explaining differences in social capital, i.e., family structure. Single parent families, in this study, did not significantly affect educational aspirations for adolescents of color but white students from single parent families had higher educational aspirations that those from intact families. This study only included high school seniors. This could explain the contradictory finding related to white students from single-parent families having higher educational expectations; single-parent families may affect younger adolescents more negatively (p. 616). Qian and Blair concluded that aspirations are “part of a set of social psychological processes” and “that the extent to which an adolescent believes that he or she should attain a higher level of education will directly affect the drive, 23 motivation, and effort that they put toward the achievement of that goal” (p. 617). The authors also found that social mobility was affected by one’s racial or ethnic background and that opportunities were structured differently for each group. Gender differences Hubbard (1999) focused on gender as an area of oppression and examined differences in African American male and female teens and their aspirations for college. The aspirations of the thirty high school students varied along gender lines; boys believed that sports would get them into college and girls believed that academics would. Hubbard (1999) used social reproduction theory, expanded to include gender, to argue that the differences in academic achievement among low-income Afiican American males and females could be explained by an intersections approach that included analyzing the effects of race, class, and gender on constructed expectations of educational outcomes. Environmental influences Environmental factors that were considered to influence expectations were involvement in health risk behaviors, living in poor, urban neighborhoods, school and home relationships and being a teen mother. Harris, Duncan, and Boisjoly (2002) evaluated the “role of ‘nothing to lose’ attitudes” (p. 1005) that lead to engaging in health risk behaviors based on adolescents’ expectations of their future and found that adolescents who view their future as negative are more likely to engage in “nothing to lose” behavior. Adolescents who felt that their educational prospects were positive were less likely to engage in risky behaviors: they felt they had more to lose. The authors found that environmental factors that promote 24 poor interpersonal factors, such as poor emotional health and depressive symptoms “diminish youths’ aspirations and hopes for their futures, reducing perceived risk associated with unhealthy or dangerous behaviors” (p. 1033). Pandey and Zhan (2000) challenged the “culture of poverty” theory by examining parent’s expectations of their children’s success across neighborhood type in urban areas. Parents’ expectations for their children’s success varied only for the age at which they expected that their children would have their own children. The parents expected success for their children regardless of the parents’ own poverty level. Pandey and Zhan’s findings support the structural perspective of poverty that states that the lower achievement of children is a reflection of lack of opportunities and that underclass behavior is an outcome of external factors not individual attributes (Pandey and Zhan, 2000). Regardless of neighborhood type, parents had high expectations for their children. Gibbs and Bankhead (2000), in studying reactions to the police beating of Rodney King, the acquittal of the police officers, and subsequent civil unrest in South Central Los Angeles, found that employment for Afiican American youth and opportunities to enhance their social mobility were severely limited by residential segregation. The youth identified five employment related concerns that influenced their expectations for their future. Lack of jobs in their neighborhoods, with lack of transportation to get to outside jobs, employer discrimination based on race, ethnicity, and address of the adolescent, employer preference for Afiican American females, few jobs with adequate wages and benefits, and lack of job training programs were listed as barriers to change in their life circumstances. Qualitative responses revealed rational decisions for getting involved in illegal activities for making money. 25 Wilson and Wilson (1992) were also interested in environmental influences affecting adolescent educational expectations that focused on family/home environment and the school environment. The sample included high school seniors living in two- parent families who were predominately white. The majority had low educational aspirations. Wilson and Wilson examined the amount of support for higher educational attainment given by both the home and the school environment. Their findings supported theories of reproduction of social class. Students whose parents had high levels of education were more likely to aspire to high levels of education. If parents had low levels of education but high aspirations for their children’s education, high parental aspirations did not mean higher aspirations by their children. The students were influenced only by the level of education that their parents actually achieved. Interaction between the home and school environments also affected student aspirations. Students who perceived mutual support from home and school for higher educational attainment most often internalized those values. Farber (1989) studied the educational and vocational aspirations of teen mothers and found that an individual teen’s aspirations could not be separated from the context of their environment. She found that how people view their lives and their opportunities are an important factor in how they live their lives. Teen moms from the lower class had high aspirations but didn’t appear to possess the wherewithal to achieve those aspirations. They had a passive acceptance of what they saw around them and doubted that it could be any different for them. The lower class teens lived and were educated in an environment where lack of hope, no sense of personal mastery, and a lack of perception that any other 26 possibilities existed for them led to negative expectations despite having high educational and vocational aspirations. The above research studies discuss expectation development and factors that influence the development of expectations. All studies discussed these influences from a race, class, gender, or family environment perspective. The following studies will discuss more specifically the influence of poverty and family structure on socio- emotional development. The influence of family income and family structure on the socio-emotional development of adolescents This section of the literature review will discuss empirical studies regarding the influence of family income and family structure on adolescents’ social and emotional development. Family income and family structure are highly correlated. Most studies discuss both issues together, but there are specific themes in the way each study addresses these issues. The first set of articles will discuss poverty as the primary focus, the second set will discuss family structure as the primary focus, and the third set will discuss family process. Em Smith, Brooks-Gunn and Klebanov (1997) explored consequences of living in poverty and found that it is highly associated with family structure (p. 135). They tested the hypothesis that income, more than family structure, accounts for a substantial portion of the negative consequences associated with single parent families. In all measures, when controlling for family income, the effects of family structure on lowered achievement scores disappeared. 27 Conger, Conger, and Elder (1997) studied economic hardship and adolescent emotional adjustment and found that cognitive development is affected in poor children and that living in poverty affects adolescents’ academic and emotional adjustment. Boys from economically disadvantaged families at mid-adolescence will have limited educational and occupational success, which will threaten their abilities as traditional breadwinners (p. 308-309), which in turn threatens stability for their children. They conclude that “to place children in seriously deprived economic circumstances creates enormous social risks” (p. 309). Guerra, Huesmann, Tolan, Van Acker, and Bron (1995) studied how economic disadvantage is linked to an increased risk for engaging in aggressive behaviors. The authors studied almost two thousand Afiican American, Hispanic, and Caucasian urban elementary boys and girls over a two-year period. They found that stressful events and positive beliefs toward aggression were influenced by a “pervasive sense of hopelessness” and that “when expectations through conventional channels are low, individuals who believe that life is hopeless may turn to more aggressive means of obtaining immediate rewards” (p. 519). Teachman, Paasch, Day, and Carver (1997) examined the link between living in poverty during adolescence with high school completion and college enrollment. Their obvious findings were that “children from the poorest families received less education, while children from the most prosperous families received more” (p. 413). IQ scores had no mediating effect on educational outcomes in this study. Children from families in the middle- income range varied less compared to other middle-income children, in their educational attainment. Middle-income families may 28 produce high school graduates but college attendance is wholly dependant on extra available income. Therefore, with a shift in governmental support for education, only those adolescents from families with considerable disposable income could afford to send their high school graduates to college without taking out student loans. The social and psychological effects of poverty on adolescents’ motivation to complete high school follows as an area of future research (p. 414). Another interesting finding from this research was that IQ had no mediating effect on educational outcomes. The social or exo environment of neighborhoods and schools “are likely to structure the opportunities and constraints perceived by adolescents as they make decisions about the relevance of continued education” (p. 415). Corcoran and Adams (1997) explored how and why poor children become poor adults. They tested four alternative explanations for poverty; 1) parents’ lack of economic resources which limits children’s resources; 2) parent’s non-economic resources, i.e., their own level of education and connection to education, family break up or non-marriage leading to psychological distress, or IQ level and genetics; 3) welfare dependency causing negative work attitudes; and, 4) structural/environmental models emphasizing labor market conditions and discrimination. Their study of 5000 included almost equal numbers of black and white, men and women, aged 25 to 35 years old in 1988. The respondents had been aged five to 15 years in 1968 and were from families who participated at that time in the Panel Study of Income Dynamics with 5000 family respondents. Results of their study supported the structural/environmental models. Findings were that “young black men’s and young black women’s adult income 29 sufficiency and adult poverty were strongly tied to the unemployment rates in their adult labor market areas” (p. 514). Axinn, Duncan, and Thornton (1997) studied whether parental income predicted completed schooling and children’s self-esteem as young adults. Consistent with other studies, they found that low-income children (those with family income below twice the poverty line) completed at least one year less of schooling than children from families above that threshold. In regards to self-esteem, parental income appeared to have little effect on their children’s self-esteem in early adulthood. Pagani, Bourlice, Vitaro, and Trembley (1999) examined the effects of poverty on academic failure and delinquency in boys. They found that “financial disadvantage reduces youngsters’ ability to compete in the school environment” (p. 1216), and that poverty was related to academic failure and extreme delinquency regardless of family configurations. Family structure Biblarz and Raferty (1999) examined the effects of alternative family structures on children’s educational and occupational success through a review of studies done over a thirty-year period. They found that for each time period of the previous studies, the relationship between family structure and child outcomes changed depending on which aspects of the respondents’ family backgrounds were used. They found support for social reproduction theory. “Children from low socioeconomic origins tend to end up in low socioeconomic destinations regardless of family type” (p. 30). They also found support for theories of domination and oppression. Single mothers had higher rates of lower- status occupational positions. The ecological systems theory was also supported in their 30 finding that fathers provided an important link to the public sphere by their employment outside the home and by their higher occupational status. Li and Wojkiewicz (1992) updated previous research of the effects of family structure on status attainment. They found that living in a mother-only family had a negative effect on a child’s educational attainment. Remarried or blended families had the same negative impact on educational attainment, refuting popular rhetoric that single or divorced mothers getting married would solve the problems associated with children in single-mother families. Their results showed that disruption in families had more of an effect on children’s education than never married families. Garis (1998) tested the assumption that increases in single parent families and children living in poverty increased the chance of engaging in risky behaviors by youth. His sample was a cross-sectional analysis of high school seniors whose family background characteristics were established in eighth grade. He tested this hypothesis by examining whether risky behaviors decreased as family income increased. His conclusion was that at-risk behavior increased when family income increased and that single parent families did not increase at-risk behaviors. Income had an indirect but unclear effect. What mattered most was the relationship with both parents. When non- custodial parents remained involved, at-risk behavior was less likely. Policy implications regarding monetary support were determined to be not an important goal unless “higher incomes and more secure employment for fathers facilitates involvement with children after divorce” (p. 1100). Environment and family process 31 Battistich, Solomon, Kim, Watson, and Schaps (1995) examined school context to test Merton’s (1957) determinants of alienation. Merton’s theory states that people suffer from alienation when their perception of specific, positive goals are not applicable to them or their group. Economically disadvantaged parents and students accept educational values and goals but perceive them as relatively unattainable. Thus, a resistance group may be formed when the educational setting doesn’t fulfill their needs for belonging. The authors found that a student’s motivation is enhanced in schools where they experience a positive, supportive community. Hanson, McLanahan, and Thomson (1997) examined the influences of income on parental practices and found that “income generally increases children’s welfare and that household debt is negatively related to a number of well-being indicators for children in one-parent and two-parent households” (p. 219). They also found that parental practices were only weakly related to income level meaning that both poor and non-poor households could include both effective and non-effective styles of parenting. McLeod and Shanahan (1993) explored poverty, mother’s parenting styles, and children’s mental health issues and found that how long a child lives in poverty makes a difference in their mental health. Children in persistent poverty exhibited more internalized symptoms of mental health issues while those experiencing current poverty displayed more external symptoms. They also found that mothers’ marital status did not “consistently predict children’s mental health” (p. 355). Astone and McLanahan (1991) also studied parenting styles and family structure but looked at high school completion rates as the outcome. Adolescents from single parent and step-parent families received much less support and encouragement for school 32 achievement than those from intact families. Living in a single-parent family or a step- parent family produced the same negative effects on educational attainment when socioeconomic status was controlled. This study focused primarily on family structure and what the different types of structures could offer children with little emphasis on how income might play a role. Research questions Empirical studies of the effects of poverty and family structure on children and adolescents primarily discuss outcomes based on being raised in low-income families or by a single mother. This research study is designed to examine the effects of being raised in or near poverty and of living with a single mother or father on adolescents’ current expectations of their future. While positive or negative expectations can be looked at as an outcome of these economic and structural situations, expectations motivate behavior. They are not just an outcome; they influence the future. From the previous studies and by predicting from the theories examined, we know that being raised in poverty affects a child’s physical and emotional development and that the likelihood for being raised in poverty is increased if you live in a single, female- headed household. Expectations are created from experience. Given this knowledge, we can examine which factors contribute to the development of either positive or negative expectations of the future for adolescents. This research explored the theories of ecological systems, domination and oppression, and social reproduction and labeling by examining the effects of race-ethnicity, income level, gender, and family structure, individually and collectively on adolescent expectation development. It builds on the 33 previous research on poverty and family structure by examining how these factors influence expectation development. Specifically this research asks, do race/ethnicity, income, family structure, and gender affect adolescents’ expectations of their future? Operationalized research questions are: 1. Do adolescents who live in mother-only families have more negative and less positive expectations than those who live in intact, two-biological- parent families? Do adolescents who live in father-only families have more negative and less positive expectations than those who live in intact, two-biological- parent families? Do adolescents who live in mother-only families have more negative and less positive expectations than those living in father-only families? Do adolescents raised in low- income families have more negative and less positive expectations than those in families with higher incomes? Do adolescents who are racial or ethnic minorities have more negative and less positive expectations than those who are white? Do female adolescents have more negative and less positive expectations than male adolescents? 34 CHAPTER 3 METHODOLOGY This chapter describes the study background and original sample, the specific sample used for this research, dependent and independent variables and how they are measured, reliability and validity information for the measures, research hypotheses, and analysis approach. Study background and sample This research uses a selected sample of respondents from the National Longitudinal Survey of Youth 1997 (NLSY97), Round 1. The NLSY97 survey is the newest national survey collected by the National Longitudinal Surveys - Bureau of Labor Statistics and is one in a set of six surveys of the National Longitudinal Surveys, begun in the mid 19605. The purpose of this particular survey is to document the transition from school to work and to assess the impact of schooling and other environmental factors on that transition. The data can be used to determine how youth experiences relate to developing a career (US Department of Labor, 2000). The total sample was drawn from a screening of 75,291 households at two different times. The first sample (n = 6,748) was cross-sectional and was designed to be representative of the United States population in 1997 of youth born between 1980 and 1984 (13 — 17 years old). The second sample (n = 2,236) was a supplemental sample of Black and Hispanic youth in the same age range. The purpose of the supplemental sample was to allow analysis across and within racial and ethnic categories. For the total sample for Round 1, 8,984 respondents were initially chosen from 6,819 households (U .S. Department of Labor, 2000). 35 A computer assisted personal interviewing system was used to gather data from any youth in the household who fell within the age group (13 - 17), and one parent, usually the mother, who resided in the same household. The youth questionnaire focused on schooling and employment activities, family background, interactions with non- resident parent, social behavior and health status. Youth who were fifteen or older (n = 3,565) by the end of the previous calendar year, were also asked about expectations for their future. The expectations questionnaire asked about the probability of certain events occurring in their lives over the next year, by the age of twenty and by the age of thirty. Parents were questioned about their personal background, marital and employment history, the family in general, the responding youth’s life, and the family’s economic circumstances. A school survey was also completed on 7,500 schools in the areas where the youth respondents lived. The schools were asked detailed information on the characteristics of the school, staff, and student body. Information included things like general practices, graduation policies, school to work programs, length of school day, facilities available at school, gender, racial/ethnic composition of faculty, staff, and students, proportion of faculty with more than a BA degree, total enrollment of school, average daily attendance, number of reported incidences of delinquent activities, percent of 1996 graduates who took the SAT or ACT tests, average scores, and percent of 1996 graduates enrolled in college or entered Armed Forces. Participating youth were given a computerized version of the Armed Services Vocational Aptitude Battery with Interest Finder. The ASVAB consists of 12 separate tests and the Interest Finder has six sub- scales to determine occupational preference. In addition to these tests, all youth were 36 given achievement tests and the Peabody Individual Achievement Test (PIAT) Math Test (US Department of Labor, 2000). Table 1 presents the gender and racial breakdown of the cross-sectional and supplemental samples. 37 Table 1. NLSY97 Gender, Race/ethnicity, and Sample Assignment Boys % Girls % Totals Total sample 4,599 51.2 4,385 48.8 8,984 Cross-sectional 3,459 51.3 3,2 89 48.7 6,748 White non-Hispanic 2,443 51.7 2,281 48.3 4,721 Black non-Hispanic 537 49.9 542 50.3 1,079 Hispanic 469 51 450 49 919 Race/ethnicity missing 10 38.5 16 61.5 26 Supplemental 1,140 51 1,096 49 2,236 Black non-Hispanic 63 1 50.3 623 49.7 1,254 Hispanic 508 51.9 471 48.1 979 Race/ethnicity missing 1 33.3 2 66.7 3 Selected sample The specific sample for this research was drawn from the NLSY97. Participants were selected from among adolescents who were 15 or older by the end of 1996 and had completed the expectations questionnaire, 11 = 3,565. Only 8.1 % of the 3565 (n = 290) did not answer either the positive or negative expectations questions. The participants were either part of a nationally representative sample of teens across the United States or part of the supplemental sample that over-sarnpled Black and Hispanic youth. Table two lists the breakdown by race/ethnicity, gender, family structure and family income level for the selected sample. Positive and Negative Expectations: Dependent variables The dependent variables were continuous, interval-level variables and were assessed using two separate scales of positive and negative expectations. The positive and negative expectations were scored separately using the total percentage of either positive or negative expectations reported by each participant. They were then divided by the number of items in each scale, resulting in average numeric values ranging from 0- 1 for both scales. Table 3 lists scale items and Cronbach’s alpha score for positive and negative expectations. Reliabilig da_ta Fischoff, Parker, Bruine De Bruin, Downs, Pamgren, Dawes, and Manski (2000) tested the reliability and validity of the expectations questions by examining the expectations reported by the teen respondents for seventeen significant life events. 38 Table 2 NLSY97 Selected Sample: 15 & l6 Year-old Respondents by Family Structure, Race/Ethnicity, Income, and Gender Boys % Girls % Totals % of Total Family Structure Father only 67 55.8 53 44.2 120 3.4 Mother only 498 48.7 525 51.3 1023 28.7 2 Bio parents 1215 51.4 1151 48.6 2366 66,4 Missing 56 1.5 Totals 1780 50.7 1729 49.3 3565 100 Race] ethnicity White non- 909 51.7 Hispanic 933 5 1 49 1 842 Black non- 26.6 Hispanic 458 48.4 489 51.6 947 Hispanic 393 52.8 351 47.2 744 20.8 Missing 32 0.9 Totals 784 50.5 1749 49.5 3565 100 Income Low income 369 49.6 375 50.4 744 20.8 Higher income 915 50.5 897 49.5 1812 50.8 Missing 1009 28.4 Totals 1284 50.2 1272 49.8 3565 100 F ischoff, et al., analyzed the data not only for the overall picture of the respondents’ expectations, but also for the validity (accuracy) and reliability (consistency) of the expectations questionnaire. Fischoff, et al., believed that the initial 39 cross-sectional analysis allowed for comparison of overall predictions with similarly aged individuals. They reported the mean and median percentage of responses, the standard deviation and the percentages of responses at the 0%, 50%, and 100% level of probability. They also reported a statistical estimate of each particular event occurring in the general population. Rank order correlations were then reported for each question, correlating the respondents’ answers with previously asked questions regarding life experiences, asked elsewhere in the questionnaire. A factor analysis was used to test for sub-sets of items that might form discrete factors. They reported two significant factors, namely “physical risks” with 23% of the variance, and “career” risks with 15% of the variance. The participants’ answers were also compared to their parents’ responses and were found to have similar beliefs about the adolescents’ future and similar response- mode usage, i.e. proportionate use of % choices for likelihood of certain events. Overall, Fischoff, et al., found the beliefs of the teens sensibly related to other aspects of their life experiences. They list several results that increased the confidence of the measurement tool. For the firll set of 17 events, the R2 between judged and statistical probabilities is .87 for untransformed data and .93 for cubic transformation. Reliability analyses using the Statistical Package for Social Sciences (SPSS) were run and a Cronbach’s alpha computed for the expectations questions specifically for this research. Because I wanted to examine positive and negative expectations separately, two scales were created and tested. Table 3 lists the resulting scales and their alpha scores. After dividing the questionnaire into positive and negative expectations and computing the alpha for each, a decision was made to create a score for each scale based on the possible total percent if a participant reported 100% chance on each item. The 40 positive scale, with four items, was scored by dividing the total percent reported by 400, the total possible percentage. The negative scale, with nine items, was divided by 900. Positive and negative expectation scores thus range from 0 - 1 for each scale. The mean positive expectations score was .8769 with a standard deviation of .1619 and n = 3492. The mean negative expectations score was .2176, with a standard deviation of .1264 and n = 3348. Using the Pearson’s R correlation, the expectations scales had a two-tailed significance of less than .001 and were negatively correlated at -.178. Although the correlation is a very weak one, it is still significant. Family Structure. Race. Income Level, and Gender: Independent variables The independent variables used were family structure, income to poverty ratio, gender, and race/ethnicity. All were categorical and were scored as follows. F_amily Structure Family Structure was divided into three categories for this research, namely biological father-only families, biological mother-only families, and two biological parent families. Biological father-only families were scored with “0", biological mother- only with “1 ", and two biological parent families with “2". Children living in other family configurations were excluded from this study. Other dynamics may be present when a child is being raised by someone other than a biological parent. Thus, adopted children, children in step-parent families, children raised by grandparents, uncle, or aunt, or children in foster care are not part of this research. 41 Table 3. Positive and Negative Expectations Positive Negative % chance work 20+ hours, % chance in school next year not in school next year % chance get high school diploma % chance pregnant or get someone by age 20 pregnant next year % chance college degree by age 30 % chance get drunk next year % chance work 20+ hours per week % chance victim of violent crime by age 30 next year % chance arrested next year % chance die next year % chance jail by age 20 % chance parent by age 20 % chance die by age 20 Chronbach’s alpha = .65 Cronbach’s alpha = .71 Income Level A ratio of family income to the poverty line for a family the size of the respondent’s family, the income to poverty ratio, was re-coded into a dichotomous variable. This represents the family’s relationship to the poverty line. In this data set the mean family income was over $47,000. The median was $38,000. Because the poverty line is so low, I used the low-income standard that the federal government uses to determine eligibility for free school lunches. Students in families with incomes above 130% of the poverty line are not eligible for free lunches according to the federal guideline. Families that had incomes at 130% of the poverty line or below are eligible for free school lunches. They are considered low income and coded as “0". Families with incomes at or above 131% of the poverty line were considered higher income and coded with “1". A large number of cases (n = 1009 or 28.3%) had missing values in the income category. I computed a cross tabulation of the variables and income score to see if any 42 one group had a higher proportion of missing cases than others. The Black, non-Hispanic group and the Hispanic group had a disproportionate number of missing data on income compared to the White, non-Hispanic group. Thecross tabulation reported that the percent of missing income data was 24.1% for the White non-Hispanic group, 33.5% for the Black non-Hispanic group, and 32.4% for the Hispanic group. Race/ethnicity The NLSY data has three variables that ask about a respondent’s race or ethnicity. One variable specifically asks if the respondent is Hispanic, one variable asks respondents to identify their race and it uses the categories of white, black, American Indians, including Eskimos and Aleuts, Asian or Pacific Islander, and “something else”. The third variable combines race and ethnicity and uses the categories of Black, Hispanic, mixed race/non-Hispanic, and Non-Black, non Hispanic. Because the respondents who identified themselves as American Indian, Eskimo, or Aleut, and Asian or Pacific Islander were a relatively small number, (n = 221), I left them out of the study. The three most represented race and ethnicity groups in the NLSY97 are White non- Hispanic, Black non-Hispanic, and Hispanic. These three groups make up the race/ethnicity categorical variable and are coded as 0 = White non-Hispanic, 1 = Black non-Hispanic, and 2 = Hispanic. Gender Gender is a categorical variable coded as 1 = boys and 2 = girls. Mediating variables Variables that were used for testing mediation effects were: 1) “teachers are good”, representing attitudes towards school, 2) “% of peers belonging to a gang”, 43 representing peer characteristics, 3) “grades received in eighth grade”, representing school achievement, and, 4) “biological father is supportive or not”, representing relationship with non-custodial father. The family structure, race/ethnicity, and gender variables were coded into dummy variables for use in the regression models. “Teachers are good” was coded as 1 = strongly agree, 2 = agree, 3 = disagree, and 4 = strongly disagree. For the variable, “Percent of peers belonging to a gang”, respondents chose from the following categories: 1 = < 10%, 2 = 25%, 3 = 50%, 4 = 75%, 5 = > 90%. “Grades received in eighth grade was coded as 1 = mostly below Ds, 2 = mostly D5, 3 = half Cs, half D3, 4 = mostly Cs, 5 = half Bs, half Cs, 6 = mostly Bs, 7 = half As, half Bs, 8 = mostly As. All mediating variables were ordinal Reseaich design and analysis approach This is a secondary analysis of cross-sectional data using a selected sample of the respondents from the National Longitudinal Survey of Youth, cohort 1997, Round 1. Using the selected sample from the NLSY97 data for youth 15 or older and who answered the expectations questionnaire, I tested six specific hypotheses regarding the effects of income, family structure, race/ethnicity, and gender on the development of either positive or negative expectations. One-way and two-way ANOVAs were computed to determine if any significant differences existed between groups that were compared and to determine if any interaction effects existed. Post hoc tests were used to determine which groups differed significantly from the other groups when the overall F value was significant in the one-way AN OVAs. Research hypotheses 44 The following is a list of research hypotheses proceeding from the research questions suggested by the empirical literature review and theory. Alternative Hypothesis #1 Adolescents living in biological mother-only families will report more negative and less positive expectations regarding future life events than adolescents living I n two- biological-parent families. Alternative Hypothesis #2 Adolescents living in biological father-only families will report more negative and less positive expectations regarding future life events than adolescents living in two-biological-parent homes. Alternative Hypothesis #3 Adolescents living in biological mother-only families will report more negative and less positive expectations regarding future life events than adolescents living in biological father-only homes. Alternative Hypothesis #4 Adolescents living in families in the low income range will report more negative and less positive expectations regarding future life events than adolescents living in families in higher income ranges. 45 Alternative Hypothesis #5 Adolescents who report their race or ethnicity as Black non-Hispanic or Hispanic will report more negative and less positive life events than adolescents who report their race as White non-Hispanic. Alternative Hypothesis #6 Female adolescents will report more negative and less positive expectations regarding future life events than males. In addition to testing the above hypotheses, I addressed the following questions in order to examine possible mediating factors: 1. Will the adolescent’s attitude towards school teachers mediate the relationship between income and positive or negative expectations? Will the adolescent’s family income level mediate the effects of race or ethnicity on the development of positive or negative expectations? Will gang members among the adolescent’s peer group mediate the effect of income level on the development of positive or negative expectations? Will family income level mediate the effect of family structure on the development of positive or negative expectations? Will the adolescent’s grades received in eighth grade mediate the effect of income on the development of positive or negative expectations? 46 CHAPTER 4 RESULTS This chapter discusses the results of each statistical analysis performed to test the impact of race, gender, income level, and family structure on positive or negative expectations. Statistical analyses were examined for each research question and hypothesis presented in chapters 2 and 3. Research questions and hypotheses are presented prior to the results of each specific analysis. One-way and two-way analysis of variance tests (ANOVA) are used to compare the means of expectation scores of the youth in the specific categories of independent variables . “Analysis of variance attempts to find significant differences between groups (or populations) by comparing the means of those groups on some variable of interest” (George & Mallery, 2003, p. 155). The variables of interest in this study are positive and negative expectations, and the groups compared are: 1) family structure, consisting of biological mother-only, biological father-only, or two biological parents; 2) race and ethnicity, consisting of White non-Hispanic, Black non-Hispanic, or Hispanic categories, 3) family income, consisting of low income (at or below 130% of poverty) and higher income categories, and 4) gender, consisting of boys and girls. family structurefi and expectations Research questions: 1. Do adolescents who live in biological mother-only families have more negative and less positive expectations than those who live in two biological parent families? 2. Do adolescents who live in biological father-only families have more 47 negative and less positive expectations than those who live in two biological parent families? 3. Do adolescents who live in biological mother-only families have more negative and less positive expectations than those living in biological father-only families? Hypothesis #1 Adolescents living in biological mother-only families will report more negative and less positive expectations regarding future life events than adolescents living in two-biological- parent families. Hypothesis #2 Adolescents living in biological father-only families will report more negative and less positive expectations regarding future life events than adolescents living in tvvo-biological-parent homes. Hypothesis #3 Adolescents living in biological mother-only families will report more negative and less positive expectations regarding firture life events than adolescents living in biological father-only homes. A one-way ANOVA was performed to test the impact of family structure on positive expectations. Family structure is measured as a categorical variable with three groups, namely, biological father-only, biological mother-only, and two biological parents in the household. The overall F score is significant at the .05 level. A post hoc 48 Sheffe test was then rim to determine which groups differed in their means. The Scheffe test indicated that there was a significant difference between biological mother-only and two biological parent household groups. The significant F score (see Table 4) tells us that there is indeed a difference between the three family structure categories and the average positive expectation score. The post hoc Scheffe test reveals that the significant difference lies between biological mother-only and two biological parent families. The mean positive expectation score for biological mother-only households was .8458 while the mean for two biological parent households was higher at .8936. The adolescents living with both biological parents had more positive expectations than those living in biological mother-only families. The null hypothesis that states there is no difference between the average positive expectation score of biological mother-only families and two biological parent families can be rejected. A second one way ANOVA was performed testing the impact of family structure on negative expectation scores. The overall F score indicates a statistically significant difference between group means at the .01 level. A post hoc Sheffe test again indicates the differences existed between biological mother-only and two biological parent families. The mean negative expectations score for biological mother-only families was .2371 and the mean for two biological parent families was lower at .2089. This means that adolescents from biological mother-only families had more negative expectations than adolescents from two biological parent families. The null hypothesis can be rejected because the Scheffe test indicates that adolescents in mother-only families have more negative and less positive expectations for their future than adolescents living in two biological parent families. Adolescents in biological father-only families do not have 49 significantly different expectations from either biological mother-only or two biological parent families for both positive and negative expectations. These two tests indicate no significant difference. See tables 4 and 5 for results. Table 4. One-Way ANOVA: Family Structure and Positive Expectations Source of Variation N Mean SD F df Total 3440 .8791 .1587 32.319* 2. 3437 Father only 1 19 .8743 .1496 Mother only 996 .8458 .1820 2 bitmarent 2325 .8936 .1457 * p <.001 Table 5. One-Way ANOVA Family Structure and Negative Expectations Source of Variation N Mean SD F df Total 3300 .2174 .1263 16.822* 2, 3297 Father only 118 .2214 .1131 Mother only 950 .2371 .1381 2 bio parent 2232 .2089 .1207 "' p < .001 Income and expectations Research question: 4. Do adolescents raised in low-income families have more negative and less positive expectations than those in families with higher incomes? 50 Hypothesis #4 Adolescents living in families in the low-income range will report more negative and less positive expectations regarding future life events than adolescents living in families in higher income ranges. A one-way AN OVA was performed testing the impact of low income (measured as at or below 130% of the poverty line: the point of eligibility for free school lunches) on positive expectations. The overall F score was significant at the .01 level. A second one-way ANOVA testing the impact of low income on negative expectations was then performed. The overall F score was also significant at the .01 level. Adolescents living in low-income families had a mean of .8220 and those living in families with higher incomes had a mean of .8999 on the positive expectations scale. This means that the adolescents living in low-income families had significantly lower positive expectations than those living in higher income families. For the negative expectations scale, adolescents living in low-income families had a mean of .2322. Adolescents living in higher income families had a mean of .2139. The significantly lower mean on positive expectations and significantly higher mean on negative expectations allows us to reject the null hypothesis for hypothesis. Low income has an impact on adolescent expectations, although the difference between low-income and higher-income categories is small. See tables 6 and 7 for results. 51 Table 6. One-Way ANOVA: Income and Positive Expectations Source of Variation 11 Mean SD F df Total 2524 .8775 .1621 125.523* 1, 2522 low income 727 .8220 .1884 high income 1797 .8999 . 1443 *p < .001 Table 7: One-Way ANOVA: Income and Negative Expectations Source of Variation 11 Mean SD F df Total 2436 .2191 .1269 10.381* 1, 2434 low income 692 .2322 .1427 high income 1744 .2139 .1 197 *p < .001 Rye/ethnicity and expectations Research question: 5. Do adolescents who identify themselves as Black or Hispanic have more negative and less positive expectations than those who identify themselves as white non-Hispanic? Hypothesis #5 Adolescents who report their race or ethnicity as Black non-Hispanic or Hispanic will have more negative and less positive life events than adolescents who report their race as White non-Hispanic. 52 A one-way ANOVA testing the impact of race/ethnicity on positive expectations was performed. There is a significant difference between the mean positive expectation scores of adolescents in the three racial/ethnic categories. A post hoc Sheffe test indicated significant differences between each pair of groups. The white non-Hispanic group average positive expectation score was .8931. Black non-Hispanics scored .8747 and the Hispanic group averaged .8345 on positive expectations for their future. A second one-way ANOVA tested the impact of race/ethnicity on negative expectations. The ANOVA shows a significant difference among the means of the three groups being compared. The post hoc Sheffe test revealed the significant difference in average negative expectations to be between White non-Hispanics and Hispanics. The mean for white non-Hispanics was .2120, and for Hispanics, it was .2289. The null hypothesis of no difference in mean positive expectation scores reported by adolescents of varying race and ethnic groups can be rejected. For negative expectations, the overall, null hypothesis is rejected but the post hoc test for a difference between white non- Hispanics and Black non-Hispanics is not significant. See tables 8 and 9 for results. Table 8. One-Way ANOVA: Race/Ethnicity and Positive Expectations Source of variation 11 Mean SD F df Total 3360 .8755 .1624 33.711* 2, 3357 White non- Hispanic 1718 .8931 .1419 Black non- Hispanic 921 .8747 .1683 Hispanic 721 .8345 .1906 *p < .001 53 Table 9. One-Way ANOVA: Race/Ethnicity and Negative Expectations Source of variation n Mean SD F df Total 3220 .2183 .1263 4882* 2, 3217 White non- Hispanic 1676 .2120 .1 187 Black non- Hispanic 860 .2220 .1312 Higanic 684 .2289 .1368 *p < .008 Gender and expectations Research question: 6. Do adolescent girls have more negative and less positive expectations than adolescent boys? Hypothesis #6 Adolescent girls will report more negative and less positive expectations regarding firture life events than adolescent boys. A one-way ANOVA testing the impact of gender on positive expectations was performed. The difference between the means of boys and girls would have been significant with a p-value of less than .001, but was the opposite of my prediction. Therefore, the null hypothesis cannot be rejected. Girls actually have higher average positive expectation scores than boys in this data set. A second one-way ANOVA was performed testing the impact of gender on negative expectations. Again, the difference would have been significant with a p-value of less than .001 but it was in the opposite direction of my hypothesis. Girls scored lower than boys did on negative expectations. The alternative hypothesis that girls would have 54 more negative and less positive expectations for their future is not supported. The difference between means is opposite to my prediction. See tables 10 and 11 for results. Table 10. One-Way ANOVA: Gender and Positive Expectations Source of variation n Mean SD F df Total 3492 .8769 .1619 23.171* 1, 3490 Boys 1762 .8639 .1648 Girls 1730 .8902 .1578 *p < .001 Table l 1. One-Way AN OVA: Gender and Negative Expectations Source of variation It Mean SD F df Total 3348 .2176 .1264 20.871 * 1, 3346 Boys 1691 .2275 .1318 Girls 1657 .2076 .1 198 p < .001 Two-Way ANOVAS to Examine Main and Interaction Effects Next, two-way AN OVAs were performed on each hypothesis with separate tests exploring the impact of combinations of independent variables on positive and on negative expectations. Family Structure and Income Family structure is a categorical variable with three levels, biological father-only, biological mother-only, and two biological parent families. Income level is a categorical variable with two levels, low income and higher income. A two-way ANOVA to test the 55 impact of family structure and income level on positive expectations for the future was computed. Income level has a main effect for income on positive expectations. Family structure and income level together show an interaction effect, but there is not significant “family structure” effect. However, the Eta2 statistic for income was .006 and for family structure by income was .009. Expressed as a percentage, family structure by income accounted for 0.9% in the variance of positive expectations. These are small effects. Although post hoc comparisons of specific cells are not appropriate with two-way ANOVA it is interesting to note that, adolescents in two biological parent, higher income families had the highest average positive expectations. This was to be expected. However, adolescents in two biological parent, low-income families had the lowest positive expectations for their future. This is contrary to the popular belief that adolescents living in single, mother-only families would have lower positive expectations. A second two-way ANOVA to test the impact of family structure and income level on negative expectations reveal a main effect for family structure but not for income. There was no interaction effect for family structure and income on the development of negative expectations. The Eta2 for the main effect of family structure was extremely small at .008 or 0.8% of the variance explained. The corrected model Eta2 was still small at .015, or 1.5% of the variance explained. Family structure has a main effect on negative expectations. The Eta2 was small at 008, or 0.8% of the variance explained. See tables 12 and 13 for results. 56 Table 12. Two-Way ANOVA: Family Structure, Income Level, and Positive Expectations Source of variation Father only Mother only 2 bio parent Totals .8630 .8300 .8130 Low income n = 25 n = 363 n = 336 n = 724 .8790 .8620 .9130 Higher income n = 67 n = 358 n = 1372 n = 1797 Totals N=92 n=721 n= 1708 N=2521 Income F = 13985", Family Structure by Income F = 11.322“, *p < .001 Table13 Two-Way ANOVA: Family Structure, Income Level, and Negative Expectations Source of variation Father only Mother only 2 bio parent Totals .2066 .2420 .2222 Low income 11 = 24 n = 334 n = 319 n = 677 .2280 .2422 .2060 Higher income n= 67 n= 345 n= 1318 n= 1730 Totals N = 91 n = 679 n = 1637 n = 2407 Family Structure F = 10257“, *p < .001 Family Structure and Race/Ethnicity A two-way ANOVA tests the impact of family structure and race/ethnicity on positive expectations. Both family structure and race/ethnicity have a main effect on positive expectations. There is no interaction effect between the two variables on positive expectations. The Eta2 scores for family structure and race/ethnicity main effects were .014 and .005 respectively. The corrected model score was .036, explaining 3.6% of the variance in positive expectations. For negative expectations, the picture was similar. Controlling for race/ethnicity, there was a main effect for family structure on negative expectations. The Eta2 for the 57 main effect of family structure was .007, with a corrected model score of .013, explaining only 1.3% of the variance. See tables 14 and 15 for results. Table 14. Two-Way ANOVA: Family Structure, Race/Ethnicity, Positive Expectations. Source of variation Father only Mother only 2 bio parent Totals White non- .8704 .8507 .9068 Hispanic n=62 n=321 n=1319 n=1702 Black non- .9179 .8629 .8883 Hispanic n = 29 n = 455 n = 419 n = 903 .8377 .8010 .8540 Hispanic 11 = 23 n = 198 n = 486 n = 707 Totals N= 114 n=974 n=2224 n=3312 Race/Ethnicity F = 23.916*, Family Structure F = 7.717*, *p < .001 Table 15. Two-Way AN OVA: Family Structure, Race/Ethnicity, Negative Expectations Source of variation Father only Mother only 2 bio parent Totals White non- .2313 .2336 .2052 Hispanic n=62 n=314 n= 1284 n=1660 Black non- .1807 .2297 .2164 Hispanic 11 = 29 n = 424 n = 392 n = 845 .2376 .2545 .2187 Hispanic n = 22 n = 190 n = 459 n = 671 Totals N=113 n=928 n=2135 n=3176 Family Structure F =11.781*, *p < .001 Family structure and gepder A two-way ANOVA tests the effects of family structure and gender on positive expectations for the future. Family structure and gender each had significant main effects but there was no interaction effect. Although the F scores were significant, the Eta2 scores for family structure and gender were .019 and .005, respectively. For the corrected model, the Eta2 score was .028, representing a very small amount of variance explained. The two-way ANOVA test of the impact of family structure and gender on negative expectations produced similar results; both variables had significant main effects 58 but there was not a significant interaction effect. Even smaller Eta2 scores were obtained. The corrected model only explains 1.7% of the variance in negative expectations. See tables 16 and 17 for results. Table 16. Two-Way ANOVA: Gender, Family Structure, Positive Expectations Source of variation Male Female Totals .8395 .9177 Father only n=66 n= 53 n= 119 .8366 .8546 Mother only 11 = 485 n = 511 n = 996 .8786 .9095 2bioparent n= 1193 n- 1132 n=2325 Totals n = 1744 n = 1696 ' n = 3440 Family Structure F = 33.401*, Gender F = 16.562*, *p < .001 Table 17. Two-Way ANOVA: Gender, Family Structure, Negative Expectations Source of variation Male Female Totals .2278 .2133 Father only n=66 n= 52 n=118 .2480 .2267 Mother only n = 463 n = 487 n = 950 .2191 .1982 2 bio parent n = 1145 n = 1087 n = 2232 Totals n = 1674 n = 1626 n =3300 Family Structure F = 17.543*, Gender F = 5.064“ , *p < .001, "p < .05 Race/ethnicity and Income The two-way AN OVA test of the impact of race/ethnicity and income level on positive expectations reveals a significant main effect for both race/ethnicity and income and a significant interaction effect for race/ethnicity and income together on positive future expectations The Eta2 score for the main effect of race/ethnicity was .009. For income it was .032. And, for the interaction effect of race/ethnicity and income, the Eta 2 was .004. For the corrected model the Eta2 score was .060, explaining 6% of the 59 variance. A surprising post hoc observation was that the mean for low income Black non-Hispanic families for positive expectations was higher than the white non-Hispanic group, .8529 and .8133, respectively. This is consistent with Mickelson’s (1990) study and MacLeod’s (1995) study, both of which found black teenagers having higher expectations than the white teens in their studies. The two-way AN OVA using negative expectations as the dependent variable, however, showed no main effect for race and no interaction effect of race/ethnicity and income but a significant main effect of income on negative expectations. The Eta2 for the main effect of income was .002 with a corrected model of .006. See tables 18 and 19 for results. Table 18. Two-Way ANOVA: Race/Ethnicity, Income Level, Positive Expectations Source of White non— Black non- Hispanic Totals variation Hispanic Hispanic .8133 .8529 .7939 Low income 11 = 183 n = 275 n = 255 n = 713 .9081 .8913 .8643 Higher income n = 1146 n = 341 n = 239 n = 1726 Totals n=1329 n=616 n=494 n=2439 Race/Ethnicity F = 10.639*, Income Level F = 79.420*, Race/Ethnicity by Income Level F = 4.972", * p < .001, "p < .01 Table 19. Two-Way ANOVA: Race/Ethnicity, Income Level, Negative Expectations Source of White non— Black non- Hispanic Totals variation Hispanic Hispanic .2323 .2308 .2350 Low income n = 181 n = 253 n = 244 n = 678 .2109 .2163 .2293 Higher income n = 1124 n = 323 n = 227 n = 1674 Totals n = 1305 n = 576 n = 471 n = 2352 Income Level F = 4.941*, p < .03 60 Gender and income A two-way ANOVA testing the impact of gender and income level on positive expectations was performed. Both gender and income level had main effects on positive expectations but there was no interaction effect. Eta2 scores were .008 for gender main effect and .048 for income level. The corrected model score was .059. For negative expectations, there was an interaction effect as well as main effects for both variables. Eta2 scores were: income level, .005, gender, .007, and the interaction of income level and gender, .002. The corrected model was .011. In post-hoe observations, low-income boys had the highest average score for negative expectations at .2502. Low-income girls and higher income boys had identical means on negative expectations, with a mean of .22. Higher income girls had the lowest average negative expectations score of .2075. Income appears to effect low-income girls and high-income boys similarly when it come to negative expectations of the future. See Tables 20 and 21 for results. Table 20. Two-Way ANOVA: Gender, Income Level, and Positive Expectations Source of variation Boys Girls Totals .8112 .8325 Low income n = 359 n = 368 n = 727 .8803 .9198 Higher income n = 906 n = 891 n = 1797 Totals n= 1265 n= 1259 n=2524 Income Level F = 127.946*, Gender F = 19.361*, p < .001 61 Table 21. Two-Way ANOVA: Gender, Income Level, and Negative Expectations Source of variation Boys Grrls Totals .2502 .2151 Low income 11 = 338 n = 354 n = 692 .2201 .2075 itigher income n = 883 n = 861 n = 1744 n=1221 n=1215 n=2436 Totals Income Level F = 11.004*, Gender F = 1768]”, Income Level by Gender F = 3.930" *p < .001, “p < .05 Mediating variables Five additional questions were posed for this research study in order to examine the impact of possible mediating variables. Mediating variables account for the relation between the predictor and the criterion variable. (Baron & Kenny, 1986). Using multiple regression allows for the impact of independent variables on dependent variables to be tested for the mediation of intervening variables. The process uses three regression equations to examine: first, the effect of the independent variable on the potential mediating variable, second, the effect of the independent variable on the dependent variable, and, third, the effect of the independent on the dependent, controlling for the effect of the mediator on the dependent variable. If the potential mediator fully mediates the relationship between the predictor and the criterion variable, then the effect in the third equation must be smaller than in the second equation. If the first relationship was statistically significant, then it should no longer be significant in the third equation, if there is full mediation. If all relationships are significant, then only partial mediation exists between the variables. (Baron & Kenny, 1986). For example, the slope in the second equation for the independent variable must have a significant t value. In the third 62 equation, if the mediator fully mediates, then the slope for the independent variable will be smaller and the t value no longer significant. Question #1 Will an adolescent’s attitude towards school mediate the relationship between family income level and positive or negative expectations? Attitude towards school is measured by using the variable “Teachers are Good”. Respondents were to report on their agreement with this statement by choosing from the following Likert scale: 1 = strongly agree, 2 = agree, 3 = disagree, and 4 = strongly disagree. Higher scores reflect that the respondent feels that teachers are not good. Two groups of three regression equations were performed to examine whether attitudes towards teachers mediated the effect of income level on adolescents’ positive and negative expectations. For positive expectations, the first equation describes the impact of Income Level on Teachers are Good. The second equation describes the impact of Income Level on Positive Expectations. The third equation includes the independent variable, Income Level, and possible mediator, Teachers are Good, and describes their impact on the dependent variable, Positive Expectations. Comparing the results to the mediation requirements explained above, the independent variable, Income Level, affects the mediator, Teachers are Good, in the first equation, such that when Income Level increases, Teachers are Good goes down. As ‘61,, explained above, Teachers are Good is coded ranging from equaling the strongest or highest support of teachers up to “4,” signifying the strongest disagreement with the statement “teachers are good”. Income Level and Teachers are Good have a negative or 63 inverse relationship. Adolescents in higher income families are less likely to disagree with the statement “teachers are good”. Equation two regressed Positive Expectations on Income Level. The relationship strength between Positive Expectations and Income Level has an adjusted R2 of .047, meaning that Income Level predicts 4.7% of the variance in Positive Expectations. The standardized beta for Income Level is .218 and had a significant t value. Because the standardized beta for Income Level was positive, this means that when families have a higher income, Positive Expectations scores are likely to be higher. The third equation regressed Positive Expectations against both Income Level and Teachers are Good. The standardized beta for Income Level is lower (.218 in the second equation to .210 in the third) and the t value remains significant. This indicates that attitude towards school, as measured by the variable, Teachers are Good, does not fully mediate the effect of Income Level on Positive Expectations. For negative expectations, equation two (Table 23) found a standardized beta of -.065 for Income Level. For those adolescents from families with higher incomes there is a lower negative expectation score. For equation three, the standardized beta for Income Level is -.058 and the t value remains significant. The slope decreased in the third equation but remains significant so, attitudes towards school do not fully mediate the effects of income on negative expectations. See Tables 22 and 23 for results. 64 Table 22. Multiple Regression: Mediating Effects - Attitude Toward School, Income Level, Positive Expectations Equation 1 B Std. Error Std. Beta t sig. Constant 2.096 .022 93.308 .001 Income Level -8.845E-02 .027 -.066 -3.317 .001 Adjusted R2 = .004, F = 11.000, p < .001, dependent variable = teachers are good Equation 2 B Std Error Std. Beta t sig. Constant .822 .006 140.045 .001 Income Level 7.793E-02 .007 .218 1 1.203 .001 Adjusted R2 = .047, F = 125.513, p < .001, dependent variable = positive expectations Equation 3 B Std. Error Std. Beta t sig. Constant .91 l .012 76.001 .001 Income Level 7.412E-02 .007 .210 10.919 .001 Teachers are good -4.225E-02 .005 -.161 -8.388 .001 Adjusted R2 = .074, F = 101.052, p < .001, dependent variable = positive expectations Table 23. Multiple Regression: Mediating Effects - Attitude Toward School, Income Level, Negative Expectations Equation 1 B Std. Error Std. Beta t Sig Constant 2.096 .022 93.308 .001 Income Level -8.845E-02 .027 -.066 -3.317 .001 Adjusted R2 = .004, F = 11.000, p < .001, dependent variable = teachers are good Equation 2 B Std. Error Std. Beta t sig. Constant .232 .005 48.230 .001 Income Level -1.833E—02 .006 -.065 -3.222 .001 Adjusted R2 = .004, F = 10.381, p < .001, dependent variable = negative expectations 65 Equation 3 B Std. Error Std. Beta t sig. Constant .170 .010 16.970 .001 Income Level -1.637E-O2 .006 -.058 -2.898 .004 Teachers are good 3012E-02 .004 .144 7.166 .001 Adjusted R2 = .024, F = 31.240, p < .001, dependent variable = negative expectations Qpestion 2 Will the adolescent’s family income status mediate the effects of race/ethnicity on the development of positive or negative expectations? Two groups of three regression equations were performed using dichotomous variables for each minority racial or ethnic category. Black and Hispanic were each re- coded to reflect two “dummy regression” variables. For “Black”, the Black non-Hispanic 661,, adolescents were coded as and all other race/ethnicity categories were coded as “0”. For the “Hispanic” group, Hispanic adolescents were coded as “1” and the other categories were coded as “0”. In the first equation Black and Hispanic, as two separate dummy variables, were regressed against Income Level. Again, Income Level is a dichotomous variable where “0” equals at or below 130% of the poverty level used to determine eligibility for free or 661,, reduced school lunches and equals 131% and above. The adjusted R2 was .094, indicating that the two predictor variables together accounted for 9.4% of the variance in Income Level. The second equation with Black and Hispanic regressed against Positive Expectations produces a standardized beta for Black of -.040, with a significant t value, and a standardized beta for Hispanic of -. l 56, also with a significant t value, indicating that Positive Expectation scores decrease when respondents are Black or Hispanic. In the third equation, when adding the mediating 66 variable, both standardized betas decreased (Black =.004 and Hispanic = -. 127). The t value for Black is no longer significant but for Hispanics, the t value is significant. This indicates that family income, as measured by Income Level, fully mediates the effect of Race/ethnicity for Blacks but only partially mediates the effect of race/ethnicity for Hispanics on positive expectations. Three regression equations also explored possible mediating effects of income status on the relationship between racial and ethnic categories and negative expectations. The standardized betas for both Black and Hispanic decreased in equation three, from equation two. For Blacks, the t value was not significant at .05 for either equation two or three. For Hispanics, the t value remains significant after adding the mediating variable. Income does not mediate the effects of race/ethnicity for Blacks on negative expectations. Income has a partial mediating effect for Hispanics on negative expectations. See Tables 24 and 25 for results. Table 24. Multiple Regression: Mediating Effects — Income Level, Race/ethnicity, Positive Expectations Equation 1 B Std. Error Std. Beta t sig Constant .811 .011 73.556 .001 Black -265 .021 -.253 -12.920 .001 Hispanic -3 36 .028 -.236 -12.036 .001 Adjusted R2 = .094, F = 128.656, p < .001, dependent variable = income level Equation 2 B Std. Error Std. Beta t sig Constant .889 .004 248.131 .001 Black -1 .449E-02 .006 -.040 -2.268 .023 Hispanic -7.580 e-02 .009 -.156 -8.902 .001 Adjusted R2 = .022, F = 39.640, p < .001, dependent variable = positive expectations 67 Equation 3 B Std. Error Std. Beta t sig Constant .834 .007 1 16.1 10 .001 Black 1.507 e-03 .008 .004 .195 .846 Hispanic -6.408E-02 .010 -.127 -6.141 .001 Income Level 6.915E-02 .007 .194 9.412 .001 Adjusted R2 = .061, F = 54.050, p < .001, dependent variable = positive expectations Table 25. Multiple Regression: Mediating Effects — Income Level, Race/ethnicity, Negative Expectations Equation 1 B Std. Error Std. Beta t sig. Constant .811 .011 73.556 .001 Black -.265 .021 -.253 -12.920 .001 Hispanic -.336 .028 .236 -12.036 .001 Adjusted R2 = .094, F = 128.656, p < .001, dependent variable = income level Equation 2 B Std. Error Std. Beta t sig. Constant .213 .003 74.564 .001 Hispanic 8.780E-O3 .005 .031 1 .700 .089 Black 2.085E-02 .007 .055 3.056 .002 Adjusted R2 = .003, F = 5.188, p < .01., dependent variable = negative expectations Equation 3 B Std. Error Std. Beta t sig. Constant .227 .006 38.481 .001 Hispanic 2.046E-02 .009 .052 2.393 .017 Black 4.196E-O3 .006 .014 .658 .510 Income Level -1.474E-02 .006 -.053 -2.443 .015 Adjusted R2 = .005, F = 52850, p < .001, dependent variable = negative expectations Question 3 Will the adolescent’s peer group mediate the effect of income level on the development of positive or negative expectations? The adolescent’s peer group characteristics are measured by using the variable, % of Peers Belonging to a Gang. Respondents chose from a range of possible percentages 68 as follows: 1 = < 10%, 2 = about 25%, 3 = about 50%, 4 = about 75%, and 5 = > 90%. For equation one (Table 26), the possible mediator, % of Peers Belonging to a Gang was regressed on Positive Expectations. The standardized beta for Income Level is -.227 and had a significant t value. When % of Peers Belonging to a Gang is high, the family Income Level for those adolescents is likely to be lower. The adjusted R2 was.051, indicating that 5.1% of the variance in peer group characteristics was predicted by Income Level. Equation two regressed Positive Expectations on Income Level. The. standardized beta for Income Level was .218 and had a significant t value, indicating higher average Positive Expectation scores among adolescents in higher income families. In equation three, the standardized beta for Income Level decreased to .192 but remains significant. The decrease is too slight to indicate full mediation but because there was some effect, % of Peers Belonging to a Gang has only a partial mediating effect on the relationship between income and positive expectations. For Negative Expectations the decrease in the standardized beta for Income Level in the third equation is larger and the t value is no longer significant, indicating that % of Peers Belonging to a Gang has a mediating effect on income and negative expectations. See Tables 26 and 27 for results. Table 26. Multiple Regression: Mediating Effects-Peer Characteristics, Income Level, Positive Expectations Equation 1 B Std. Error Std. Beta t sig Constant 2.045 .037 55.039 .001 Income Level -.515 .044 -.227 -1 1.679 .001 Adjusted R2 = .051, F = 136.392, p < .001, dependent variable = % peers belonging to a gang 69 Equation 2 B Std. Error Std. Beta t Sig Constant .822 .006 140.045 .001 Income Level 7.793E-02 .007 .218 1 1.203 .001 Adjusted R2 = .047, F = 125.513, p < .001, dependent variable = positive expectations Equation 3 B Std. Error Std. Beta t Sig Constant .864 .009 100.33 1 .001 Income Level 6.834E-02 .007 .192 9.670 .001 % Peers in gang -2.090E-02 .003 -.133 -6.706 .001 Adjusted R2 = .066, F ’= 88.347, p < .001, dependent variable = positive expectations Table 27. Multiple Regression: Mediating Effects - Peer characteristics, Income Level, Negative Expectations Equation 1 B Std. Error Std. Beta t siL Constant 2.045 .037 55.039 .001 Income Level -.515 .044 -.227 -1 1.679 .001 Adjusted R2 = .051, F = 136.392, p < .001, dependent variable = % of peers belonging to a gang Equation 2 B Std. Error Std. Beta t sig Constant .232 .005 48.230 .001 Income Level -1.833E-02 .006 -.065 -3 .222 .001 Adjusted R2 = .004, F = 10.381, p < .001, dependent variable = negative expectations Equation 3 B Std. Error Std. Beta t sig Constant .187 .007 26.542 .001 Income Level -6.891E-03 .006 -.025 -1 .193 .233 % Peers in _gang 2.266E-02 .003 .183 8.882 .001 Adjusted R2 = .035, F = 44.886, p < .001, dependent variable = negative expectations 70 Question 4 Will family income mediate the effect of family structure on the development of positive or negative expectations? Two dummy regression variables, Father-only and Mother-only , are used as the predictor variables and Income Level as the potential mediating variable. Equation one regressed Income Level against the two predictor variables and obtained a standardized beta for Father-only of -.031 and for Mother-only, -.306, indicating that with Father or Mother as the predictor the standardized beta for Income Level would decrease by -.031 and -.306, respectively. Having a single parent reduces income. For equation two, Positive Expectations is regressed against the two family structure indicators. The standardized beta for Father-only is -.022 and for Mother-only, -.137. The Positive Expectation score decreased when Mother-only families was the predictor variable. The t value for F ather-only is not statistically significant. In equation three, the standardized beta for Mother-only became less negative, .072, and remained significant. Partial mediation only occurred for Mother- only families. For negative expectations, the t value for Father-only families is not significant in equation two or three and the standardized beta for Mother-only increased very slightly from .101 to .108, when income level is controlled. Income level does not mediate the effects of family structure on negative expectations. See Tables 28 and 29 for results. 71 Table 28. Multiple Regression: Mediating Effects — Income Level, Family Structure, Positive Expectations Equation 1 B Std. Error Std. Beta T sig Constant .803 .010 76.976 .001 Mother- _ only -.307 .019 -.306 -16.015 .001 Father-only -7.502E-02 .046 -.03 1 -1 .625 .001 Adjusted R2 = .082, F = 113.598, p < .001, dependent variable = income level Equation 2 B Std. Error Std. Beta T sig Constant .894 .003 274.065 .001 F ather-only -1.929E-02 .015 -.022 -1 .306 .192 Mother- only -4.783E-02 .006 -. 137 -8.033 .001 Adjusted R2 = .018, F = 32.319, p < .001, dependent variable = positive expectations Equation 3 B Std. Error Std. Beta T sig Constant .836 .007 120.851 .001 Father-only -1.372E-02 .017 -.016 -.822 .41 1 Mother- only -2.550E-02 .007 -.072 -3.507 .001 Income Level 7.065E-02 .007 .200 9.790 .001 Adjusted R2 = .053, F = 47.295, p < .001, dependent variable = positive expectations Table 29. Multiple Regression:- Mediating Effects - Income Level, Family Structure, Negative Expectations Equation 1 B Std. Error Std. Beta T sig Constant .803 .010 76.976 .001 F ather-only -7.502E-02 .046 .03 1 -1 .625 .104 Mother- only -.307 .019 -.306 ~16.015 .001 Adjusted R2 = .089, F = 128.297, p < .001, dependent variable = income level 72 Equation 2 B Std. Error Std. Beta T sig Constant .209 .003 78.507 .001 F ather-only 1.254E—02 .012 .018 1.056 .291 Mother- only 2.891E-02 .005 .101 5.790 .001 Adjusted R2 = .010, F = 16.822, p < .001, dependent variable = negative expectations Equation 3 B Std. Error Std. Beta T siL Constant .216 .006 37.701 .001 Father-only 1.261E-02 .014 .019 .930 .353 Mother- only 3.044E-02 .006 .108 5.062 .001 Income Level -8.47OE-03 .006 -.030 -l.418 .156 Adjusted R2 = .015, F = 13.537, p < .001, dependent variable = negative expectations Qpestion 5 Will the adolescent’s grades received in eighth grade mediate the effect of income on the development of positive or negative expectations? The variable, Grades Received in Eighth Grade, is used as an indicator of achievement level. “Grades Received” is coded on an eight-point scale, with “1” representing mostly failing grades up to “8” representing mostly “As”. Equation one regressed Grades Received against Income Level. The standardized beta for income level is .228. For every increase in Income Level, the value for the standardized beta for Grades Received increases by .228. In equation two, the standardized beta for Income Level is .209 with Positive Expectations as the criterion variable. In the third equation, the standardized beta for Income Level decreased to .147 when grades were controlled, and the t value remained significant. This indicates that Grades Received in Eighth 73 Grade did not fully mediate the effect of income on Positive Expectations but had a partial mediating effect. For negative expectations, equation two obtained a standardized beta for the effect of Income Level of -.089 (less negative expectations for higher-income teens). Equation three revealed a change to —.055 and the t value remained significant, indicating only partial mediation. See Tables 30 and 31 for results. Table 30. Multiple Regression: Mediating Effects - Grades Received in Eighth Grade, Income Level, Positive Expectations Equation 1 B Std. Error Std. Beta t sig Constant 5.284 .051 103.665 .001 Income Level 1.414E-O3 .000 .228 1 1.555 .001 R2 .052, F = 133.529, p < .001, dependent variable = grades received in eighth grade Equation 2 B Std. Error Std. Beta t sig Constant .842 .005 184.939 .001 Income Level 1.191E-04 .000 .209 10.721 .001 Adjusted R2 = .043, F = 114.937, p < .001, dependent variable = positive expectations Equation 3 B Std. Error Std. Beta t sig Constant .729 .010 71.397 .001 Income Level 8.001E-05 .000 .147 7.432 .001 Grades fl“ _grade 2.256E-02 .002 .257 12.961 .001 Adjusted R2 = .104, F = 140.690, p < .001, dependent variable = positive expectations Table 31. Multiple Regression: Mediating Effects - Grades Received in Eighth Grade, Income Level, Negative Expectations Equation 1 B Std. Error Std. Beta t sig Constant 5 .284 .051 103.665 .001 Income Level 1.414E-03 .000 .228 1 1.555 .001 Adjusted R2 = .052, F = 133.529, P < .001, independent variable = grades received 74 Equation 2 B Std. Error Std. Beta t sig Constant .23 1 .004 62.037 .001 Income Level -4.030E-05 .000 -.089 -4.410 .001 Adjusted R2 = .008, F = 19.446, p < .001, independent variable = negative expectations Equation 3 B Std. Error Std. Beta t sig Constant .297 .009 34.066 .001 Income Level -2.448E-05 .000 -.055 -2.654 .008 Grades - 8th grade -1.225E-02 .001 -. l 73 -8.295 .001 Adjusted R2 = .034, F = 42.419, p< .001, dependent variable = negative expectations The next chapter will discuss the major findings of this research and compare them with findings from the literature review, the conceptual framework of the study and implications for social work. 75 CHAPTER FIVE DISCUSSION In this study, the effects of race, class, gender, and family structure on the development of positive or negative expectations of an adolescent’s future are examined. This chapter discusses the major findings of the hypotheses tests and examine them in light of the conceptual framework previously discussed. It also discusses findings from additional analyses performed. Implications for social work education and future research needs will be discussed. To review, the conceptual which frames the research questions and hypotheses is as follows: Following from Bronfenbrenner’s Ecological Systems Theory (1979, 1989), l) adolescents interact with multiple ecological systems due to their age and developmental stage. Adolescents are students, workers, friends, volunteers, part of a social network, etc. Following from theories that attempt to explain the intersections of race, class, and gender, 2) adolescents are socially located in multiple settings based on their race, ethnicity, their socio-economic situation, or social class, and their gender. These social locations interact with their ecological systems to influence who they become. 76 Social reproduction and labeling theories, along with theories of domination and oppression help to explain how certain social constructions come to be developed and create situations that are considered less ideal, so that, 3) the family structure that an adolescent lives in is an ecological system that is part of the socially constructed context of American culture. As a part of that socially constructed context, the family in our society is affected by the larger political system and by our economic system of capitalism. The types of families discussed in this research are only three types of a myriad of forms that a family could take. Some of these forms are considered less desirable because of the greater risk for poverty and adverse conditions for the children raised in them. However, with a capitalistic economic system, wage structures have been created based on family types that don’t fit how society actually functions. Families have revolutionized and democratized based on social forces but single-female-headed households get blamed, individually for changes (Hays, 2003). The data used for this research includes biological parent families only: 120 biological father-only families, 1023 biological mother-only families and 2366 two-biological parent families. Again, from Bronfenbrenner’s Ecological Systems theory, 4) all of the above factors contribute to an adolescent’s life experiences. Futures are built from past experience and if those experiences are negative, an adolescent may develop negative expectations and set their sights low for future possibilities (Mickelson, 1990; Yowell, 2002; Gibbs & Bankhead, 2000). 77 Major findings Hypotheses l. 2, 3 Do adolescents who live in mother-only families have more negative and less positive expectations than those who live in intact, two biological parent families? Do they have more negative and less positive expectations those who live in father-only families? Do father-only families have more negative and less positive expectations than two biological parent families? The one-way AN OVA and post hoc tests performed for these questions do indicate a significant difference between mother-only and two biological parent families. Adolescents who lived in two parent biological families had significantly higher positive expectation scores than their counterparts living in mother- only families. And, adolescents living in mother-only families had more negative expectations than those living in two biological parent families. The null hypothesis regarding biological mother-only and two biological parent families is rejected for positive expectations. The F score was significant and the Scheffe post hoc test reveal the differences to be significant between mother-only and two biological parent families. The mean positive expectations score for adolescents living in mother-only families was .85; for two biological parent families it was .89. The means are not hugely different and are both very high. For negative expectations, the mean for adolescents in mother-only families was .24 and for two parent families it was .21. Again, there was a significant F score but the differences are small in reality. 78 There are no significant differences between two biological parent families and biological father-only families or between biological father-only and biological mother- only families so the null hypothesis is retained for questions two and three. Hypothesis 4 The next question asks if adolescents living in lower income families have less positive and more negative expectations than their counterparts living in higher income families. The obvious answer is yes and the one-way AN OVA tests show a statistical significance for this answer. Adolescents in low-income families had less positive and more negative expectations than those living in families with higher income levels, though the differences are small. The mean positive expectations score for adolescents in low-income families was .88 and adolescents in higher income families had a mean score of .90. For negative expectations the means were.23 for low income and .21 for higher income. Again, the differences are relatively small. The null hypothesis is rejected for question four. Hypothesis 5 Due to discrimination and oppression based on race or ethnicity the next hypothesis tested is that adolescents identified as racial or ethnic minorities would have lower positive and higher negative expectations than adolescents who identify as White non-Hispanic. The one-way ANOVA for positive expectations bears this out. All three racial and ethnic groups in this study had significantly different means from each other, such that White non-Hispanic adolescents had higher positive expectations than Black non-Hispanic adolescents, and Black non-Hispanic adolescents had higher positive expectations than Hispanic adolescents. Those most different from the mainstream 79 dominant culture of the white middle-class perceive less opportunity than others. Checking for primary language and customs that a family adheres to might explain some of this difference. The means for positive expectations for each group were as follows: White non-Hispanic, .89, Black non-Hispanic, .88, and Hispanic, .83. The null hypothesis is rejected for positive expectations because the F score was significant although the difference in means is small. For negative expectations, there is a significant impact of race/ethnicity but post hoc tests find a difference between White non-Hispanics and Hispanic adolescents only. Hispanics may experience the American culture differently due to closer ties with historical culture and language than many African Americans. The significant post hoc tests differences were between White non-Hispanic, .21 and Hispanic, .23. Hypothesis 6 A surprising finding of this study was that female adolescents had higher positive expectations and lower negative expectations than males. Women are clearly oppressed in our society by income and social status (Sands & Nuccio, 1989; Gordon, 1994; Lord, 1993; Abrarnovitz, 1996; Sidel, 1996; Hubbard, 1999; Mullings, 2001; Hays, 2003). For adolescents, girls may be more socialized than boys to fit into the role of student. Past research has looked at how education has possibly shortchanged girls by having lowered expectations for their potential performance in the math and science curriculum. The subsequent focus in education has been to remedy this by increasing math and science programs for girls. The expectations items used for this data are very limited in scope. The positive expectation questions had more to do with education and career. The negative had more 80 to do with risky behaviors. Boys may be more socialized to believe that expecting negative physical risks is part of their social identification as males. High school students support the idea of male bravado and toughness that the expectations address for boys. Possibly in a few years, the adolescent girls in this study would perceive their career opportunities as less positive than the males when they are in the world of work. If schools are designed to reproduce social class, as described by Bourdieu in Swartz (1977) and Bowles and Gintes (1976), and girls are considered second-class citizens, schools may shortchange them by socializing them to be obedient students. The more obedient the student, the more successful one is in many school systems. Although the ANOVA test reveals potentially significant differences between boys and girls and their positive and negative expectations, the differences were in the opposite direction of my hypothesis. Therefore, the null hypothesis is retained for question six. Tests for Main Effects and Interaction Effects Single-female-headed families are stigmatized in our society and blamed for any negative consequences that they may experience. One the one hand, our society supports women who stay at home with their children unless they are poor women without support from the father(s) of the children. The 1996 welfare reform legislation clearly forces women to “work first” and then worry about raising their children later. Most qualitative studies of women who use welfare support have found that poor women needing assistance don’t have different values or morals from the rest of society (see Duerr Berrick, 1995, Pelisser Kingfisher, 1996, and Stack, 1974 for qualitative studies that discuss women in poverty and using welfare). They want the same for their children that non-poor mothers do. Single parenthood is only devastating when that single parenthood 81 occurs in poverty ( Abramovitz, 1996; Sands & Nuccio, 1989; Sidel, 1996). The two- way ANOVA results revealed a main effect for income on positive expectations and an interaction effect between family structure and income. The mean positive expectations score for adolescents living in mother-only low-income families was higher than adolescents living in two biological parent, low-income homes, .830 and .813, respectively. Also, adolescents living in higher income mother-only families had higher positive expectations than lower income, two biological parent homes, .862, and .813, respectively. Therefore, when a single-female-headed family is not poor, the adolescents in that family had a higher mean for positive expectations than those in two parent families who were poor, and they also had higher positive expectations when both family types had low incomes. Income had a main effect but family structure did not. Although the effect size measure, Etaz, was small for income at .006, the corrected model increased to .062, or 6.2% of the variance explained by the main and interaction effects on positive expectations. For negative expectations, the differences in means by family structure are significant and there is no interaction effect with income. The negative expectations in this study were extremely negative. For example, risks included things like dying before the age of thirty, being in jail, and getting pregnant. Although these are realities for many adolescents, the negative expectation scores were very low overall for this group. The two-way AN OVA tests for family structure and race/ethnicity on positive expectations produced main effects for both family structure and race/ethnicity but no interaction effects. The Eta2 for the corrected model was .03 6, or 3.6% of the variance in positive expectations explained by the main effects of family structure and race/ethnicity 82 combined. For negative expectations, only family structure had a main effect, which was very small at 1.3% of variance explained. A two-way ANOVA was performed for family structure and gender with main effects for both family structure and gender on both positive and negative expectations. For positive expectations, the Eta2 was 2.8% for the corrected model and 1.7% for negative expectations. Again these results were significant but with very small effects. For race/ethnicity and income, there were main effects and interaction effects. The most interesting finding was that low income Black non-Hispanic adolescents had higher positive expectations than the other two groups. The Eta2 for the corrected model containing both main effects and the interaction effect for positive expectations was 6% of the variance explained. This result supports Mickelson’s (1990) study that showed that black adolescents had high abstract ideals about education but their concrete experiences resulted in lower academic achievement than their white counterparts. The results also support McLeod’s (1989) qualitative study of groups of boys in a housing project. The blacks had higher expectations than the white boys. The low income Black adolescents in this study had significantly higher positive expectations for their future than the whites or Hispanic teens. For negative expectations, low income Black non-Hispanics and low income Hispanics had the same mean. For gender and income on positive expectations, there were no interaction effects but main effects for both gender and income. The corrected model Eta2 score was .059 or 5.9% of the variance explained. For negative expectations, an interaction effect occurred between gender and income level as well as the main effect of gender. Males typically had more negative expectations all around but gender differences were greatest among 83 low-income adolescents. Therefore, poverty affected the negative expectations of girls in the same direction as boys, but low-income boys have the most negative expectations for their future. Gender and income also had main effects on negative expectations. The corrected Eta2 was low at .011 or 1.1% of the variance explained. Mediating variables Using multiple regression to test for mediating effects I examined an additional five questions. The possible mediating variables were all suggested by the literature. Sometimes attitudes towards school or school performance can alter the negative effects of low income for a student. Persons of color in higher income ranges experience the effects of racism differently due to their status in society. Hanging out with the “wrong” or “ri t” crowd can alter ones experiences and possibilities for the future. Qpestion 1 Will attitudes towards school mediate the effects of income on positive or negative expectations? The first regression shows a significant but weak relationship between income and attitudes towards school. Income only accounts for .4% of the variance in school attitudes. The relationship between income and positive expectations was significant but again it was weak. Income only accounts for 4.7% of the variance on positive expectations. The third equation shows only partial mediating effects for attitudes towards school because the variables are correlated. For negative expectations, the relationships between income and attitudes and income and negative expectations were even weaker. Again, since the variables are correlated, only partial mediation occurs. 84 Question 2 Will income mediate the effects of race/ethnicity on positive or negative expectations? Income fully mediates the effects of race/ethnicity for Blacks and partially mediates the effects for Hispanics. When income level is controlled in the regression equation, being Black no longer had a significant impact on reducing positive expectations. Being Hispanic maintained a significant relationship with positive expectations, but the standardized beta decreased when income level was controlled. Income had a partial mediating effect for Hispanics on negative expectations also. Question 3 Will the adolescent’s peer group mediate the effect of income on positive or negative expectations? For positive expectations, there is a slight decrease in the effect of income level when adding peer group gang membership to the regression equation, indicating partial mediation. For negative expectations, peer group gang membership fully mediates the effect of income. Low-income teens have more gang-member peers. Lots of gang peers predict more negative expectations. When peer gang membership is controlled, income level no longer has a significant impact on negative teen expectations. Percent of peers belonging to a gang is the significant explanatory variable for negative teen expeditions. The difference in the standardized beta for income between the second and third equations is larger and no longer significant, indicating full mediation. 85 Question 4 Will family income mediate the effects of family structure on positive or negative expectations? For biological mother-only families, income has a partial mediating effect on positive expectations. The standardized beta for mother-only families decreased but the t value remains significant. Biological father-only families did not have a significant impact on either positive or negative expectation scores. Income level does not mediate the impact of mother-only families on negative expectations. Question 5 Will grades received mediate the effect of income on positive or negative expectations? The standardized beta for the impact of income level decreased from the second equation to the third equation but its t value remains significant. This indicates that grades received has only a partial mediating effect on income level for positive expectations. For negative expectations, the result is similar. The standardized beta decreased in the third equation but the t value remains significant, indicating partial mediation. Discu_ssion of major findings The results of the statistical tests for hypotheses one through five support the theories and empirical literature discussed. Minority racial or ethnic background, low income level, and single-parent family structure are found to have a negative impact on future expectations for adolescents. Although the impact is small in this data set, statistically significant differences are found. 86 Tests of hypothesis nmnber six show an opposite result than that suggested by the literature and theories discussed in Chapter 2. Adolescent girls in this data set have higher positive and lower negative expectations than adolescent boys. This suggests that girls and boys experience the school setting differently. The questions on negative expectations were extreme and included risky behaviors. Girls are not socialized to expect extremely negative behavioral consequences. Further examination of negative consequences that girls might expect may have produced more negative expectations on the part of the girls in the study. Further analyses reveal more interesting and more in depth findings. When examining combinations of variables using the two-way ANOVA test, interaction effects reveal interesting relationships. An interaction effect is defined as “the idiosyncratic effect of two or more independent variables on a dependent variable over and above the independent variables’ separate (main) effects” (George & Mallery, 2003, p. 372). Income level has a main effect on positive expectations but family structure and income level together have an interaction effect that is peculiar to their individual effects. Adolescents in two biological parent, higher income families had the highest positive expectations of the three groups but because of the interaction effect of family structure and income level, the adolescents living in two biological parent, low-income families had the lowest positive expectations. This is contrary to what we would expect to find; the theory that two parents are better than one. Adolescents in biological mother-only, low-income families had higher positive expectations than adolescents in low-income two parent families. 87 Interaction effects were also found for race/ethnicity and income and for gender and income. For race/ethnicity, the interaction effect with income occurs for positive expectations. Low income Black adolescents have higher positive expectations than low income whites or Hispanics. For gender, the interaction effect occurs for negative expectations, with low—income girls and higher income boys having essentially the same average negative expectations. However, low-income boys had the most negative expectations. For variables that had significant main effects, the interesting differences occurred with family structure and race/ethnicity. From white non-Hispanic families and Hispanic families, adolescents living in two biological parent families have the highest positive expectations. However, for Blacks, adolescents living in biological father-only families have the highest positive expectations, though their numbers are small (n = 29). For negative expectations, Hispanic adolescents in mother-only families have the most negative expectations, followed by Hispanics in father-only families, and then whites in mother-only families. The 29 Blacks in father-only families have the lowest negative expectation score of all the groups. The father-only families are a very small group (n=120) of a total sample of more than 3000. Limitations The primary limitation of this study is the inability to directly question its respondents. This research analyzes a large secondary data set to examine the effects of race, income, gender, and family structure on the development of adolescents’ positive and negative expectations for the future. The questions asked of the respondents were not asked with the intention of looking at macro level issues affecting individuals. The 88 expectations questions were designed to elicit how an adolescent perceives their future opportunities. Without further elaboration, it’s hard to tell what respondents meant or why they chose a specific likelihood of a certain event occurring. What exactly were these adolescents basing their expectations on? This research suggests a smaller, in- depth qualitative study of adolescents and their expectations. The researcher could ask the respondents directly about how their expectations were developed and whether race/ethnicity, socioeconomic class, gender, or family structure influence how they see their future. Missing income information was a problem in this 1997 data set. Approximately one third of the income information was missing for each analysis. See appendix C for demographics of missing income information. Another issue of concern in this data set was that the Hispanic group was not categorized into country of origin or ancestral origin. Hispanics in America are very diverse representing a total population in the United States in 1997 of over 29 million people. The 1997 statistics report that out of that 29,703 million, over 18 million were of Mexican descent, over 3 million were of Puerto Rican descent, over lmillion of Cuban descent, over 4 million, Central or South American descent, and over 2 million from “other” areas. Sixty-seven% of Cuban Americans live in Florida, 60% of Puerto Rican Americans live in New England or the mid-Atlantic states, and the majority of Mexican Americans live in the west and southwest with some in the mid-west (US Census Data, 1997). The data in the current study is not differentiated, which misrepresents the diversity of Hispanics in the United States. The cultures of Mexico, Cuba, Puerto Rico, and other Central or South American countries are all different and the adolescents from 89 those cultures may have very different expectations than the other groups but this data set did not allow comparison across country of origin. The Latino National Political Survey, designed to elicit information on a variety of political, economic, and social issues, from Latinos themselves, portray the various culturally diverse thoughts from these groups (see de la Garza, DeSipio, Garcia, Garcia, & Falcon, 1992). This data set does not differentiate Hispanic adolescents from various ancestral origins. Implications for Social Work This research is informative for social work practice, theory, and policy. For social work practice this research discusses macro versus micro issues and can inform the education and practice of direct work. Social work education stresses micro and macro practice but the heavy emphasis in most social work education programs is on micro or direct practice with individuals, small groups, and families. Subsequently, fieldwork is most often in a direct practice setting. This research examines the effects of social categories — race/ethnicity, income level, gender, and family structure - on individual expectations. The results, although slight, indicate that focusing on these specific categories is important for the practitioner. In the area of theory and theory development as taught in Human Behavior in the Social Environment, this research can be used to inform the human behavior curriculum regarding an “intersections approach”. The social work human behavior curriculum describes populations of diverse people as if they exist in a vacuum. It talks little about intersecting oppressions. C. Wright Mills once commented that social workers have an “occupationally trained incapacity to rise above a series of ‘cases’” (Mills, as quoted in Brandon, 1979). Mills was writing in the late fifties and early sixties. Most human 90 behavior texts do not discuss the intersections approach as it is discussed in other disciplines such as sociology. Policy implications are rampant in this research. Expectations for the future are affected by family structure and poverty but poverty interacts with family structure in its effect on adolescent expectations. If families are poor, the impact of family structure is overwhelmed by the impact of poverty. Single parenthood, gender, and race/ethnicity are all linked to poverty but they are not the cause of poverty. As Sharon Hays writes, “(T)heir membership in these social groups is no accident, no mere historical footnote. Their membership in these social groups is, at this historical juncture, the primary basis of their poverty. The backdrop for their poverty is persistent gender and race inequalities (Hays, 2003, p. 128). With the 1996 welfare legislation, low-income people have once again been punished and blamed for their condition, as if poverty exists outside of the system and belongs entirely to the individual. The “work first” requirement of the PRWOR Act of 1996 and the 2002 amendments requiring that recipients must work forty hours at any job they can get or be assigned to public service for no wages, places the blame squarely on the individual and doesn’t discuss labor market issues. Policy advocacy in the area of welfare and economic reform is an on-going need. Political advocacy for the rights of human citizenship is an on-going need. Implications for future research: Although this data set is very rich in detail, a qualitative study with a small group of adolescents that would have a chance to explain why they believe a certain event would happen to them, would give more insight into the development of positive or negative expectations. Qualitative responses could explain why they believe as they do. It would allow testing of theories of oppression and discrimination more directly. 91 The hypothesis testing regarding girls having more negative and less positive expectations for their future than boys produced the opposite results. In this data set, girls have more positive and less negative expectations than boys. Since this is a longitudinal data set, future research could examine if these expectations continue beyond high school. Women are clearly discriminated against in employment. Do girls perceive their opportunities diminishing as they grow older? Future research regarding how boys and girls are treated in schools is also needed. Are girls expected to cope better than boys in an academic setting? If so, why are girls expected to accomplish less in the math and science areas? What is it about the school setting that differentiates between gender so that boys and girls have different perceptions of what their future could hold? How are boys and girls socialized to develop expectations? Future research could also explore additional family types. This research only examined biological parent families. It would be interesting to know how other family types — step-parent, adoptive, or extended family members in the parent role — would impact the development of positive or negative expectations 92 APPENDICES 93 APPENDIX A Descriptive Statistics for Positive and Negative Expectation Scales 94 Descriptive Statistics for Positive and Negative Expectations Scales Std. N Minimum Maximum Mean Deviation Positive Expectations 3492 .00 1 .00 .8769 . 1619 Negative Expectations 3348 .00 .89 .2176 .1264 95 APPENDIX B Descriptive Statistics for Individual Expectations Questions 96 Descriptive Statistics for Individual Expectations Questions Std. Positive N Minimum Maximum Mean Deviation % chance in school next year 3528 O 100 92.09 22.83 % chance high school diploma by 20 3523 0 100 93.42 18.72 % chance college degree by 30 3511 0 100 72.78 31.48 % chance work 20+ hours/week by 30 3508 0 100 92.31 17.05 97 Negative Minimum Maximum Mean Std. Deviation % chance work 20+ hours, not in school next year 3485 100 79.91 29.44 %chance get drunk next year 3516 100 19.79 30.83 % chance victim of violent crime next year % chance arrested next year 3486 3508 100 100 14.70 10.22 20.77 18.17 % chance die next year 3408 100 18.64 22.49 % chance in jail by 20 3519 100 5.31 13.18 % chance a parent by 20 3489 100 17.76 26.00 % chance die by 20 3431 100 20.33 22.48 % chance pregnant or get someone pregnant next year 3500 100 7.90 17.60 98 APPENDIX C Demographics of cases with missing income 99 Demographics of Cases with Missing Income Not Missing Missing Income Income Data Data Totals Male 1284 5 1 7 1 801 Female 1272 492 1764 Totals 2556 (72%) 1009 (28%) 3565 White non-Hispanic 1337 402 1739 Black non-Hispanic 630 317 947 Hispanic 503 241 744 Totals 2470 (72%) 960 (28%) 3430 Father only 92 28 120 Mother only 721 302 1023 2 Biological Parent 1708 658 2366 Totals 2551 (72%) 988 (28%) 3509 100 REFERENCES Abramovitz, Mimi. (1996). Regulating the lives of women: Social welfare policy from colonial times to the present. Boston: South End Press. Annie E. Casey Foundation. (2003). State profiles of child well-being: Kids count data book. Baltimore: Annie E. Casey Foundation. Armstrong, Patrick Ian, and Crombie, Gail. “Compromises in adolescents’ occupational aspirations and expectations from grades 8 to 10". Journal of Vocational Behavior, 56, 82-98 (2000). Astone, Nan Marie and McLanahan, Sara S. “Family structure, parental practices and high school completion”. American Sociological Review, 1991, Vol. 56 (June:309-320). Axinn, William, Duncan, Greg J ., and Thornton, Arland. “The effects of parents’ income, wealth, and attitudes on childrens’ completed schooling and self-esteem” in Consequences of growing up poor. Greg J. Duncan and Jeanne Brooks-Gunn, editors, 1997. Bandura, Albert, Barbaranelli, Claudio, Caprara, Gian Vittorio, & Pastorelli, Concetta. “Self-efficacy beliefs as shapers of children’s aspirations and career trajectories”. Child Development, January/February 2001, Vol 72, No. I, pp. [87-206. Baron, Reuben M., and Kenny, David A. “The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations”. Journal of Personality and Social Psychology, 1986, Vol. 51, No. 6, 11 73-1182. Battistich, Victor, Solomon, Daniel, Kim, Dong-i1, Watson, Marilyn, & Schaps, Eric. “Schools as communities, poverty levels of student populations, and students’ attitudes, motives, and performance: a multilevel analysis”. American Educational Research Journal, Fall 1995, Vol. 32, N0. 3, pp. 62 7-658. Bergman, Manfred Max, and Scott, Jacqueline. “Young adolescents’ wellbeing and health-risk behaviours: gender and socio-economic differences”. Journal of Adolescence, 2001, 24, 183-197. Biblarz, Timothy J ., and Raftery, Adrian E. “Family structure, educational attainment, and socioeconomic success: rethinking the ‘Pathology of Matriarchy’”. American Journal of Sociology, Vol. 105, No. 2, (September 1999):.321-65. 101 Bolger, Kerry E., Patterson, Charlotte J ., Thompson, William W., & Kupersmidt, Janis B. “Psychosocial adjustment among children experiencing persistent and intermittent family economic hardship”. Child Development, 1995, Vol. 66, pp. 1107-1129. Bowles, Samuel and Gintis, Herbert (1976). Schooling in capitalist America: Educational reform and the contradictions of economic life. New York: Basic Books. Brandon, David. (1976). Zen and the art of helping. New York: Delta. Bronfenbrenner, Urie. (1979). The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press. Bronfenbrenner, Urie. “Ecological systems theory”. Annals of Child Development, Vol. 6,1717. 187-249, 1989. Brooks-Gunn, Jeanne, Duncan, Greg J. and Maritato, Nancy. “Poor families, poor outcomes: the well-being of children and youth”, in Consequences of growing up poor. Greg J. Duncan and Jeanne Brooks-Gunn, editors, 1997. Bush, D. M. & Simmons, R. G. (1981). Socialization process over the life course. In M. Rosenberg and R. H. Turner (Eds), Social psychology: Sociological perspectives, pp. 133-143. New York:Basic Books. Center for Human Resource Research (2000) NLS Y 97 user ’s guide. Columbus, OH: Ohio State University. Cohen, A. K. (1955).Delinquent boys: The culture of the gang. New York: Free Press. Conger, Rand D., Conger, Katherine Jewsbury, and Elder, Glen H., Jr. “Family economic hardship and adolescent adjustment: mediating and moderating processes”, in Consequences of growing up poor. Greg J. Duncan and Jeanne Brooks-Gunn, editors, 1997 Cook, Thomas D., Church, Mary B., Ajanaku, Subira, Shadish, Jr., William R., Kim, Jeong-Ran, and Cohen, Robert. “The development of occupational aspirations and expectations among inner-city boys”. Child Development, 1996, 67, 3368- 3385. Corcoran, Mary and Adams, Terry. “Race, sex, and the intergenerational transmission of poverty” in Duncan, Greg J. and Brooks-Gunn, Jeanne (Eds.) Consequences of growing up poor. New York: Russell Sage Foundation. De la Garza, Rodolfo 0., DeSipio, Louis, Garcia, F. Chris, Garcia, John, & Falcon, Angelo. (1992). Latino voices: Mexican, Puerto Rican, & Cuban perspectives on American Politics. Oxford: Westview Press. 102 Dominitz, Jeff, Manski, Charles F., & Fischhoff, Baruch. “Who are youth ‘at risk’? Expectations and evidence in the NLSY97", in Social Awakening: Adolescent behavior as adulthood approaches. Robert T. Michael, editor, 2001 . Duncan, Greg J. and Brooks-Gunn, Jeanne (eds).(1997). Consequences of growing up poor. New York: Russell Sage Foundation. Eamon, Mary Keegan. “The effects of poverty on children’s socioemotional development: an ecological systems analysis”. Social Work, Vol. 46, No. 3, July 2001. Erikson, E. H. (1963).Childhood and society. New York: W. W. Norton. Farber, Naomi B. “The significance of aspirations among unmarried adolescent mothers”. Social Service Review, December 1989. Felner, Robert D., Brand, Stephen, Dubois, David L., Adan, Angela M., Mulhall, Peter F ., Evans, Elizabeth G. “Socioeconomic disadvantage, proximal environmental experiences, and socioemotional and academic adjustment in early adolescence: Investigation of a mediated effects model”. Child Development, 1 995, Vol. 66, pp. 774- 792. Fischhoff, Baruch, Parker, Andrew M., Bruine De Bruine, Wandi, Downs, Julie, Palmgren, Claire, Dawes, Robyn, Manski, Charles F. “Teen expectations for significant life events”. Public Opinion Quarterly, Summer 2000, 64,2: 189-205. Gans, Herbert J. (1995). The war against the poor: The underclass and antipoverty policy. New York: Basic Books. Garis, Dalton. “Poverty, single-parent households, and youth at-risk behavior: an empirical study”. Journal of Economic Issues, Vol. my, No. 4, December 1 998. Gibbs, Jewelle Taylor, and Bankhead, Teiahsha. “Joblessness and hopelessness: The case of Afiican American youth in South Central Los Angeles.” Journal of Ethnic and Cultural Diversity in Social Work, Vol. 9(1-2), 2000. George, Darren & Mallery, Paul (2003). SPSS for windows: Step by step 11.0 update. Boston: Allyn and Bacon. Golden, Renny. (1997). Disposable children: America ’s child welfare system. Belmont, CA: Wadsworth Publishing Company. Gordon, Linda. (1994). Pitied but not entitled: Single mothers ad the history of the welfare state. Cambridge, MA: Harvard University Press. 103 Guerra, Nancy G., Huesmann, L. Rowell, Tolan, Patrick H., Van Acker, Richard, & Eron, D. Leonard. “Stressful events and individual beliefs as correlates of economic disadvantage and aggression among urban children”. Journal of Consulting and Clinical Psychology, 1995, Vol. 63, No. 4, 518-528. Hanson, Thomas L., McLanahan, Sara, and Thomson, Elizabeth. Economic resources, parental practices, and children’s well-being”, in Consequences of growing up poor. Greg J. Duncan and Jeanne Brooks-Gunn, editors, 1997. Harris, Kathleen Mullan, Duncan, Greg J ., and Boisjoly, Johanne. “Evaluating the role of ‘nothing to lose’ attitudes on risky behavior in adolescence”. Social Forces, March 2002, 80(3) :1 005-1 039. Hauser, Robert M. and Sweeney, Megan M. “Does poverty in adolescence affect the life chances of high school graduates?” in Consequences of growing up poor. Greg J. Duncan and Jeanne Brooks-Gunn, editors, 1997. Herr, Edwin L. “Perspectives on ecological context, social policy, and career guidance”. The Career Development Quarterly, September 1996, Vol. 45, pp. 5-19. Hays, Sharon (2003). Flat broke with children: Women in the age of welfare reform. New York: Oxford University Press. Hill-Collins, Patricia (1990). “Black feminist thought and the matrix of domination” in Lemert, Charles (Ed.).Social theory: The multicultural and classic readings. Boulder, CO: Westview Press. Hubbard, Lea. “College aspirations among low-income African American high school students: gendered strategies for success”. Anthropology & Education Quarterly 30(3): 363-383. 1999. Huston, Aletha C., Duncan, Greg J ., Granger, Robert, Bos, Johannes, McLoyd, Vonnie, Mistry, Rashmita, Crosby, Danielle, Gisbson, Christina, Magnuson, Katherine, Romich, Jennifer, & Ventura, Ana. “Work-based antipoverty programs for parents can enhance the school performance and social behavior of children. Child Development, January/February 2001, Vol 72, No. I, pp. 318-336. Jodl, Kathleen M., Michael, Alice, Malanchuk, Oksana, Eccles, Jacquelynne S., & Sarneroff, Arnold. “Parents’ roles in shaping early adolescents’ occupational aspirations”. Child Development, July/August 2001, Vol 72, No. 4, pp. 1247- 1265. Kozol, Jonathan. (1991). Savage inequalities: Children in America’s schools. New York: Harper Perennial. 104 Krein, Sheila Fitzgerald. “Growing up in a single parent family: the effects on education and earnings of young men”. Family Relations, 1986, 35, 161-168. Kroger, Jane. (1993). Identity in adolescence: The balance between self and other. London: Routledge. LeBlanc, Adrian Nicole. (2003). Random family: Love, drugs, trouble, and coming of age in the Bronx. New York:Scribner. Lerman, Robert I. “The impact of the changing US family structure on child poverty and income inequality.” Economica (1986) 63, S1 1 9-S1 3 9. Li, Jiang Hong, and Wojtkiewicz, Roger A., “A new look at the effects of family structure on status attainment”. Social Science Quarterly, Vol. 73, No. 3, September 1992. Lord, Shirley A. (1993). Social welfare and the feminization of poverty. New York: Garland Publishing, Inc. MacLeod, Jay (1995). Ain ’t no makin’ it: Aspirations and attainment in a low-income neighborhood. Boulder, CO: Westview Press. Markus, Hazel, and Nurius, Paula. “Possible Selves”. American Psychologist, September, 1986, Vol. 41, No. 9, 954-969. McLanahan, Sara and Sandefur, Gary (1994). Growing up with a single parent: What hurts, what helps. Cambridge, MA: Harvard University Press. McLanahan, Sara S. “Parent absence or poverty: which matters more?” in Consequences of growing up poor. Greg J. Duncan and Jeanne Brooks-Gunn, editors, 1997. McLeod, Jane D. & Shanahan, Michael J. “Poverty, parenting, and children’s mental health”. American Sociological Review, 1993, Vol. 58 (June:351-366). Merton, R. L. (1957). Social theory and social structure. Glencoe, IL: Free Press Michael, Robert T. (ed). (2001). Social awakening: Adolescent behavior as adulthood approaches. New York: Russell Sage Foundation. Mickelson, Roslyn Arlin. “The attitude-achievement paradox among black adolescents.” Sociology of Education, 1990, Vol. 63 (January): 4461. Mills, C. Wright. (1959). The sociological imagination. London: Oxford University Press. 105 Mueller, Daniel P. and Cooper, Philip W. “Children of single parent families: how they fare as young adults”. Family Relations, 1986, 169—1 76. Mullings, Leith. “Households headed by women: the politics of race, class, and gender” in The new poverty studies: The ethnography of power, politics, and impoverished people of the United States. Judith Goode & Jeff Maskovsky, editors, 2001. National Research Council (1993). Losing generations: Adolescents in high-risk settings. Washington, DC: National Academy Press. Paa, Heidi K., and McWhirter, Ellen Hawley. “Perceived influences on high school students’ current career expectations”. The Career Development Quarterly, September 2000, V0. 49, 29-44. Pagani, Linda, Boulerice, Bernard, Vitaro, Frank, & Tremblay, Richard. “Effects of poverty on academic failure and delinquency in boys: a change and process model approach.” Journal of child Psychology and Psychiatry, Vol. 40, No. 8, pp. 1209- 1219, 1999. Pagani, Linda, Boulerice, Bernard, and Tremblay, Richard E. “The influence of poverty on children’s classroom placement and behavior problems” in Consequences of growing up poor. Greg J. Duncan and Jeanne Brooks-Gunn, editors, 1997. Pandey, Shanta, and Zhan, Min. “Effect of urban poverty on parents’ expectation of their children’s achievement”. Advances in Social Work, Vol. I, No. 1, Spring, 2000, 1 07-125. Pelissier Kingfisher, Catherine (1996). Women caught in the American welfare trap. Philadelphia: University of Pennsylvania Press. Qian, Zhenchao, & Blair, Sampson Lee. “Racial/ethnic differences in educational aspirations of high school seniors”. Sociological Perspectives, 1999, Vol 42, No. 4. pp. 605-625. Rhea, Anisa, & Otto, Luther B. “Mothers’ influences on adolescents’ educational outcome beliefs”. Journal of Adolescent Research, Vol. 16, No. 5, September 2001, pp. 491-510. Rist, Ray C. “On understanding the processes of schooling: The contributions of labeling theory” in Karabel, Jerome, and Halsey, A. H. (Eds). (1997). Power and ideology in education. New York: Oxford University Press. Rodgers, Harrell R., Jr. (1996). Poor women, poor children: American poverty in the 1990s. Arrnonk, NY: ME. Sharpe. 106 Sanders, Christopher E., Field, Tiffany M., and Diego, Miguel A. “Adolescents’ academic expectations and achievement”. Adolescence, Vol. 36, No. 144, Winter, 2001. Sands, Roberta G. and Nuccio, Kathleen. “Mother-headed single-parent families: a feminist perspective”. Afi‘lilia, Vol. 4 No. 3, Fall 1989, 25-41.’ Shaw, Daniel S., Winslow, Emily B., and Flanagan, Clare. “A prospective study of the effects of marital status and family relations on young children’s adjustment among Afiican American and European American families”. Child Development, May/June 1999, Vol. 70, No. 3, pages 742- 755. Sidel, Ruth. (1996).Keeping women and children last: America’s war on the poor. New York: Penguin Books Smith, Judith R., Brooks-Gunn, Jeanne, and Klebanov. “Consequences of living in poverty for young childrens’ cognitive and verbal ability and early school achievement”, in Consequences of growing up poor. Greg J. Duncan and Jeanne Brooks-Gunn, editors, 1997. SPSS, Inc. SPSS: Statistical Package for the Social Sciences. Chicago: SPSS, Inc. 1999. Swartz, David. “Pierre Bourdieu: The cultural transmission of social inequality.” Harvard Educational Review. Vol, 47. No. 4. November 1977. Teachman, Jay D., Paasch, Kathleen, Day, Randal D., and Carver, Karen P. “Poverty during adolescence and subsequent educational attainment” in Consequences of growing up poor. Greg J. Duncan and Jeanne Brooks-Gunn, editors, 1997. Trusty, Jerry. “Family influences on educational expectations of late adolescents”. The Journal of Educational Research, May/June, 1998, Vol. 91, No. 5, pp. 260-270. Trusty, Jerry. “High educational expectations and low achievement: stability of educational goals across adolescence”. The Journal of Educational Research, August 2000 [V0]. 93 (No. 6} ], pp. 356-365. US. Department of Labor, Bureau of Labor Statistics. (2000). The national longitudinal surveys handbook. Wigfield, A. and Eccles, Jacquelynne. “Expectancy-value theory of achievement motivation.” Contemporary Educational Psychology, 25, 68-81 (2000). Wilson, Patricia M. & Wilson, Jeffrey R., “Environmental influences on adolescent educational aspirations: a logistic transform model”. Youth and Society, Vol. 24, No. 1, September 1992, 52- 70. 107 Young, Iris Marion. (1990)Justice and the politics of difference. Princeton, NJ: Princeton University Press. Yowell, Constance M. “Dreams of the future: the pursuit of education and career possible selves among ninth grade latino youth”. Applied Developmental Science, 2002, Vol. 6, No. 2, 62-72. Zinn, Howard (1995). A people ’s history of the United States: 1492 - present. NY: Harper Perennial. 108 IIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 11111111111IllLlljjjlljjjjll