<%BANNER%>

Level of Voice among Urban and Rural African-American Middle-School Students

Permanent Link: http://ufdc.ufl.edu/UFE0010523/00001

Material Information

Title: Level of Voice among Urban and Rural African-American Middle-School Students
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0010523:00001

Permanent Link: http://ufdc.ufl.edu/UFE0010523/00001

Material Information

Title: Level of Voice among Urban and Rural African-American Middle-School Students
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0010523:00001


This item has the following downloads:


Full Text

PAGE 1

LEVEL OF VOICE AMONG URBAN A ND RURAL AFRICAN-AMERICAN MIDDLE-SCHOOL STUDENTS By YASHICA JARIECE CRAWFORD A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS IN EDUCATION UNIVERSITY OF FLORIDA 2005

PAGE 2

ii ACKNOWLEDGMENTS This is the initial step toward reaching my educational goals, and the journey would not be complete without the s upport of my family, friends, a nd professors. I would like to thank my family for their constant words of encouragement. I thank my friends, especially Evelyn Chiang and Aisha Wood-Jackson who supported me through this project. I hold them in high esteem. My professors (Dr. Bridge t Franks, Dr. Patricia Ashton, and Dr. David Miller) have been a source of support and guidance through this process. Last, but not least, I thank my husband, Cesnae, who has held my hand the entire time.

PAGE 3

iii TABLE OF CONTENTS page ACKNOWLEDGMENTS..................................................................................................ii LIST OF TABLES...............................................................................................................5 ABSTRACT....................................................................................................................... vi CHAPTER 1 INTRODUCTION........................................................................................................1 2 REVIEW OF THE LITERATURE..............................................................................4 Gilligan’s Description of Loss of Voice.......................................................................4 Harter’s Description of Level of Voice........................................................................6 Level of Voice and Silencing: The Similarities between Two Concepts.....................9 Leander’s Description of Silencing.......................................................................9 Michelle Fine and Silencing................................................................................10 African-Americans and Voice.............................................................................13 Inclusion of African-American Rural Students...................................................15 Purpose.......................................................................................................................1 5 3 METHODS.................................................................................................................17 Participants.................................................................................................................17 Instrument...................................................................................................................17 Procedure....................................................................................................................19 Data Analysis..............................................................................................................20 4 RESULTS...................................................................................................................22 Research Question I....................................................................................................22 Research Question II...................................................................................................23 Analysis of Variance...........................................................................................23 Parents.................................................................................................................24 Teachers...............................................................................................................25 Male Classmates..................................................................................................26 Female Classmates..............................................................................................27

PAGE 4

iv Research Question III.................................................................................................28 Parents.................................................................................................................28 Teachers...............................................................................................................29 Male Classmates..................................................................................................30 Female Classmates..............................................................................................31 Simple Multiple Regression................................................................................32 Parents.................................................................................................................33 Teachers...............................................................................................................33 Male Classmates..................................................................................................34 Female Classmates..............................................................................................34 5 DISCUSSION.............................................................................................................36 Internal Consistency and Reliabilities........................................................................36 Level of Voice............................................................................................................37 Level of Voice with Parents................................................................................37 Level of Voice with Teachers..............................................................................38 Level of Voice with Male Classmates.................................................................39 Level of Voice with Female Classmates.............................................................40 Similarities and Differences from Previous Studies...................................................40 Limitations..................................................................................................................42 Future Research..........................................................................................................43 LIST OF REFERENCES...................................................................................................45 BIOGRAPHICAL SKETCH.............................................................................................47

PAGE 5

5 LIST OF TABLES Table page 4-1 Coefficient alpha of Teenage Voice Survey.......................................................23 4-2 Mean scores for level of voice with parents.......................................................25 4-3 Three-way analysis of variance for level of voice with parents.........................25 4-4 Mean scores for leve l of voice with teachers......................................................26 4-5 Three-way analysis of variance for level of voice with teachers........................26 4-6 Mean scores for level of voice with male classmates.........................................27 4-7 Three-way analysis of variance fo r level of voice with male classmates...........27 4-8 Mean scores for level of voice with female classmates......................................27 4-9 Three-way analysis of variance for level of voice with female classmates........28 4-10 Pearson correlations for level of voice with parents...........................................29 4-11 Pearson correlations for level of voice with teachers.........................................30 4-12 Pearson correlations for level of voice with male classmates.............................32 4-13 Pearson correlations for level of voice with female classmates.........................32 4-14 Level of voice with parent s multiple regression estimates.................................33 4-15 Level of voice with teachers multiple regression estimates................................34 4-16 Level of voice with male classm ates multiple regression estimates...................34 4-17 Level of voice with female classmates multiple regression estimates................35

PAGE 6

vi Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Ma ster of Arts in Education LEVEL OF VOICE AMONG URBAN A ND RURAL AFRICAN-AMERICAN MIDDLE-SCHOOL STUDENTS By Yashica Jariece Crawford May 2005 Chair: Bridget Franks Major Department: Educational Psychology Level of voice is described as people’s abilit y to express their true opinions and say what is really on their minds. A lack of voi ce has been linked in the literature to having depressed affect, lower self-esteem, and a depr essed global self-worth. Carol Gilligan is credited with describing voice, and found that females are prone to having a lower level of voice, because of submitting to the “good woman” stereotype, which proposes that women should keep their opinions to themselves and remain quiet. Females also acquire a lower level of voice as they mature deve lopmentally. Susan Ha rter and colleagues added to this literature by creating a Teenag e Voice Survey, which was used to quantify the level of voice that adoles cents have with various pe ople in their lives. They concluded that both males and females who s ubscribe to more feminine attributes are more likely to exhibit a de pressed level of voice. My study adds to this body of literature by including African-American middle-school students from rural and urban schools. Using Harter’s Teenage Voice

PAGE 7

vii measure, level of voice was assessed among 6th, 7th, and 8th grade middle-school students. Their levels of voice with their teachers, pare nts, male classmates, and female classmates were examined. Seventy-seven urban students and thirty-five rural students participated in the study. A three-way analysis of varian ce showed a significant difference in level of voice with parents based on grade level and type of school attended. Level of voice increased at each subsequent grade level. At the school level, rural students perceived their level of voice with their parents to be higher than did urban students. Level of voice with teachers differed significantly based on th e sex of the participant. Male students perceived their level of voice with their teach ers to be higher than female students did. Level of voice also differed significantly with male classmates and the type of school that the student attends. Students in the rural sc hools perceived their le vel of voice to be higher with their male classmates than urban school students did. Multiple regression analysis was used to assess whether two subscales of the survey, Reasons for not Saying What I Think (Reasons) and Support for Voice were adequate predictors of level of voice. Results were significant for the Reasons subscale but not for the Support for Voice subscale for parents, teachers, and male classmates. Given this, the Reasons subscale appeared be tter for assessing level of voice among the students.

PAGE 8

1 CHAPTER 1 INTRODUCTION Level of voice can be described as people’ s ability to express their opinions, and say their true thoughts and feelings to other pe rsons in their lives. A lack of voice has been linked in the literature to having depressed affect, lower self-esteem, and a depressed global self-worth (Harter, Waters, & Whitesell 1997). The purpose of my study is to explore African-American middleschool students’ percep tions of their level of voice with four others in their lives: pare nts, male classmates, female classmates, and teachers. Preliminary research on this concept can be credited to Gilligan (1982), who described loss of voice as a person’s inabi lity to express his or her true opinions. Gilligan’s qualitative research showed that this primarily happens to females who as adolescents, submit to the “good woman” stereo type, which is the commonly held belief that women should refrain from sharing th eir opinions, and that a good woman avoids conflict at all costs. Taylor, Gilligan, and Sullivan (1995) conti nued the discussion of level of voice by analyzing the influence of race and voice am ong females who were at risk for dropping out of school and for early pregnancies. In the Understanding Adolescence study, Taylor et al. (1995) interviewed 26 minority girls (including African-American girls) over a 3-year period. They examined this populat ion because most research on this population is cast based on participants being in some wa y maladaptive or deviant. They said the label of “at risk” is dangerous because of its ab ility to divert attention away from the very

PAGE 9

2 conditions that place a person at risk. Placing the attention on the person, “thus relieves the larger society of the responsibility of a ddressing the inequities of race, class, and gender that create conditions of risk” (p. 21). They conclude d that the participants may experience loss of voice differently than wome n who are Caucasian and come from more privileged backgrounds. The voice that the partic ipants used to speak on the injustices of their communities may have been a contributi ng factor in placing them at odds with that community, which ultimately cast the part icipants as “at risk” and deviant. Taylor et al. (1995) proposed that th e influence of race among African-American females adds complexities to the examinati on of voice. Through inte rviews, they learned that the participants (and African-American females generally) were taught the traditional feminine roles, such as being a nurturing moth er; and also were encouraged to be strong and outspoken and to provide for the famil y, which are traditionally masculine roles. These qualities may be contradictory in traditional setti ngs, and place the African-American female at risk of being silenc ed or in conflict for being too outspoken. Taylor et al. (1995) observed this during thei r study, in which they interviewed 26 girls over the course of 3 years. Harter (1996) revisited th e idea of level of voice by cr eating a quantifiable measure (in the form of a survey) to determine the level of voice among adolescents. Harter included males in the study to provide sufficien t evidence for Gilligan’s initial claims that loss of voice is most prevalent among female s. She concluded that loss of voice depends more on the gender roles that males and fe males subscribe to than simply on sex classification. Females who subscribed to more feminine roles were more likely to have a loss of voice than females who were androgynous. In contrast, males who were more

PAGE 10

3 androgynous were more likely to experience loss of voice than males who subscribed to more masculine roles. Female loss of voice was related to complying with the stereotypical role for females, while male loss of voice was related to noncompliance with the stereotypical role for males. My goal was to bridge the gap between Gilligan’s (1982) an d Harter’s (1996) research by examining the interview study of ethnic minority girls that included African-Americans described by Taylor et al. (1995) and contrasting it with Harter’s (1996) survey research. Most of the participants in Harte r’s (1996) study were Caucasian, making it impossible to analyze the influen ce of level of voice among African-American students. It is unclear whethe r Harter’s survey is a reliable measure of level of voice in these students, or whether Harter’s populat ion of students would respond to the measure in the same way as Caucasian students have. My study offers a significant contribution to the literature by using Harter’s Teenage Voice measure with a previous ly ignored population, African-American teenagers. I established the reliability of the measure with a sample of African-American middle-school students, and de scribed a number of results that differ from those of previous studies. The results have implicat ions for future path s of study, including gender differences, whether the “good woman” stereotype applies to African-American girls, and how level of voice could positivel y or negatively affect students’ academic achievement and overall motivation for educatio n. I hope that my study will give this population of students a voice by serving as a pre liminary step in a la rger exploration of African-Americans and level of voice in the classroom.

PAGE 11

4 CHAPTER 2 REVIEW OF THE LITERATURE The classroom environment is oftentimes a safe haven for children, particularly those who are not receiving the guidance and sup port they need at home. For this reason the classroom setting represents an educationa l arena and also a pow erful social context in which the psychological adjustment of children and adolescents can be affected (Harter, 1996). One of the many attributes adolescents acquire through socialization with their family and in the classroom is a level of voice. Level of voice has been described as a person’s ability to express his or her true opinions (Gilligan, 1982), and is an extension of one’s true self (Harter, 1996). Although th e contributions of Gilligan and Harter was a primary focus for my study, other views on le vel of voice are also presented. While Oldfather and Dahl (1994) described an honoured voice that should be present in the classroom, Fine (1987) described the debilita ting effects that a silencing environment could have on a person’s ability to express his or her true voice. Gilligan’s Description of Loss of Voice Gilligan (1982) is credited with first descri bing a person’s inability to express his or her true opinions as loss of voice. In adol escence girls are more prone to loss of voice than their male counterparts are. Gillig an (1982) found that the suppression of thoughts and opinions, or loss of voice, is particular ly problematic for females beginning with the onset of adolescence. Gilligan said that as females enter adolescence, they begin to acquire the identity of a woman in our societ y. Gilligan said that as part of this new identity, girls perceive that women’s opinions are not valued or s upported, and that the

PAGE 12

5 “good woman” in our society is one who is kind, polite, and (most important) quiet. Incorporation of this “good wo man” stereotype into the female adolescent’s identity (coupled with the feminine values of staying connected and preserving relationships at all costs) leads the female adolescent (as she b ecomes a woman) to suppress and eventually lose a once-str ong childhood voice. Gilligan (1993) said that girls have “insid er knowledge” of their views of life when they are at the edge of adolescence. Gi rls have this knowledge based on their own perspective. However, as they reach adolesce nce, they realize that the information that they were privy to while they were younger is in jeopardy of being washed away. Girls now begin to focus their attention on thei r insider knowledge and the perspective of others, or outsiders’ views of life. This focu s of attention allows girls to negotiate and determine which voices people want to listen to from girls and which voices are deemed to reflect rudeness or stupidity. Girls rece ive lessons every day about what must be spoken versus what must be kept in, and what to suppress so as not to be told that they are bad or wrong. Further research shifted Gilligan’s attent ion from exploring female loss of voice generally, to exploring specific populations of women who may be more susceptible to experiencing it. Taylor et al. (1995) conducted a re search project called the Understanding Adolescence Study. The purpose of the study was to interview 26 girls annually over a 3-year period who were deem ed at risk for early pregnancy and dropping out of high school. They chose this populati on of participants because they wanted to listen and understand their voices and also add to the body of literature on a population that is inadequately represented or missing in theories of adolescent development. While

PAGE 13

6 a few of the girls did go to college, most of them experienced early pregnancy, dropping out of school, and working at low-wage jobs which undoubtedly secure d their position in the working class. During the interview proce ss, the girls were asked about their feelings toward school, their relationships, their fu ture outlook, and themselves. The group was made up of students who the authors consid ered ethnic minorities. These included African-American, Caribbean, Latina, Portugue se, Irish, and Italian Americans. Taylor et al. (1995) said that one of the ke y resources that girls who were at risk for academic failure had was a meaningful relati onship with a woman. This relationship was associated with the girls who did not have an early pregnancy, gr aduated high school, and for some, pursued higher education. It was w ithin these relationships that the girls’ voices were validated and taken seriously. These relationships are important because girls’ and women’s experiences are often devalu ed within patriarchical societies, and this devaluation begins to jeopardize their sense of themselves a nd relationships. When girls’ experiences come into conflict w ith their relationships, or thei r sense of self is in conflict with the “good woman” stereo type, it is the voice of wome n that can be psychologically necessary for girls. Furthermore, the experiences of sexuality, relationships, and work are made more complicated when race, cla ss, and sexual orientat ion are taken into consideration. The women that were important in the lives of these girls named women who were similar to themselves as role models. Harter’s Description of Level of Voice Harter (1996) defines level of voice as a person’s abil ity to express his or her opinions. According to Harter, Waters, White sell, and Kastelic (1998), level of voice describes the opportunity, encouragement, and validation that students receive when they choose to express their true opi nions and express what they re ally feel. This description

PAGE 14

7 of level of voice stems from previous resear ch on false self-behavior. This involves acting in ways that does not reflect one’s true self as a person, the “real me.” This includes “not saying what you think,” “expressi ng things you don’t real ly believe or feel, “not saying your true opinion,” “saying what you think other people want to hear.” Harter et al. (1998) ha ve concluded that among both adol escents and adults, those who do not receive validation for the expression of their true or authentic selves engage in “false self behavior, which in turn is associated with low self-esteem, hopelessness, depressed affect, and the acknowledgement that one doe s not even know one’s true self” (p. 37). They also said that children create a sense of self during adolescence. They also realize that depending on whom they are associ ating with at the time, be it their parents, teachers, or classmates, their conceptualizati on of self can take different forms. The influence of multiple selves naturally encourages adolescents to search for their true identity that describes them. The true self is described as the “real me inside,” and “what I really think and feel”, wh ereas the false self includes “being phony” and “putting on an act.” The true self during adolescence is often verbali zed using the aforementioned phrases. This exemplifies true self-behav ior. However, not saying what you think exemplifies false self-behavior. Harter (1996) examined the impact of t eachers and classmates on three constructs that represent different indices of adjustment within the scho ol context. These include: (a) intrinsic and ex trinsic motivation for classroom lear ning, (b) self-esteem, and (c) level of voice. Harter (1996) described the fi ndings of an unpublished manuscript (Harter, Waters, and Gonzales, 1994) in which she and other researchers creat ed a systematic way of analyzing level of voice among female and ma le students in grades 6, 7, and 8 though

PAGE 15

8 the creation of the Teenage Voice Survey. Th ey were most interested in the level of voice that students perceived that they had in three different relational contexts: with close friends, with peers, and with teachers in the classroom. The study also explored the potential causes of individual differences in level of voice by proposing that the endorsement of gender role stereotypes in the form of femininity, masculinity, and androgyny may be important indicators. In m easuring the prevalence of level of voice among female and male students, the rese archers found that students were most outspoken with their close friends, followed by peers, and then teachers. Their findings coincided with Gilligan’s observation that it is within one’s rela tionship with closest friends that voice should mo st likely be expressed. Di ffering from the findings of Gilligan, however, was their observation by Harter et al. (1994) that males and females had identical levels of voice in their relations hips with close friends and peers, which is counter to Gilligan’s claim that loss of voi ce is a particular problem for females. Consistent with Gilligan’s claim, however, was their finding that female students showed significantly less voice than males with teachers. In reference to the endorsement of gender role stereotypes, Ha rter et al. (1994) concluded that gender role orientation is a more powerful predictor than gender. Their findings showed that female students who were high in femininity and low in masculinity reported significantly lower levels of voice than those who were androgynous. The males who endorsed a feminine orient ation reported lower levels of voice than those endorsing either an androgynous or a masculine orientatio n. They concluded that those of either gender who adopt feminine sex-role stereotype d behaviours are at greater risk for lower levels of voice.

PAGE 16

9 Harter et al. (1994) also studied the reasons for low level of voice among the middle-school students. Through piloting t echniques and criteria proposed by Gilligan (1982), four factors were used to measure low level of voice: lack of validation of self (others won’t listen to me, won’t take me seriously); threats to the relationship (will cause conflict, will lead to tension in the relations hip); affective reactions (I’d feel embarrassed, others would make fun of me ); and lack of opinion (I don’t have an opinion, I’m not sure what to think). The researchers found that the primary motive for students’ feelings of loss of voice with teachers was lack of validation. In relati on to classmates, the students’ reasons for lack of voice included lack of validation and negative affective reaction. Level of Voice and Silencing: The Similarities between Two Concepts There are two concepts described in the li terature that have similar meanings, but one is usually discussed wit hout acknowledgement of the other. While level of voice refers to individuals’ belief in their ability to express their opinions, silencing includes the environmental effects that may prohibit the e xpression of certain types of talk. There may be a link between level of voice (or lack thereof) and silencing that has not been made in the existing literature. The inability of individuals to say what is on their minds or speak their true feelings is called lack of voice. This pa rallels what other researchers have described as silencing. Leander’s Description of Silencing Leander (2002) described the phenomenon of silencing based on the negotiation of social space. He sought to explain this phenomenon through an analysis of movement, the creation of social space, such as the coordination of group formations, and using microlevel individual movements, such as gest ures and gazes. He described the processes that may be in play when one is being silenced in a 10-month ethnographic study which

PAGE 17

10 examined the relations of school-related disc ourse to the production of social space. Although he acknowledges that silence is a “rich medium of communication,” it becomes problematic when people or groups are not allowed, for any reason, to express themselves (p. 194). Leander describes sile ncing as a disabling act that inhibits a person’s ability to interact with others. Examples of how silencing occurs include “talking over another, raising one’s volume interruption, and criticism” (p. 194). Michelle Fine and Silencing Although level of voice has been described as a silencing mechanism that separates individuals from group activities, other author s have described the overarching negative effects that silencing can ha ve in students, particularly in lower socio-economic environments. Leander cites the seminal work by Fine (1987), who further described what silencing entails when in the confines of a school environment. Fine (1991) said that there are certain conversa tions that are not allowed in the classroom. As a result, students cannot critically analyze their presen t economic conditions and the reasons that may be associated with their present status. By not allo wing this dialogue to exist, administrators and teachers systematically e xpel the discourses and social lives of many of their students. Fine used the example of a girl who is in the classroom. While the girl who says nothing is often times seen as being obedient, this may have different implications for young African-American girls whose lack of voice in the classroom is “directly linked to the institutional control of allowable discourses. The discourse of the girls is not permitted, thus they are silenced” (p. 198). This description of silencing parallels Ta ylor et al. (1995), who discussed voice and the influence of race. They said that voice and silence are used among girls to navigate contradictory situations that they experience at home and at school. If either voice or

PAGE 18

11 silence is used too much, each may have its advantages and disadvantages. Outspoken girls will be in conflict with peers and parents. Silence in order to keep in touch with their culture and communities will get them branded as good girls, but at the expense of being overlooked or unnoticed. Fine (1991) said that public schools a nd schools with large percentages of low-income students are more likely to avoi d conversations which involve talking about the difficult conditions of stude nt’s lives and the failure of schools to help them. Silencing shapes low-income public schools mo re intimately than relatively privileged ones. Silencing signifies a terror of words, a fear of talk. Fine sought to determine whom the silencing protects, identify the practices by which silencing is institutionalized, and explore how the muting of students and their communities undermines a project of educational empowerment. She further said that “low-income schools officially contain rather than explore social and economic cont radictions, condone ra ther than critique prevailing social and economic inequities, and usher children and adolescents into ideologies and ways of interpreting social ev idence that legitimate, rather than challenge conditions of inequity” (p. 61). When silence is prevalent in a school, the students take on a ca ste-like character, in which those that are silenced the most are una ble to progress in a positive direction. Fine (1992) attributes this form of caste-like character, and the ending of critical consciousness among students in low-income schools to the silencing in urban public schools. The criteria which describe a silencing school environment include The preservation of the ideology of equal opportunity and access while obscuring the unequal distribution of resources and outcomes.

PAGE 19

12 The creation within a system of severe asymmetric power relations of the impression of democracy and collaborat ion among “peers” (e.g. between white, middle-class school administrators an d low income African-American and Hispanic parents or guardians.) The quieting of student voices of differen ce and dissent so that such voices, when they burst forth, are rendered deviant and dangerous. The removal from public discourse of the tensions between (a) promises of mobility and the material realities of students’ lives; (b) explicit claims to democracy and implicit reinforcement of power asymmetries; (c) schools as an ostensibly public sphere and the influen ces wrought on them by private interests, the church, and the military; and (d) th e dominant language of equal educational opportunity versus the undeni able evidence of failure as most experience for low-income adolescents. A classroom that shows these characteristics will stifle the intellectual growth of an at-risk African-American stude nt and also will not provi de a communal ground to share discourses about the injustices that this population experi ences daily. Without this avenue, the students will not be able to rela te the daily experiences of the classroom to themselves. A culturally relevant classroom, on the other hand, is essential for increasing talk across cultures. This includes provi ding a curriculum that allows teachers and students alike to directly in clude positivity for a multicultural community. A culturally relevant classroom also involves engagement of students in discussions and activities that increase diversity, choice and student em powerment in the school (Wardle & CruzJanzen, 2004). Another view of voice and silencing is offered by Diaz-Greenberg (2001), who studied the ramifications of having an ab sence of voice among Latino students. In a description of why the Latino st udents in his study felt that th ey were being silenced, the author described the concerns of one student : “Sometimes they (teachers) don’t take the time to listen, they think we are just kids . I also think that, at other times, when you have an opposing point of view, it can seem threatening to someone else if you speak

PAGE 20

13 against them, and they feel out of control when another voice is heard” (p. 58). The student seems to be describing the disadvantages of voicing opinions to teachers. If the student feels that their voice is not being received by the te acher, the student may choose to silence themselves instead to avoid conflict. African-Americans and Voice Descriptive research concerning urban adoles cent girls, particul arly those who are at risk for academic failure, is important because it gives a voice to a population of students who are rarely included in the c onversation about impr oving their academic progress. The social location of class, ge nder, and race for these students creates for them a position in society that does not garn er them a public hearing for voicing their experiences (Taylor et al., 1995). It is doubly important to understand how this population of students experiences their classroom environments. African-Ame rican students, particularly males, are disproportionately represented in special education categories, school suspensions, and the juvenile justice system (Leone, Christ le, Nelson, Skiba, Frey, & Jolivette, 2003). Additionally, African-American students are al so more likely to be instructed by less qualified teachers who lack cer tification in specific areas, such as math (Barton, 2003). Harter and colleagues (1997) said that the media a nd popular culture help to perpetuate the stereotype of the desirabl e woman who should be seen and not heard. While adolescents often learn the rules and cu ltural norms from media and other sources, this can have a troubling effect on African-Ame rican adolescents. Tatum (1997) said that while African-American girls are bombarde d with devaluing messages on who they are and who they will become, African-American boys receive very similar devaluing messages. While African-American girls are often portrayed in the media as school drop-

PAGE 21

14 outs and teenage welfare mothers, young Africa n-American males are often seen with handcuffs and being arrested for a violent cr ime. Tatum (1997) believes that these adolescents often times do not find voice in expressing their experiences with other students who are not in their racial group or by those that represent the mainstream culture, but they do find solace with persons that are their same age a nd also in their same race. Taylor et al. (1995) said th at African-American girls ar e often described as being socialized less stereotypically than their Caucasian counterpa rts. Given this, the “good woman” stereotype may mean something diffe rent to African-Ameri can girls. Although African-American girls are given the traditiona l female role teachings, such as being a nurturing mother, they are also encouraged to be strong, self sufficient, and fight back (verbally and physically) when th ey believe that an injustice has occurred against them. This influences the voice or silence that is used in a given situation. In their concept of voice, Oldfather and Dahl (1994) used the term “honoured voice” to denote the responsiveness in the cl assroom environment to student’s self expression. “Through honoured voice, the comm unity of learners invites, listens, responds to, and acts upon student’s thoughts, fee lings, interests, and needs” (p. 143). In light of the previous discussions of silenc ing and culturally releva nt classrooms, it is interesting to consider the relationship that th ese environments have with level of voice. They contend that if knowledge is shared with in the classroom, then the teacher does not position herself as the sole source of inform ation, but rather, invites a diverse array of opinions.

PAGE 22

15 Inclusion of African-American Rural Students There is a need to bring a ttention to the issues and c oncerns of the rural school. Despite the fact that 43% of the nation’s public schools are in rural areas, and 31% of the nation’s children attend these schools, literature is lacking in the area that may determine the needs that may be unique to this co mmunity (Beeson & Strange, 2003). There are some issues that are prevalent in rural schools that urban schools may not have to address. This includes the recruitment a nd retention of quality teachers who receive a lower salary than those teachers in urban distri cts. Rural schools also have to address the long bus rides that students must endure wh ich may reduce study time. While the large transportation costs deplete district funds, the sparsity of these schools make it difficult to connect them with the digital world to the level of their urban counterparts (Beeson & Strange, 2000). One of the central issues re garding this community is poverty; rural poverty rates are as high as in urban cente rs. Because of this economic fact, these children are at risk of not be ing provided with a quality education (Beeson & Strange, 2003). Additionally, those rural school dist ricts with a high percentage of minority students are often classified as an at risk population. St ates in the South, including Florida, rank high (Beeson & Strange, 2003). Purpose The purpose of my study was to explor e level of voice among rural and urban African-American students. The goal was to replicate establishe d findings on level of voice among teenagers and also to explore whet her the sex of the student and type of school he/she attends affects the level of voice that a teenager may e xhibit. Since this population of students has yet to be analyzed using a quantifiable measure, it may be erroneous to conclude that th e level of voice among these students will resemble the level

PAGE 23

16 of voice among students in majority populati on. Given this, the following research questions were answered: Research question 1. Do the three sections of the Teenage Voice survey ( Level of Voice Reasons and Support for Voice ) produce reliable results with AfricanAmerican middle-school students? Research question 2. Does the mean score on level of voice vary as a function of a student’s grade level, se x, and type of school? Research question 3. Are the two scales of the Teenage Voice survey ( Reasons and Support for Voice ) adequate predictors of leve l of voice in African-American students?

PAGE 24

17 CHAPTER 3 METHODS Participants One urban middle school located in northeas t Florida and 2 rural middle schools in northeast Florida participated in the study. Seventy-seven ur ban students (57 females and 20 males) and 35 rural students (22 females a nd 13 males) participated in the study. All of the 112 students who participated in th e study classified themselves as being African-American. In the urban school, 26 si xth graders (17 females and 9 males), 14 seventh graders (10 females and 4 males), a nd 37 eighth graders (30 females and 7 males) participated. The urban school was predom inately African-American, and consisted of low to medium income students. Becau se of the limited amount of enrolled African-American students at each rural sc hool, the two school populations were combined to create a compos ite rural school population. In all, 18 sixth graders (11 females and 7 males), 8 seventh graders (4 fe males and 4 males), and 9 eighth graders (7 females and 2 males) from the rural school s participated. Both of the rural schools mirrored each other based on the percentage of African-American students and the population of the surrounding community. Instrument The Teenage Voice survey (Harter et al ., 1994) was used to evaluate level of voice. Harter, Waters, and Whitesell (1997) re ported internal consistency reliabilities from three studies, ranging from .82 to .91. Th e survey consisted of three sections. The first section assessed the depende nt variable, level of voice. The questions of the survey,

PAGE 25

18 entitled “Saying what I think” were designed to measure the extent that teenagers have the ability to express their opi nions to four different groups of people: parents, teachers, male classmates, and female classmates. Part icipants answered five questions for each group of people. After reading two contrasting statements, st udents were asked to choose which statement they were most like, and th en chose how true that statement was for them. An example of two statements from the parent section is “some teenagers share what they are really thinking with their pare nts BUT other teenagers find it hard to share what they are thinking with their parents.” If the former statement was chosen, a score of 4 was given if the student chose “very true for me” and a score of 3 was given if the student chose “sort of true for me.” If the latter statement was chosen, a score of 1 would be given if the student chose “very true fo r me” indicating the lowe st level of voice for that question and a 2 would be given for “sort of true for me.” The next two sections of the survey se rved as independent variables to predict level of voice. Section two of the survey is entitled “Reasons for not saying what one thinks” (hereinafter referred to as “Reasons ”). The questions for this portion of the survey were used to address the specific reasons for students to not say what they think. For each statement, students chose how true th e statement was for them for each group of people: parents, teachers, male classmates, and female classmates. The students were given the option of choosing very tr ue (score of 4), sort of true (score of 3), not very true (score of 2), or not at all true (score of 1). At least four statements were used to evaluate categories used to assess specific reasons. These categories were lack of validation, threat to the relationship, don’t have an opinion, would feel em barrassed/ stupid, and good woman stereotype.

PAGE 26

19 Section three of the survey is entitled “Pe ople in my life” (hereinafter referred to as “Support for Voice”). The questions for this por tion of the survey were used to address the level of support that students receive for voicing their opinion to their parents, teachers, male classmates, and female classmat es. Five items for each of the four types of people were used to assess support of voi ce. For each item, students chose how true the statement was for them. Section thr ee mirrored section one based on the way the questions were asked and scored. An example of two statements from the parent section is “my parents do not listen to my opinions and take them seriously OR my parents do listen to my opinions and take them seriously.” Procedure Each individual school principal determined the method used to solicit requests to participate from students. In the urban sc hool, language arts teach ers agreed to help inform students about the study. Given the fact that most of the students were African-American, a researcher with the project was able to address e ach individual class that the language arts teache rs had and gave a brief synops is of the project and asked students for assistance in the in itial part of class. Student s were given parental consent forms, and those who returned them were elig ible to participate in the study. Students were given a lollipop for returning a signed pa rental consent form, regardless of whether or not they chose to complete the survey. In one rural school, the principal reques ted that all the teachers who have a homeroom in the beginning of the school da y receive permission forms. They were instructed to give the permission forms to African-American students and explain to them the nature and importance of the study. St udents were given a lollipop for returning a signed parental consent form, regardless of wh ether or not they chose to complete the

PAGE 27

20 survey. In the other rural school, 4 language arts teachers and 1 world history teacher agreed to help inform students of the study. A researcher with the project was allowed to address the students in each of the teache r’s classes about the nature of the study. Students were given a lollipop for returning a signed parental consent form, regardless of whether or not they chose to complete the survey. In each of the schools, students were admini stered the survey in a group setting. A researcher with the project read the child as sent, which informed students of their rights with regard to participation and privacy. The reading levels of the students were not assessed before their completion of the surve y. To ensure that st udents completed the survey, did not answer questions in error and understood the proce dures for completing the survey, they survey was read aloud by the researcher, and students were asked to answer each question after it was read. At the end of the administration, students were given a lollipop for completing the survey. Data Analysis For section one of the survey, each subscal e was given a mean score for voice for each set of people in the students’ lives (par ents, teachers, male classmates and female classmates). A composite mean score was used in the analysis to assess level of voice for the participants as a whole. For secti on two, a composite score was given for each category assessing the reasons for not saying op inions. A composite score was also given for each set of people. In section three, a composite support for voice score was given for each set of people. Composite scores were derived by adding the scores from each question of the survey, and dividing them by the total number of respondents, creating a mean score for the question.

PAGE 28

21 To evaluate if there is a difference in level of voice among the participants based on sex, grade level, and type of school, mean sc ores were used to describe the composite scores. A three-way ANOVA was used for each relational context to determine whether differences in level of voice shown by the m ean scores were signifi cant. Reliability and item analysis were used to determine wh ether or not the items measured a single construct. Additionally, to assess whether Reasons and Support for Voice were good predictors of level of voice w ith this population of students, simple multiple regression analysis was used.

PAGE 29

22 CHAPTER 4 RESULTS Research Question I Coefficient alpha was used to determine the internal consistency estimates of reliability for each of the three sectio ns of the Teenage Voice survey. The Level of Voice Reasons and Support for Voice scales included responses fo r each relational context: teachers, parents, male classmates, and female classmates. Coefficient alphas for each scale are included in Table 4-1, and means and standard deviations for the sample as a whole. The internal consistency of the scales of the survey ranged from marginal (.60) to high (.84). Item Discrimination. There were three ques tions based on the item-total statistics that showed an improved coeffici ent alpha if the questi on was deleted, although improvements were negligible. Within the Reasons subscale, analysis showed that removal of the question “I don’t express my opinion because it will cause tension in the relationship” for the male classmates relational context would increase threat to relationship from 0.69 to 0.75. The analysis also showed that removal of the question, “Often I don’t really have an opinion, so I don’t say anything” for the male classmates relational context within the Reasons subscale would increas e don’t have an opinion from 0.64 to 0.67. Lastly, the analysis show ed that removal of the question, “Some teenagers share what their thinking But other t eenagers find it hard to share what they are thinking their parents” would increase Level of Voice with parents alpha from .66 to .68.

PAGE 30

23 Table 4-1. Coefficient alpha of Teenage Voice Survey Subscale Relational M SD context Level of Voice Saying What I Think Teachers .75 2.95 .76 Parents .66 2.89 .66 Male Classmates .69 2.79 .69 Female Classmates .66 2.81 .70 Reasons Lack of Validation Teachers .75 2.11 .83 Parents .80 1.93 .83 Male Classmates .72 2.16 .74 Female Classmates .74 2.09 .76 Threat to the Relationship Teachers .64 2.00 .70 Parents .73 2.08 .81 Male Classmates .69 2.14 .75 Female Classmates .77 2.27 .86 Don’t have an Opinion Teachers .69 1.96 .72 Parents .69 1.83 .71 Male Classmates .64 1.91 .66 Female Classmates .73 1.89 .73 Would Feel Stupid Teachers .73 1.72 .75 Parents .75 1.60 .70 Male Classmates .79 2.06 .86 Female Classmates .78 1.96 .83 Good Woman Stereotype Teachers .60 1.72 .60 Parents .70 1.74 .66 Male Classmates .72 1.83 .70 Female Classmates .70 1.70 .64 Support for Voice People in my Life Teachers .77 3.11 .66 Parents .84 3.34 .74 Male Classmates .81 2.73 .74 Female Classmates .83 2.96 .75 Research Question II Analysis of Variance To determine whether there were significant differences in the mean scores for level of voice for the participants, a thr ee-way analysis of va riance (ANOVA) based on grade level, sex, and type of school attended wa s used for each relationa l context. Tables

PAGE 31

24 4-2 and 4-3 contain the mean scores and ANOVA for parents, respectively. Tables 4-4 and 4-5 contain the mean scores and ANOVA for teachers, Tables 4-6 and 4-7 contain the mean scores and ANOVA for male classmat es, and Tables 4-8 and 4-9 contain the mean scores and ANOVA for female classmates. Parents The first analysis evaluated the relationshi p between level of voice with parents and grade level, sex of student, a nd the type of school the student attended and is represented in Table 4-3. The ANOVA wa s significant for grade, F (2, 97) = 3.50, p = .03, and school, F (1, 97) = 10.45, p = .00, indicating that students di ffered in their perceived level of voice with parents based on the grade le vel and type of school they attended. However, a Tukey HSD post hoc analysis fo r grade level showed that there was no significant difference in perceived level of voice based on grade level. The Tukey HSD is a more conservative post hoc comparison than the F-test because the alpha level is adjusted. The strength of relationshi p between perceived level of voice with parents and grade level, as assessed by 2, was moderate, with grade le vel accounting for 7% of the variance in perceived level of voice with parents. The stre ngth of relationship between perceived level of voice with pa rents and school, as assessed by 2 was moderate, with type of school attended accounting for 10% of the variance in perceived level of voice for parents. There was no main effect for sex and there were no significant interactions. As shown in Table 4-2, rural students’ percei ved level of voice with their parents was significantly higher than that of urban students. No other main effects or interactions for the indicators were observed.

PAGE 32

25 Table 4-2. Mean scores for level of voice with parents Grade 6th grade 7th grade 8th grade M 2.79 2.84 3.01 SD .61 .62 .72 School Urban Rural M 2.77 3.14 SD .64 .65 Sex Male Female M 2.75 2.95 SD .56 .69 Table 4-3. Three-way analysis of vari ance for level of voice with parents SS MS DF F p 2 Grade 2.80 1.40 2 3.50 .03* .07 Sex .25 25 1 .63 .43 .01 School 4.16 4.16 1 10.45 .00* .10 Grade*Sex 1.44 .72 2 1.80 .17 .04 Grade*School .62 .31 2 .78 .46 .02 Sex*School .00 .00 1 .01 .91 .00 Grade*Sex*School .53 .27 2 1.80 .17 .04 Error 38.67 .40 97 p < .05 Teachers The second analysis evaluated the relati onship between perceived level of voice with teachers and grade level, sex of student and type of school th e student attended and is represented in Table 4-5. The ANOVA was significant for sex of student, F (1, 100) = 4.74, p = .03, indicating that male students’ perc eived level of voice with their teachers was significantly higher than that of female students. Refer to Table 4-4 for means and standard deviations. The strength of relati onship between perceived level of voice with teachers and sex of student, as assessed by 2, was moderate, with the sex of the student accounting for 5% of the variance in perceive d level of voice with teachers. No other main effects or interactions for the indicators were observed.

PAGE 33

26 Table 4-4. Mean scores for level of voice with teachers Grade 6th grade 7th grade 8th grade M 2.89 3.01 2.97 SD .81 .69 .75 School Urban Rural M 2.89 3.07 SD .73 .81 Sex Male Female M 3.26 2.82 SD .69 .69 Table 4-5. Three-way analysis of vari ance for level of voice with teachers____________ SS MS DF F p 2 Grade .54 .27 2 .48 .62 .01 Sex 2.67 2.67 1 4.74 .03* .05 School .81 .81 1 1.44 .23 .01 Grade*Sex .27 .14 2 .24 .79 .01 Grade*School 1.07 .53 2 .95 .39 .02 Sex*School .00 .00 1 .19 .73 .00 Grade*Sex*School .00 .00 2 .06 .94 .00 Error 56.28 .56 100 p < .05 Male Classmates The third analysis, represente d in Table 4-7, evaluated pe rceived level of voice with male classmates and grade level, sex of student, and type of school attended. The ANOVA was signifi cant for type of school, F (1, 100) = 9.07, p = .00, indicating that rural students’ perceived leve l of voice with male classm ates was significantly higher than that of urban students (see Table 4-6 for means and standard deviations). In determining the strength of relationship between level of voice with male classmates and type of school attended, the 2 showed that 8% of the variance in perceived level of voice with male classmates can be attributed to the type of school attended. No other main effects or interactions for the indicators were observed.

PAGE 34

27 Table 4-6. Mean scores for leve l of voice with male classmates Grade 6th grade 7th grade 8th grade M 2.63 2.91 2.89 SD .68 .74 .67 School Urban Rural M 2.68 3.05 SD .67 .68 Sex Male Female M 2.61 2.87 SD .66 .70 Table 4-7. Three-way analysis of variance for level of voice with male classmates___ SS MS DF F p 2 Grade 2.55 1.27 2 2.88 .06 .06 Sex .76 .76 1 1.71 .19 .02 School 4.01 4.01 1 9.07 .00* .08 Grade*Sex .75 .38 2 .85 .43 .02 Grade*School .00 .00 2 .04 .96 .00 Sex*School .00 .00 1 .00 .94 .00 Grade*Sex*School .00 .04 2 .11 .90 .00 Error 44.15 .44 100 *p< .05 Female Classmates The fourth analysis, represented in Tabl e 4-9, evaluated perceived level of voice with female classmates and grade level, sex of student, and type of school the student attended. The analysis yielded no signifi cant findings, indicating that there was no difference in students’ perceived level of voice with their female classmates based on grade level, the sex of the student, and the t ype of school the student attended (see Table 4-8 for means and standard deviations). Table 4-8. Mean scores for level of voice with female classmates Grade 6th grade 7th grade 8th grade M 2.76 2.84 2.84 SD .64 .71 .75 School Urban Rural M 2.75 2.94 SD .72 .63 Sex Male Female M 2.76 2.83

PAGE 35

28 SD .59 .74 Table 4-9. Three-way analysis of variance for level of voi ce with female classmates___ SS MS DF F p 2 Grade .39 .20 2 .39 .68 .01 Sex .23 .23 1 .47 .50 .01 School .19 .19 1 .37 .13 .02 Grade*Sex .58 .29 2 .57 .57 .01 Grade*School .40 .20 2 .40 .67 .01 Sex*School .05 .05 1 .10 .75 .00 Grade*Sex*School .67 .34 2 .67 .52 .01 Error 50.28 .50 100 Research Question III Parents Table 4-10 includes the correlations between level of voice with parents and the predictor variables. Based on the correlation matrix, Level of Voice was correlated the most, although moderately, with would feel stupid, don’t have an opinion, and good woman stereotype within the Reasons scale. This means that students whose level of voice with parents is low are more likely to cite would feel stupid, don’t have an opinion, or the good woman stereotype (if female) as a reason for not expressing their opinions. Level of Voice was not correlated with the second subscale Support for Voice Support for voice was not significantly correlated with good woman stereotype or don’t have an opinion However, Support for Voice was significantly correlated with lack of validation threat to the relationship and would feel stupid This means that students’ who perceive their pa rents as not supporting their vo ice are more likely to cite lack of validation threat to the relationship or would feel stupid as reasons for not expressing their true opinions. Within the Reasons subscale, the highest inte rcorrelations were between lack of validation and threat to the relationship would feel stupid and threat to the relationship and would feel stupid and good woman stereotype This means that students who cited

PAGE 36

29 threat to the relationship as a reason for not expressing th eir true opinions were likely to also cite lack of validation and would feel stupid Similarly, students who cited would feel stupid as a reason for not expres sing their true opinions were likely to also cite good woman stereotype Table 4-10. Pearson correlations for level of voice with parents Subscale 1 2 3 4 5 6 7 1. Level of Voice --.26** -.30**-.40**-.41** -.34** .12 2. Lack of Validation -.68**.46**.62** .48**-.59** 3. Threat to the Relationship -.48**.55** .51**-.32** 4. Don’t Have an Opinion -.51** .65**-.15 5. Would Feel Stupid -.68**-.28** 6. Good Woman Stereotype --.10 7. Support for Voice -** p =.001 Teachers Table 4-11 includes the correlation matrix for level of voice with teachers and the predictor variables. Based on the correlation matrix, level of voice was correlated the most, although moderately, with don’t have an opinion and good woman stereotype This means that students whose level of voice with teachers is low are more likely to cite don’t have an opinion or good woman stereotype as a reason for not expressing their opinions. Level of Voice was not significantly correlate d with the second subscale Support for Voice. Support for Voice was not significantly correlated with don’t have an opinion and good woman stereotype Support for Voice was significantly correlated with lack of validation threat to the relationship and would feel stupid This means that students who perceive their teacher s as not supporting their voice are more likely to cite lack of validation threat to the relationship or would feel stupid as reasons for not expressing their true opinions.

PAGE 37

30 Table 4-11. Pearson correlations for level of voi ce with teachers Subscale 1 2 3 4 5 6 7 1. Level of Voice --.17 -.21* -.30** -.17 -.34** .14 2. Lack of Validation -.60**.51**.66** .54**-.49** 3. Threat to the Relationship -.50**.63** .57**-.36** 4. Don’t Have an Opinion -.52** .59**-.10 5. Would Feel Stupid -.56**-.40** 6. Good Woman Stereotype --.15 7. Support for Voice -* p = .05. ** p =.001. Within the Reasons subscale, the highest interc orrelations occurred between lack of validation and threat to the relationship lack of validation and would feel stupid and threat to the relationship and would feel stupid This means that students who cited lack of validation as a reason for not expressing their tr ue opinions are likely to also cite threat to the relationship and would feel stupid Similarly, students’ who cited threat to the relationship as a reason for not expressing their true opinions are likely to also cite would feel stupid Male Classmates Correlations between level of voice that st udents had with their male classmates and the predictor variables are reported in Table 4-12. The results showed that Level of Voice was correlated most, although moderately, with lack of validation would feel stupid and good woman stereotype. This means that students whose level of voice with male classmates is low are more likely to cite lack of validation, would feel stupid and good woman stereotype as a reason for not expressing th eir opinions. Unlike all other relational contexts, Level of Voice was correlated, although mode rately, with the second subscale Support for Voice with male classmates. Support for Voice was not significantly correlated with good woman stereotype However, it was significantly correlated with all other predictor vari ables. The highest

PAGE 38

31 significant correlations for Support for Voice were with lack of validation and would feel stupid This means that students’ who perceive their male classmates as not supporting their voice are more likely to cite lack of validation or would feel stupid as reasons for not expressing their true opinions. Within the Reasons subscale, the highest interc orrelations occurred between lack of validation and would feel stupid threat to the relationship and would feel stupid and don’t have an opinion and good woman stereotype This means that students who cited would feel stupid as a reason for not expressing their true opinions were likely to also cite lack of validation and threat to the relationship Similarly, students who cited don’t have an opinion as a reason for not expressing their true opinions were likely to also cite good woman stereotype Female Classmates Table 4-13 includes the correlation coefficien ts of student’s perceptions of their level of voice with their female classmates and the predictor variables. Based on the correlation matrix, Level of Voice was correlated most, a lthough moderately, with lack of validation, would feel stupid, and good woman stereotype This means that students whose perceived level of voice with female clas smates was low were more likely to cite lack of validation would feel stupid or good woman stereotype (if female) as a reason for not expressing their opinions. Level of Voice was not significantly correlated with the second subscale Support for Voice Support for Voice was not significantly correlated with don’t have an opinion and good woman stereotype However, Support for Voice was significantly correlated with lack of validation threat to the relationship and would feel stupid This means that students who perceived their female classmat es as not supporting their voice were more

PAGE 39

32 likely to cite lack of validation threat to the relationship or would feel stupid as reasons for not expressing their true opinions. Within the Reasons subscale, the highest interc orrelations occurred between lack of validation and would feel stupid lack of validation and threat to the relationship and good woman stereotype and don’t have an opinion This means that students who cited lack of validation as a reason for not expres sing their true opinions we re likely to also cite would feel stupid and threat to the relationship Also, students who cited don’t have an opinion as a reason for not expressing their tr ue opinions were like ly to also cite good woman stereotype Table 4-12. Pearson correlations for le vel of voice with male classmates Subscale 1 2 3 4 5 6 7 1. Level of Voice --.32** -.24* -.23* -.31** -.28* .28** 2. Lack of Validation -.63** .60**.68** .57**-.51** 3. Threat to the Relationship -.56**.70** .53**-.34** 4. Don’t Have an Opinion -.64** .67**-.26** 5. Would Feel Stupid -.58**-.46** 6. Good Woman Stereotype --.19 7. Support for Voice -* p = .05. ** p = .001. Table 4-13. Pearson correlations for le vel of voice with female classmates Subscale 1 2 3 4 5 6 7 1. Level of Voice --.31** -.22** -.06 -.30** -.23** .14 2. Lack of Validation -.65**.57**.72** .63**-.45** 3. Threat to the Relationship -.41**.60** .50**-.39** 4. Don’t Have an Opinion -.56** .70**-.12 5. Would Feel Stupid -.54**-.37** 6. Good Woman Stereotype --.14 7. Support for Voice -** p = .001 Simple Multiple Regression A multiple regression analysis was conducte d to examine the degree of association between level of voice and the explanatory variables, Reasons and S upport for Voice for

PAGE 40

33 students’ relationships with their parents, teachers, male classmates, and female classmates. The proceeding tables report the unstandardized regression coefficients (B), the standardized regression coefficients ( ), the observed t-values (t), the significance level (p) and the squared semi-partial correlations (r2) for each relationship. Parents A multiple regression analysis was conducte d for each of the two predictors for each relationship. The first analysis evalua ted students’ perceived level of voice with their parents and included the Reasons and Support for Voice scales. Table 4-14 shows the results of the analysis. The analysis showed that the model represented by the Reasons and Support for Voice scales was significant, R2 = .32, adjusted R2 = .26, F (6, 63) = 4.89, p = .00. However, based on the t-statis tic, there were only two predictor variables within the Reasons scale that made a significant contribution to the predictive measures. The predictor variable don’t have an opinion t(69) = -2.27, p = .03, and the category would feel stupid t(69) = -2.37, p = .02 were the only significant categories. Table 4-14. Level of voice with pa rents multiple regression estimates Predictors B t p sr aLack of Validation .04 .04 .21 .83 .03 aThreat to the Relationship -.06 -.07 -.41 .68 -.05 aDon’t have an opinion -.39 -.37 -2.27 .03* -.28 aWould feel stupid -.38 -.37 -2.37 .02* -.29 aGood woman stereotype .14 .17 1.02 .31 .13 Support for voice -.02 -.01 -.09 .93 -.01 p < .05. sr = semipartial correlati on. a = categories represent the Reasons scale. Model: R = .57, R2 = .32, adjusted R2 = .26, F = 4.89, df = 6, p = .00*. Teachers Table 4-15 shows the multiple regression analysis that evaluated the strength of the predictors in determining stude nts’ perceived level of voice wi th their teachers. The two predictors were significantly related to level of voice for teachers R2 = .21, adjusted R2 =

PAGE 41

34 .13, F (5, 65) = 3.13, p = .01. However, based on the evaluation of the t-statistics, the category, don’t have an opinion was the only predictor variable that resulted in a significant outcome, t(69) = -1.98, p = .05. Table 4-15. Level of voice with teac hers multiple regression estimates___________ Predictors B t p sr aLack of Validation -.09 -.10 -.60 .55 -.07 aThreat to the Relationship .23 .21 1.30 .20 .16 aDon’t have an opinion -.30 -.23 -1.98 .05* -.24 aWould feel stupid -.08 -.01 -.05 .96 -.01 aGood woman stereotype -.23 -.24 -1.55 .13 -.19 Support for voice .14 .13 .97 .34 .12 p < .05 sr denotes semipartial correla tion. a = categories represent the Reasons scale. Model: R = .45, R2 = .21, adjusted R2 = .13, F = 2.76, df = 6, p = .02* Male Classmates The third analysis (Table 4-16) evaluated the strengt h of the predictors in determining students’ perceived level of voice with their male classmates. The Reasons and Support for Voice scales were significantly relate d to perceived level of voice for teachers R2 = .19, adjusted R2 = .11, F (6, 63) = 2.39, p = .04. However, based on the evaluation of the t-statistics, there were no ot her categories that re sulted in significant outcomes. Table 4-16. Level of voice with male cl assmates multiple regression estimates______ Predictors B t p sr aLack of Validation -.03 -.03 -.14 .89 -.02 aThreat to the Relationship .16 .16 .96 .34 .12 aDon’t have an opinion .01 .01 .06 .95 .01 aWould feel stupid -.24 -.28 -1.52 .13 -.19 aGood woman stereotype -.09 -.12 -.77 .44 -.10 Support for voice .20 .22 1.58 .12 .20 p < .05. sr denotes semipartial correla tion. a = categorie s represent the Reasons scale. Model: R = .43, R2 = .19, adjusted R2 = .12, F = 2.39, df = 6, p = .04* Female Classmates The fourth analysis (Table 4-17) evalua ted the strength of the predictors in determining students’ perceived level of voice with their female classmates. The analysis

PAGE 42

35 showed that the Reasons and Support for Voice predictors were not significantly related to level of voice for female classmates R2 = .17, adjusted R2 = .09, F (6, 63) = 2.20, p = .06. Additionally, based on the evaluation of the t-statistics, there were no categories that resulted in significant outcomes. Table 4-17. Level of voice with female classmates multiple regression estimates____ Predictors B t p sr aLack of Validation -.23 -.20 -1.06 .29 -.13 aThreat to the Relationshi p -.04 -.04 -.30 .77 -.04 aDon’t have an opinion .29 .26 1.54 .13 .19 aWould feel stupid -.24 -.25 -1.48 .15 -.18 aGood woman stereotype -.10 -.11 -.64 .52 -.08 Support for voice .07 .07 .56 .58 .07 p = .05 sr denotes semipartial correla tion. a = categories represent the Reasons scale. Model: R = .42, R2 = .17, adjusted R2 = .09, F = 2.20, df = 6, p = .06.

PAGE 43

36 CHAPTER 5 DISCUSSION My study had two primary goals. The first goal was to give a preliminary review of level of voice among a population of students (African-Americans in early adolescence) that have yet to be considered. The second goal was to assess whether the Teenage Voice survey provided scales that were good predictors of level of voice for these students. Internal Consistency and Reliabilities Research question 1 evaluated the internal consistency and reliability estimates for the three predictors of level of voice in the survey; Level of Voice Reasons and Support for Voice Harter and colleagues (1997) reported that across three studies using the Teenage Voice measure, internal consiste ncy was high, ranging from .82 to .91. The results of my study yielded moderate internal consistency and reliability results. The internal consistency of the survey ranged from .66 to .75 for the Level of Voice scale, .64 to .80 for the Reasons scale and .77 to .84 for the Support for Voice scale. For most of the items in the three scales, item analysis determined that most of questions did not increase the coefficient alpha level if one or more items were deleted. Based on the item analysis, there was one ques tion from the Level of Voice scale and two questions from the Reasons scale that if deleted would ha ve yielded an increased alpha level; however, these increases would have been negligible. Given this, it seems that the Teenage Voice survey was an appropriate measure for this population of students.

PAGE 44

37 Level of Voice There were some interesting findings that resulted from the analysis of research question 2 which determined whether or not level of voice was different for students based on grade, sex, and type of school attended, and whether there were any interactions. Based on the analyses, there we re differences in level of voice depending on the relational context. Level of Voice with Parents In determining the student’s level of voice w ith their parents, analysis showed that there were significant differences in level of voice depending on a student’s grade and school; however, there was no interaction between the two. Although there was a significant effect because of grade level, a dditional post hoc analysis showed that there were no significant differences between grad es. Perceived level of voice with parents does appear to increase graduall y with age. More research is needed, possibly with a larger sample size or broader age range, to determine if significant age-related differences in perceived level of voice with parents do exist. Additiona lly, students in rural schools perceived their level of voice with parent s to be higher than did students from urban schools. Research question 3 asse ssed whether or not the Reasons and Support for Voice scales were good predictors of level of voice for each relational context. The first part of the analysis of the third research ques tion evaluated level of voice using Pearson correlation coefficients. Based on correlation coe fficients, linear trends for level of voice and the predictor scales were different depending on the relati onal context. In evaluating level of voice with parents, Level of Voice had the highest significant correlations with would feel stupid and don’t have an opinion within the Reasons scale. This means that

PAGE 45

38 students whose perceived level of voice with parents was low were more likely to cite would feel stupid or don’t have an opinion as a reason for not expressing their true opinions. Level of Voice with Teachers For level of voice with teacher s, the analysis showed that there were differences in level of voice based on the sex of the student Male students percei ved a higher level of voice with their teachers than female students. This finding is consistent with research which suggests that in the classroom setti ng, boys receive more attention from teachers than girls (AAUW, 1992). Additionally, for African-American girl s, there are fewer interactions with classroom teachers when compared to th eir Caucasian counterparts, even though there is evidence to suggest that African-America n girls attempt to initiate interactions with their teachers more fre quently than their Caucasian counterparts (AAUW, 1992). In evaluating level of voice with teachers, Level of Voice had the highest significant correlations with don’t have an opinion and good woman stereotype within the Reasons scale. This means that students whose perc eived level of voice with teachers was low were likely to cite don’t have an opinion or good woman stereotype as a reason for not expressing their true opinions. It was an in teresting finding that female students in my study reported a low level of voice with teache rs in comparison with male students and also cited good woman stereotype as a reas on for not expressing their true opinions. Further examination is needed for g ood woman stereotype and its impact on African-American females. As Taylor et al. (1995) said, African-American girls are often described as being socia lized less stereotypically than their Caucasian counterparts. The good woman stereotype entail s girls being taught a traditi onal female role, to refrain

PAGE 46

39 from sharing their opinions, and to a void conflict at all costs. Although African-American girls are taught this traditional view, they are also encouraged to be strong, self sufficient, and fight back (verbally and physically) when they believe that an injustice has occurred against them, which may be contradictory to th e traditional view in many instances. Further examination with African-American females and good woman stereotype is warranted. With the excepti on of one teacher, all of the students across schools had Caucasian teachers. It is unclear whether or not this cultural difference played a role in whether African-American gi rls used the traditiona l role or the role supported by culture in the classroom. Level of Voice with Male Classmates With male classmates, perceived leve l of voice was significantly different depending on the type of school that students attended. Stude nts in the rural schools felt they had a higher level of voice with their male classmates than those in the urban school. It is interesting to note that there was no significant intera ction with leve l of voice and sex, indicating that male students did not percei ve a greater level of voice with their male classmates than female students did. Previous literature from Harter (1996) suggests that level of voice is highest among close frie nds, followed by classmates. Future study should examine whether the addition of close friends as a relational context would yield significant differences in how students perceive their level of voice with their peers. For level of voice with male classmates, the highest sign ificant correlations were for lack of validation and would feel stupid within the Reasons scale. This means that students who feel their level of voice with male classmates is low are more likely to cite lack of validation or would feel stupid as a reason for not e xpressing their opinions.

PAGE 47

40 Level of Voice with Female Classmates In determining level of voice with female classmates, the analysis showed that there were no significant differe nces in level of voice based on sex, grade, or type of school attended. Similar to level of voice with male classmates, it is interesting to note that female students did not have a significantly higher leve l of voice with their female classmates. For level of voice with female classmates the highest correlations within the Reasons scale were for lack of validation and would feel stupid which mirrors the correlations for male classmates. These reas ons may have played a role in why there were no significant differences for sex for either male or female classmates. As previously mentioned, it would be interesti ng to include an examination of student’s relationships with close friends and determine if students w ith a low level of voice with close friends would cite lack of validation and would feel stupid as reasons for suppressing their true opinions. For level of voice with parents, teache rs, and female classmates, there was no significant correlation between Level of Voice and the subscale, Support for Voice This is an interesting result, cont radicting the claim of Harter and her colleagues (1997) that lack of support, particularly from parents, is a major factor that causes children of both genders to suppress their true level of voice. Additionally, for level of voice with parents, teachers, male classmates, and female cla ssmates, there was no significant correlation between the subscale Support for Voice and the good woman stereotype Similarities and Differences from Previous Studies One of the major conclusions made by Hart er et al. (1996, 1998) was that perceived support for voice was predictive of level of voi ce. However, evidence from this research

PAGE 48

41 showed that for this population of stud ents, perceived support for voice was not predictive of level of voice. This differen ce in conclusions can be explained in a few ways. First, since this is the first tim e that this measure has been used on an African-American population, this is the only ev idence to suggest that this population of students may derive their level of voice in a nother way. A different set of items that would explore other ways that level of voice is derived is needed. Secondly, these students may be defining support for voice di fferently than Caucasian populations who have used the measure before. Despite the f act that internal cons istencies from my study showed that the Support for Voice subscale had some of the highe st alpha levels in out of the entire survey, this did not translate into significant correlations between level of voice and support for voice. And lastly, my study did not have as many participants as previous studies by Harter, which may have impacted overall outcomes from the measure. Similar to conclusions found by Gilligan (1982), there is some preliminary evidence in my study to suggest an age trend in perceived level of voice with parents. As grade level increased, perceive d level of voice with parent s increased, though the group means were not significantly different fr om each other. With future study using additional participants and a broader age ra nge, it would be interesting to examine whether or not there is a true diffe rence in level of voice based on age. Harter’s previous studies (1994, 199 6, 1998) concluded that based on her quantifiable measure, there are no sex differen ces in perceived level of voice except for a sex difference in perceived level of voice w ith teachers. Female students showed significantly less voice than male students with teachers. However, there were

PAGE 49

42 differences in perceived level of voice ba sed on gender orientation. Those students who had a more feminine orientation, regardless of sex, exhibited a lowe r level of perceived level of voice than those students who ha d an androgynous orientation. Based on the results of my study, male and female students differed in their perceived level of voice with teachers. Males perceived their level of voice to be high er with their teachers than female students. The gender orientation type s described by Harter were not assessed in my study. Given previous discussions on the good woman stereo type and the less stereotypical nature of African-American female socialization, it would be interesting to examine whether these students perceived gender orientation in the same way as Caucasian students from previous studies. Unlike previous studies of level of voi ce, my study included students from rural schools. These students, based on the cont ext of their environment and the small community atmosphere, may perceive their leve l of voice differently than urban students. Results from my study showed that students in the rural schools perc eived their level of voice to be higher with parents and male classmates than did students in the urban school. Although this difference may be attribut ed to smaller school size and smaller communities as a whole, additional research is needed to determine the factors that increase students’ perceived leve l of voice with this population. Limitations My study involved recruiting participants whose voices are currently not sufficiently acknowledged in th e literature. There were limitations to adequately studying this population of students. A lthough there was an adequate amount of participation from students who attended urba n schools, the recruitm ent of rural students should have been greater so th at level of voice for these st udents could be shown through

PAGE 50

43 a more representative population. However, based on the large number of African-American students in the urban school versus the small pool of eligible students in the rural school, it could be expected that the urban school would be better represented in the study, thereby increasing the likelihood of generalizability. Similarly, there were a greater amount of female students who partic ipated in the study. To increase the likelihood of generalizability, special attention should be gi ven to increasing the amount of male participation. Future Research Harter (1996) believes that the media play a role in perpetuating the good woman stereotype. Given the percepti ons of African-American female s that are portrayed (or not portrayed) in the media, it may be the case th at this group of students is not submitting to a global good woman stereotype, but rather a stereotype of African-American woman. Although there is an absence of significan t theoretical litera ture concerning African-American sex-role classification and stereotypes, Dade and Sloan (2000) cited research to support the claim that within African-American families, male and female socialization is interchangeable, flexible, and fluid. Harter an d colleagues (1997, 1998) gave supporting evidence to suggest that an drogynous males and females, versus those that submit to feminine and masculine roles, are more likely to have a higher level of voice. What may be seen as feminine in the European tradition may be seen as androgynous for African-Americans. Future re search could examine the prevalence of a gender-specific stereotype in African-Americans, whether this plays a role in lower levels of voice, and whether Harter’s initial findings are genera lizable to this population. Another study that would be of interest is to evaluate the silencing mechanisms that take place in classrooms, as defined by Fine ( 1993) against the tenets of level of voice as

PAGE 51

44 described by Harter (1996). Specifically, what is the climat e and culture of the classroom environment and how does it impact perceived le vel of voice in students? Since literature on a quantifiable measure on silencing in the classroom is absent, such a study would need to use a mixed-methods approach, us ing Harter’s (1996) Teenage Voice Survey with the qualitative methods used by Fine (1992). The classroom environment is increasingly evolving as various legislativ e measures are calling for more accountability in the classroom. Given the prevalence of standardized testing and the greater accountability that schools have in increasing scores, and th e limited time that teachers have to teach the skills needed to pass thes e exams, the environment may invite dialogue that is guided toward the goal of test prep aration and silence othe r forms of dialogue. Based on the results of my study, it is ev ident that studying level of voice among this population of students dese rves attention. Although ther e were many questions that were answered, there are many more that ne ed further study. It is hoped that this preliminary research will serve as th e starting point for more studies about African-American students and level of voice.

PAGE 52

45 LIST OF REFERENCES American Association for University Women. (1992). How schools shortchange girls. Washington, DC: American Association of University Women Educational Foundation. Barton, P.E. (2003). Parsing the achievement gap: Baselines for tracking progress (Policy Information Center, Education Te sting Service). Retrieved March 8, 2005, from http://www.ets.org/research/pic/parsing.pdf Beeson, E. & Strange, M. (2000). Why rural ma tters: The need for every state to take action on rural education. Journal of Research in Rural Education, 16(2), 63-140. Beeson, E. & Strange, M. (2003). Why rura l matters 2003: The continuing need for every state to take ac tion on rural education. Journal of Research in Rural Education, 18(1), 3-16. Dade, L.R. & Sloan, L.R. (2000). An investig ation of sex-role ster eotypes in African Americans. Journal of Black Studies, 30(5), 676-690. Diaz-Greenberg, R. (2001). The emergence of voice in Latino/a high school students. New York: Peter Lang Publishing. Fine, M. (1987). Silencing in public schools. Language Arts, 64(2), 157-174. Fine, M. (1991). Framing Dropouts: Notes on the po litics of an urban public high school. Albany, NY: State University of New York Press. Fine, M. (1992). Disruptive voices: The possibil ities of feminist research. Ann Arbor, MI: University of Michigan Press. Gilligan, C. (1982). In a different voice: Psychological theory and women’s development. Cambridge, MA: Harvard University Press. Gilligan, C. (1993). Joining the resistance: Psychology, politics, girls, and women. In Lois Weis and Michelle Fine (Eds.). Beyond Silenced Voices (pp. 143-168). Albany, NY: State University of New York Press. Harter, S. (1996). Teacher and classmate influences on scholastic motivation, selfesteem, and level of voice in adolescents. In J. Juoven & K.R. Wentzel (Eds.). Social motivation: Understandi ng children’s school adjustment. Cambridge, MA: Cambridge University Press.

PAGE 53

46 Harter, S., Waters, P.L., & Whitesell, N.R. ( 1997). Lack of voice as a manifestation of false self-behavior among adolescents: The school setting as a stage upon which the drama of authenticity is enacted. Educational Psychologist, 32(2), 153-173. Harter, S., Waters, P.L., Whitesell, N.R., & Ka stelic, D. (1998). Level of voice among female and male high school students: Relational context, support, and gender orientation. Developmental Psychology, 34, 892-901. Leander, K. (2002). Silenci ng in classroom interaction: Producing and relating social spaces. Discourse Processes, 34(2), 193-235. Leone, P.E, Christle, C.A., Nelson, C.M., Skiba, R., Frey, A. & Jolivette, K. (2003). School failure, race, and disability: Pr omoting positive outcomes, decreasing vulnerability for involvement w ith the juvenile delinquency system (The National Center on Education, Disability and Juve nile Justice). Retrieved March 8, 2005, from http://www.edjj.org/Publica tions/list/leon e_et_al-2003.pdf Oldfather, P. & Dahl, K. (1994). Toward a so cial constructivist reconceptualization of intrinsic motivation for literacy learning. Journal of Reading Behavior, 26(2), 139158. Quiroz, P.A. (2001). The silencing of Latino student “voice”: Puerto Rican and Mexican narratives in eighth grade and high school. Anthropology & Education Quarterly, 32(3), 326-349. Tatum, B.D. (1997). “Why are all the Black kids sitting together in the cafeteria?”: And other conversations about race. New York: Basic Books. Taylor, J.M., Gilligan, C., & Sullivan, A.M. (1995). Between voice and silence: Women and girls, race and relationship. Cambridge, MA: Harvard University Press. Wardle, F. & Cruz-Janzen, M.I. (2004). Meeting the needs of multiethnic and multiracial children in schools. Boston: Allyn and Bacon.

PAGE 54

47 BIOGRAPHICAL SKETCH Yashica Crawford is a 4th-year doctoral student studyi ng educational psychology at the University of Florida. Originally fr om San Francisco, she completed a B.A. in political science and a B.S. in psychology magna cum laude at Howard University (Washington, D.C.) in 2001. She has devoted her research program and personal endeavors to understanding and supporting at-risk and resilient students. She resides in Tampa with her husband, Cesnae; and their dog, Bison.