A Longitudinal Model of School Climate, Social Justice Orientation, and Academic Outcomes Among Latina/o Students
by Leyla M. Pérez-Gualdrón & Janet E. Helms - 2017
Background: Social justice orientation (SJO) is the motivation to promote justice and equity among all in society. Researchers argue that students of Color with high SJO can resist structural racism in their schools/society and have positive academic outcomes.
Purpose: In the present study, a longitudinal model of cultural and environmental predictors (i.e., school relational climate, school language climate, Spanish language background, and English proficiency) and civic/educational outcomes (i.e., community engagement, grades, school engagement, school dropout) of SJO among Latina/o youths was developed and tested.
Participants: The study was conducted with a subsample of Latinas/os taken from the National Education Longitudinal Study of 1988. Participants were enrolled in eighth grade (N = 1,472), sampled from different schools and regions in the U.S., and followed through three waves of data collection from 8th through 12th grade.
Research Design: A longitudinal, correlational design was used to explore the association among the constructs studied.
Data Collection and Analysis: Secondary data analyses were conducted. Structural equation modeling techniques were used to analyze the data.
Results: Early school relational climate (8th grade) was a positive predictor of SJO, which in turn predicted more community and school engagement, higher grades, and decreased likelihood of dropping out of school (12th grade) via personal agency. In addition, school language climate and language skills predicted a greater sense of personal agency, which in turn predicted higher grades and a decreased likelihood of dropping out.
Conclusions: The results of the present study underscore the importance of strengths-based and cultural approaches in education in a sample of Latina/o students. Specifically, close attention should be paid to school cultural climate variables in which positive relational climates and cultural language climates are addressed in schools. The integration of sociopolitical context, critical consciousness, and SJO may be key factors in improving the educational and counseling experiences of Latina/o youths.
Latina/o children under 18 years of age are the second largest group of students in the educational system in the U.S. after Whites (Dolan, 2009; National Center for Education Statistics [NCES], 2015). Despite the presence and growth of Latinas/os in U.S., Latina/o youths are among the most socioeconomically and educationally disenfranchised groups, facing significant educational barriers and deficient school environments (Fine, Burns, Payne, & Torre, 2004; Miller & Garran, 2008). These factors may promote low expectations and/or deficit perspectives when understanding the educational experiences of Latina/o students. However, it is imperative to consider the role of context and Latina/o students cultural strengths in their academic achievement when informing counseling and educational research and practice.
Specifically, Latina/o students are powerful agents of change for educational justice, as they have led important movements to promote educational equity and success. Examples of this include the 1968 Chicano/a students school walkouts in East Los Angeles, where students demanded equitable educational access and resources; and more currently, Latina/o students involvement in activism for the Development, Relief and Education of Undocumented Minors Act (DREAM Act). Youths movements like these are a form of sociopolitical development and social justice orientation (SJO), which have been associated with greater educational engagement and positive outcomes among students of Color (Oakes, Rogers, & Lipton, 2006; Watts & Flanagan, 2007). Therefore, when considering Latina/o educational achievement, it is central to study the role of cultural strengths and their potential to be powerful agents of change to promote educational/social justice.
Watts, Williams, and Jagers (2003) proposed a model to explain how Black adolescents may be able to resist educational barriers and become academic achievers by recognizing oppression and resisting it. Applying Watts et al.s model to Latina/o students may imply that if students have acquired the knowledge, analytical skills, emotional faculties, and the capacity for action in political and social systems necessary to interpret and resist oppression, they will be academically successful (p. 185). At the core of Watts et al.s model of sociopolitical development is the capacity to acknowledge unfairness and oppression (i.e., critical consciousness) and being motivated to change them for promoting social justice (i.e., SJO). Namely, sociopolitical development consists of the following four components: (a) critical awareness of unfairness and oppression; (b) motivation for change for social justice; (c) emotional and personal maturity and self-confidence; (d) action, activism, and resistance to oppression and community engagement.
In the present study, we adopted sociopolitical development theory (Watts, Griffith, & Abdul-Adil, 1999) to investigate the question of whether SJO (a component of sociopolitical development) may enable Latina/o students to resist factors that lead to poor academic outcomes. Watts et al. (1999) proposed a relevant theoretical model for SJO in which they elaborated on the importance of critical consciousness and action to resist oppression during adolescence. This theoretical model was further elaborated as a framework to study SJO in Latina/o youths in the present study.
SOCIAL JUSTICE ORIENTATION
There are multiple definitions of SJO in the educational and psychological literature (North, 2008). Congruent with Watts et al.s (2003) sociopolitical development theory, in the current study, SJO is defined as the desire and motivation to end societal inequalities and unfairness for the collective benefit of ones community.
Researchers have suggested that having an SJO would facilitate students capacity to resist structural racism in their school environments (Cammarota, 2004; Diemer, Hsieh, & Pan 2009; Ginwright & Cammarota, 2007; Watts et al., 1999; Watts et al., 2003). Furthermore, SJO has been associated with positive academic outcomes and personal agency beliefs in youths of Color with a history of oppression and negative school experiences, including Latina and Latino students (Cammarota, 2004; OConnor, 1997; Solórzano & Bernal, 2001). Accordingly, Solórzano and Bernal (2001) suggest that when Latina/o students are critical of systems of oppression and are motivated for social justice, they may be more engaged educationally and civically, especially if education is seen as a way to end oppression.
ENVIRONMENTAL BARRIERS OR FACILITATORS: SCHOOL RELATIONAL AND LANGUAGE CLIMATES
Over the years, theorists and researchers have proposed different definitions of school climate (Cohen, 2014; Thapa, 2013). Furthermore, with recent federal attention to and support for school climate reform, there has been increased theoretical and practical attention to this construct (Cohen, 2014). However, there continues to be great diversity in the way in which school climate is understood (Thapa, Cohen, Guffey, & Higgins-DAlessandro, 2013). A widely used definition of school climate proposes that school climate refers to the quality and character of school life. It is based on patterns of school life experiences and reflects norms, goals, values, interpersonal relationships, teaching, learning and leadership practices, and organizational structures (National School Climate Council [NSCC], 2007). Though this definition includes many important aspects of school climate, we also propose an ecological approach to assess school climate. Thus, we consider school climate to also be embedded in, and bidirectionally related to, sociopolitical contexts and culture (Bronfenbrenner, 1994). Therefore, in addition to the factors listed in the definition above, specific cultural aspects of school climate should be considered when understanding Latina/o students experiences.
To better assess and understand cultural aspects of school climate, as it pertains specifically to Latina/o students, the NSCCs (2009) perspective on school climate dimensions is helpful. They propose that school climate has four distinct dimensions: (a) safety, which refers to rules and norms in the school as well as physical, emotional, and social safety; (b) relationships, respect for diversity, social support from adults and from students; (c) teaching and learning, support for learning, social and civic learning, career and professional development; and (d) institutional environment, physical surroundings as well as school connectedness and engagement (Cohen, Pickeral, & McCloskey, 2009). All of these dimensions are impacted by context, public policy, and resources. As Fine, Burns, Payne, & Torre (2004) note, the systems of educational inequities have deep psychological and educational impacts on children and youths. In the present study, we focus on school relational climate (school climate relational dimension) and school language climate (school climate teaching/learning supports dimension) as pertinent to Latina/o students.
School Relational Climate
From the school climate dimensions previously listed, school relationships (i.e., school relational climate) may be of particular relevance for Latina/o students. Theorists and researchers have consistently described Latina/o groups as collectivistic and relationally oriented (e.g., Helms & Cook, 1999; La Roche & Shriberg, 2004; Marín & Marín, 1991; Triandis, 1994). That is, although there are individual differences, generally Latinas/os may tend to emphasize personal interdependence and sensitivity to their social environments over individual interests (Helms & Cook, 1999). Further, Latinas/os may tend to define their identities and behaviors in relation to others, prioritizing relationships with others and group goals rather than individual goals (La Roche & Shriberg, 2004).
In the present study, relational school climate is defined as the students perceptions of the quality of the relationships within their schools, including relationships of students with teachers and school staff, as well as with their fellow students and peers. For Latinas/os, issues of fairness, respect for diversity, and teachers support might be important aspects of the school relational climate. For instance, in a random sample of diverse middle school students, Schneider and Duran (2010) found that Latina/o students considered their relationships with teachers a very central aspect of school climate. This was different from the findings observed in the other groups surveyed in the same study (i.e., Asians, Whites). Therefore, it is important to further evaluate the association between relational school climate and Latinas/os academic engagement and achievement.
Little is known about how school relational climate/environments may (or may not) foster SJO development. However, school environments play a key role in the lives and socialization of Latina/o students. In schools, learning and important developmental tasks such as racial identity and personal agency beliefs occur (Fine et al., 2004; Helms, 2003). Several researchers also have suggested that early school climate has a long-term impact on students academic and socioemotional outcomes (e.g., Fine et al., 2004; Lindholm-Leary & Borsato, 2005; Quintanar-Sarellana, 2004). One perspective suggests that a positive relational climate, where students feel they belong and there are social justice-oriented mentors, may be conducive to greater motivation and desire to change social unfairness and inequities (Rubin, 2007). Another perspective suggests that a negative school climate, in which students directly experience oppression, may also foster a motivation to change such environments and/or make them less unfair due to the students own experiences and awareness (Cammarota, 2004). Therefore, we explore the association between school relational climate and SJO development.
Another important NSCC factor we consider in the present study is supports for teaching and learning. Particularly pertinent for Latina/o students is the consideration of cultural variables, such as language. The debate in different states in the U.S. about language of instruction and/or bilingual education has important implications for students and their families. Specifically, if another language (e.g., Spanish) is considered a deficit to overcome rather than a strength, students cultural and academic identities may be compromised. Therefore, we propose that language resources devoted to learning are relevant school climate aspects to assess when understanding the experiences of some Latina/o students.
School Language Climate and Language Proficiency/Background
A significant number of Latinas/os preserve Spanish language through generations (Portes & Hao, 2002; M. M. Suárez-Orozco & Páez, 2008). Therefore, language of instruction and language supports become a significant issue when considering Latina/o educational needs and experiences. Availability of bilingual education and education that is congruent with students cultural backgrounds is associated with positive academic outcomes for Latinas/os (Quintanar-Sarellana, 2004; Thomas & Collier, 2002). Authors have suggested that early school language climate has long-term academic and socioemotional outcomes (e.g., Quintanar-Sarellana, 2004). Likewise, positive language climates (e.g., bilingual education programs and resources) in schools have been associated with Latina/o students long-term development of language skills in Spanish and English (e.g., Lindholm-Leary & Borsato, 2005; Quintanar-Sarellana, 2004).
Skills in two different languages have been associated with positive outcomes among Latinas/os. Although this may not be true for all Latinas/os, Gonzalez (2009) argued that having proficiencies in both Spanish and English may be related to the development of bicultural identities and skills. These language skills may also translate into positive personal and mental health outcomes, especially when a native language (e.g., Spanish) is considered as an asset to be promoted rather than a deficit to overcome at their schools. For instance, Buriel, Perez, DeMent, Chavez, & Moran (1998) found that students who practiced language brokering, facilitating communication in two languages, felt greater personal agency beliefs in dealing with two different cultures.
Culture and language are often key factors for Latinas/os identities. Gonzalez (2009) contended that when Spanish proficiency is seen as a competency, students may show a commitment to their racial/ethnic group and what happens in society, which is a crucial factor in the development of an SJO (Watts et al., 1999). Portes and Hao (2002) found that fluent bilingual proficiency among Latina/o students was associated with positive outcomes, such as positive psychosocial adjustment and self-esteem. This is perhaps not true for all Latinas/os, as other studies have shown that English proficiency and acculturation in Latinas/os was associated with greater self-efficacy and engagement (e.g., Martínez, DeGarmo, & Eddy, 2004). It is important to also note that many Latinas/os in the United States with strong cultural identities may be monolingual English speakers and/or may identify more with other indigenous languages, which may be different from Spanish.
PERSON-LEVEL PERSONALITY CONSTRUCTS
Social Justice Orientation and Community Engagement
Community engagement or service intended to help others in the community can also be a way of promoting social changes to reflect social justice. Authors have suggested personal and societal benefits from more civically engaged and empowered youths in the face of societal inequities (e.g., Cammarota, 2004; Diemer, 2009; Ginwright & Cammarota, 2007; Solórzano-Bernal, 2001; Watts et al., 1999; Watts et al., 2003). However, little is known about the mechanisms by which SJO may translate into community engagement and action. More empirical evidence is needed to understand the mechanisms by which SJO leads to community engagement. Authors have suggested that the benefits of community engagement are key in the successful transition from late adolescence to early adulthood in adolescents of Color (e.g., Cammarota, 2004; Ginwright & Cammarota, 2007; Watts et al., 2003).
Personal agency is defined as students beliefs in their capabilities to pursue their plans for action and have control over their environments (Bandura, 2001). Theorists and researchers have proposed that emotional and motivational factors (e.g., SJO) may or may not translate into community engagement. For example, Watts and Flanagan (2007) suggested that personal agency beliefs could be a mediating variable to study when observing the relationship between SJO and actual engagement in social action, which is sociopolitical participation for change. Based on her qualitative research, OConnor (1997) also suggested a link between the awareness of collective struggle, SJO, and greater personal agency beliefs in adolescents of Color. Specifically, she argued that high school students experienced greater personal agency beliefs when they were motivated for social justice (SJO).
EDUCATIONAL OUTCOMES: ACADEMIC ENGAGEMENT, GRADES, AND SCHOOL DROPOUT
Latinas/os as a group have among the lowest records of educational attainment, marked by low grades and high dropout rates (Alfaro, Umaña-Taylor, Gonzales-Backen, Bámaca, & Zeiders, 2009; Martínez et al., 2004). These statistics tend to promote deficit-oriented views in education without acknowledging the impact of contextual and sociopolitical issues and Latina/o cultural strengths (Fine, 1986). Furthermore, some authors have suggested that, despite facing multiple oppressive barriers, Latina/o student academic engagement and success may be attributed, in part, to SJO. For example, Cammarota (2004) suggested that students who see education as a way to make a difference and produce societal change may be more academically engaged and successful. Researchers and theorists in counseling and education have also documented the positive association between personal agency beliefs and academic/school engagement (e.g., Bandura, 2001; Close & Solberg, 2008).
Furthermore, several authors have suggested that school engagement is a critical factor to prevent school dropout among Latina/o students (e.g., Conchas, 2001; Martínez et al., 2004; Ream & Rumberger, 2008; Rumberger & Rotermund, 2012). However, school engagement is a multidimensional construct defined and measured in multiple ways by researchers (Christenson, Reschly, & Wylie, 2012; Finn & Rock, 1997; Fredricks, Blumenfeld, & Paris, 2004; Glanville & Wildhagen, 2007). In a review of the educational literature, Fredricks et al. (2004) identified three dimensions of school engagement: (a) behavioral engagement, (b) emotional engagement, and (c) cognitive engagement. Each of these dimensions has been defined and measured in different ways in research. For instance, Fredricks et al. suggested that the operationalization of behavioral engagement has included the measurement of positive conduct in school (i.e., adherence to school norms) and/or the presence/absence of disruptive or negative behaviors (e.g., absenteeism, skipping classes, or getting in trouble).
Similarly, in a 5-year longitudinal study of recent immigrant students in middle/high school (N = 409; Latinas/os n = 240), different kinds of school engagement (i.e., cognitive, relational, and behavioral) were found to be associated with higher grades (C. Suárez-Orozco, Suárez-Orozco, & Todorova, 2008). Overall, behavioral engagement and English proficiency were the most important contributors to academic outcomes in C. Suárez-Orozco et al.s model of academic achievement for recent immigrant youths. The authors also reported that the youths behavioral disengagement increased over time. Although boys and girls presented similar levels of engagement at the beginning of the study, girls were more likely to maintain their levels of engagement while boys engagement was most likely to decrease over time. This pattern of decreased school behavioral engagement over time is consistent with the increased school disengagement observed in inner-city youths in the context of poor quality of education/schools available (e.g., Fine et al., 2004). Therefore, school behavioral disengagement may be a key factor to assess when studying the association of school climate and Latina/o students academic outcomes.
For instance, Solórzano and Bernal (2001) studied school behavioral disengagement, which they described as a form of self-defeating resistance among Latina/o youths when confronting unfairness and injustice. They found that this form of school behavioral disengagement may be associated with having a critique for systems of oppression but a lack of motivation for social justice and change. Additionally, in support of the possibility that Latina/o youths disengage from negative school climate, Cammarota (2004) observed in a longitudinal ethnographic study (N = 6) that some Latina/o students who experienced their schools as oppressive and discriminatory (i.e., experienced negative relational climates) engaged in school truancy and disengagement, which eventually led them to drop out of school. In contrast, other students in his sample, who also experienced their school as oppressive, viewed their education as an instrument for social justice and a way to end oppression. Thus, the latter students were more academically engaged and had higher grades. Cammarota also found that these students had a social justice agenda and were engaged in activism or doing service in their communities.
Hence, for some students, the negative school environments were seen as so oppressive that they avoided them and became behaviorally disengaged from their schools. On the other hand, other students were engaged in community social actions, such as Latina/o community organizations, which emphasized engagement in their Latina/o congruent values, academic work, and civic activities. Congruent with this, researchers have also found that ethnic identity and cultural factors are associated with greater school engagement (e.g., Bingham & Okagaki, 2012). These cultural factors are particularly salient in the face of discrimination or negative school climates as presented above. Taken together, the findings of C. Suárez-Orozco et al. (2008), Cammarota (2004), and Solórzano and Bernal (2001) suggest that school behavioral disengagement may be a key factor to further investigate when assessing academic outcomes of Latina/o students in relation to their school environments and SJO.
THE PROPOSED MODEL: SOCIAL JUSTICE ORIENTATION OF LATINA/O YOUTHS
The purpose of the present study was to develop and test a longitudinal model of factors leading to SJO as well as its civic/academic outcomes. Specifically, school climate variables relevant to Latina/o students were assessed to evaluate their impact on personal-level variables (i.e., Spanish and English proficiencies, SJO, and personal agency), which in turn were also assessed to explore their association with long-term academic and civic outcomes. The hypothesized model is shown in Figure 1. More specifically, the associations between 8th grade school climate variables relevant to Latina/o students (i.e., school relational and language climates) and cultural skills (i.e., Spanish background and English proficiency) were proposed to be associated with the development of SJO in 10th grade. In addition, we proposed that SJO in 10th grade would be associated with community engagement and various academic outcomes in 12th grade (i.e., school behavioral disengagement, grades, and school dropout) via personal agency beliefs. A two-group social justice orientation model for Latina/o youths (SJOLY) was also assessed to explore gender differences in the association among the constructs.
Figure 1. Predictors and Mediators of SJO and Its Academic/Civic Outcomes for Latina/o Students
Single arrow lines represent regression paths among the constructs. The double arrow line represents a correlation path.
The participants were eighth-grade Latina/o students (N = 1,472) of diverse ethnic and racial backgrounds sampled from different schools and regions in the U.S. The sample was taken from the National Education Longitudinal Study of 1988 (NELS:88), a large-scale longitudinal study conducted by the National Center for Education Statistics (NCES). The age range of the participants was 13 years to 16 years (M = 14.46, SD = .65), of whom 49.6% were girls and 50.4% were boys. In addition, 82.7% of the participants were born in the U.S.; the majority self-identified as Mexican, Chicana/o (68.1%). Most (72.6 %) of the students were classified in the low-income quartiles 1 and 2 from the NELS:88 sample. With respect to English proficiency, as deemed by the school records and self-report, the majority of the sample was proficient in English (92.4%), and 14.5% of the students in the sample belonged to a non-English only (e.g., Spanish) household background, while 36% of the participants belonged to a non-English dominant household (e.g., Spanish), according to parents reports. For more information about the sample, please see Table 1. For more information about the NELS:88 sample and procedures, please refer to the National Education Longitudinal Study of 1988: Base Year to Fourth Follow-Up Data File Users Manual (Curtin, Ingels, Wu, & Heuer, 2002).
Table 1. Participants' Demographic and Background Information (N = 1,472)
Note: LEP= Limited English proficiency as determined by the NELS:88 based on student self-report and teachers assessments of the students proficiency in English.
Demographic and Background Information
The participants demographic and background information was obtained from the NELS:88 student, parent, and school administrators surveys, which included self-report questions about place of birth, ethnicity, school location, etc. In addition, the NELS:88 socioeconomic status (SES) composite and limited English proficiency (LEP) composite were used to determine the English proficiency of the research participants. The NELS SES composite is a standardized combination of ratings of parental educational attainment, income, and parents occupation prestige, prepared by NCES to describe their samples socioeconomic characteristics (Ingels et al., 1994). The base year LEP composite was constructed by combining students self-reported English proficiencies and teachers assessments of students English proficiency (Ingels et al., 1994).
School Relational Climate
This variable was defined as the quality of the school environment relationships among teachers and students as determined from the students appraisal. Seven items reliably used in previous studies (e.g., Freidlin & Salvucci, 1995; Gregoire & Algina, 2000) as indicators of school climate were taken from the NELS:88 student base-year questionnaire and analyzed to evaluate their adequacy to assess school relational climate in this sample. Examples of these items are Students get along well with teachers and There is a real school spirit. The participants responded to the items using Likert-type scales (ranging from 1 = strongly agree to 4 = strongly disagree). For the present study, items were reverse scored so that higher scores represented a more positive school relational climate. Cronbachs alpha coefficient for these items was .74 in this sample.
School Language Climate
This variable was defined as the availability of academic instruction and support provided in the students native language at school. School language climate refers to the congruency of the school environment and resources with the students linguistic skills. In the present study, nine items from the eighth grade school administrators surveys pertaining to the availability of bilingual education, courses taught in students native language (i.e., Spanish), percentage of students who were English language learners, and the number of teachers available for English language learner support in eighth grade were used as indicators of language climate. Responses to each item were based on frequencies or percentages available at the schools. Higher item scores or frequencies indicated more language resources and more English language learners at the schools. The items used were consistent with the way other researchers have assessed the presence of language resources available in schools (e.g., Quintanar-Sarellana, 2004; Thomas & Collier, 2002). Cronbachs alpha coefficient obtained in the current sample was .90.
Spanish Language Background
Spanish language background was defined as whether Spanish language was spoken in the students home. To assess this construct, students were asked to respond to the question, What other language is spoken in your home? Students responded to this item by selecting a language from multiple-choice options. Item responses were dummy coded to indicate whether Spanish was spoken at home (0 = No language other than English was spoken; 1= Spoke Spanish at home). Other researchers have used self-report measures similar to the one used in the present study to assess language background of students from their own perspectives (e.g., De la Piedra, 2011; Guglielmi, 2008).
This variable was defined as perceived ability in and mastery of English. To assess this construct, the students responded to four self-evaluation items to indicate how well they understood, spoke, read, and wrote English (ranging from 1 = very well to 4 = not at all). Items were reverse coded so that higher scores were equivalent to higher reported proficiency in English. In the present sample, Cronbachs alpha coefficient of .92 was obtained for these items. Researchers have used similar self-report measures of language proficiency reliably (e.g., Guglielmi, 2008; Portes & Hao, 2002).
Social Justice Orientation
SJO is defined as the motivation to end social inequities/unfairness and help others in ones community. SJO involves student awareness of societal inequalities and the desire to change them. In the present study, four items taken from the NELS:88 first follow-up student questionnaire (10th grade) were used as indicators of the construct SJO. Two of the items, It is o.k. to make racist comments, and It is o.k. to make sexist comments, were assessed on a 4-point frequency scale (1 = often; 2 = sometimes; 3 = rarely; 4 = never). The two other items, important to help others in the community, and important to work to end inequalities, used a 3-point rating scale (1 = not important; 2 = somewhat important; 3 = very important). Higher levels of item endorsement represent higher levels of SJO. Cronbachs alpha coefficient for these items in this sample was .55. Diemer (2009) used the same items to assess SJO/sociopolitical development in a sample of students of Color taken from the NELS:88.
Personal Agency Beliefs
Personal agency is defined as a persons beliefs in her/his ability to accomplish and have control over what the person desires or plans (Bandura, 2001). To assess personal agency beliefs, four items were taken from the NELS:88 students survey at the 1st follow-up (i.e., 10th grade). Students responded to these items using a 4-point, Likert-type scale (ranging from 1 = strongly agree to 4 = strongly disagree) to indicate the extent to which the statements described them. For three of the items, higher scores indicated a greater sense of personal agency (e.g., When I make plans, I am almost certain I can make them work); one of the four items (i.e., Every time I try to get ahead, something or somebody stops me) was reverse scored so that its direction was consistent with the other three items. Similar items have been used reliably in previous studies to assess general personal efficacy and agency in Latinas/os (e.g., Locke, Newcomb, Duclos, & Goodyear, 2007; Weisman de Mamani, Rosales, & Navarro, 2007). In the current study, Cronbachs alpha coefficient of .69 was obtained.
This variable was defined as student involvement in community service. A single indicator which was a self-report item taken from the NELS:88 second follow-up (12th grade) student survey was used to assess community engagement. Specifically, the item assessed students weekly frequency of involvement in community service. Students responded to the item using a 4-point frequency scale (ranging from 1 = never/rarely to 4 = every day/almost every day). Researchers have used similar assessments of youths community engagement by observing frequency of involvement in community organizations (e.g., Cammarota, 2004; Ginwright & Cammarota, 2007).
School Behavioral Disengagement
This variable was defined as the presence of problem behaviors or consequences that are detrimental to a students academic performance. Six items, taken from the NELS:88 second follow-up student questionnaire (12th grade), were used to assess school behavioral disengagement. Items assessed the students self-reported frequency of school absences/truancy (e.g., I cut or skipped classes) and disciplinary actions (e.g., I was put on an in-school suspension) during the last school semester/term attended by the student. Students responded to items using a 7-point frequency scale (ranging from 0 = never to 6 = over 15 times). Higher scores on the items indicate higher levels of disengagement. Glanville and Wildhagen (2007) found good interitem reliability when using the same NELS:88 items to assess school (dis)engagement in a large sample, as used in the current study. Cronbachs alpha coefficient of .73 was obtained in the current sample.
School dropout was defined as students leaving/opting out of their school. In the present study, this construct was operationalized using two indicators from the NELS:88. The student academic transcripts obtained from their school indicated whether the student had ever dropped out of school. The second indicator was obtained based on whether a dropout questionnaire was administered to the student during the NELS:88 second follow-up data collection (i.e., 12th grade). A dichotomous variable was created to indicate whether or not the student was administered the dropout questionnaire. Researchers have used similar measures of school dropout (e.g., Carpenter & Ramirez, 2007; Stearns & Glennie, 2006). Cronbachs alpha for the school dropout items in the present sample was .94.
This variable was operationally defined as the grade point average (GPA) provided by the NELS:88 second follow-up school transcripts in the subjects of English, math, social studies, and sciences. For consistency purposes, all grades obtained from the schools were recoded by the NCES researchers to a 1 to 13 scale (where 1 = A+ and 13 = F) (Ingels et al., 1995). In the present study, GPAs in each subject were reverse coded so that higher grade points would indicate better grades (e.g., 1 = F and 13 = A).
Participants were selected from the NELS:88 dataset if they met the criteria of the present study: (a) students self-identified as Latina/o; (b) students data were available for the following waves of data collection (using NELS:88 terminology): base year (1988, 8th grade), first follow-up (1990, 10th grade), and second follow-up (1992, 12th grade); and (c) data from school administrators were available for the base year of data collection. Also, cases were selected if they had a school transcript from which grades and dropout status could be obtained. Another inclusion criterion was the availability of the base-year parent questionnaire that was used to assess student characteristics (e.g., place of birth of student and home language background).
Data were screened, before and after estimating missing data, to ensure that the assumptions for multivariate analyses were met. There were 3.2% missing values, which falls in the low range for missing values in moderate to large datasets (Tabachnick & Fidell, 2013). Therefore, missing data were imputed using the expectation maximization (EM) algorithm in the PRELIS program (Jöreskog & Sörbom, 2009). No multivariate outliers were identified using Mahalanobis and Cooks distance. There was no evidence of multicollinearity or singularity. Bivariate correlations among the observed variables (i.e., indicators) also ruled out singularity (see Table 2). Preliminary analyses indicated that there were no differences among Latinas/os of different ethnicities and/or generational statuses, so they were included in the study as a single group.
To facilitate the analyses in the LISREL statistical package, all variables were rescaled to a common metric, the percent of maximum scale (POMS) (Little, 2013). POMS rescaling does not change the shape of the variables distributions, but rescales the variables scores so that they are all on a common metric, ranging from 0 to 1 (Little, 2013). In addition, some of the constructs indicators were parceled (i.e., grouped) to facilitate the main analysis (Little, Cunningham, Shahar, & Widaman, 2002). In structural equation modeling (SEM), it is generally recommended that latent constructs have between three to four indicators for stability of the results in the latent space. Parceling (i.e., grouping) the indicators has been recommended to facilitate analysis in constructs with more than five indicators (Little et al., 2002). To determine how to parcel the indicators, exploratory maximum likelihood factor analyses were conducted with the indicators for the latent constructs that had more than four indicators (i.e., school relational climate, school language climate, and school behavior disengagement). Parcels were created by using the mean of the scores of the items that loaded on the same factor.
Preliminary analyses indicated that age, SES, and gender were associated with several indicators of the latent outcome variables. Therefore, they were included as covariates in the model. In addition, analysis of a two-group model was conducted to observe whether the associations among the constructs in the model occurred differently for girls and boys. Girls and boys in this sample differed with respect to their age (i.e., boys tended to be older than girls).
A confirmatory factor analysis was conducted to assess the fit of the measurement model for all the latent constructs and their indicators. Also, the fit of the data to the structural model (with hypothesized paths among the constructs) was evaluated. All analyses were conducted using the LISREL 8.72 statistical package. In addition, direct paths from the variables age, SES, and baseline grades to the outcome variables in Time 3 were added to the model because of their potential role as nuisance variables.
The model was estimated from the raw data in PRELIS using the covariance matrix and maximum likelihood estimation. A total of 29 indicators were used for 13 latent constructs, including the covariates (i.e., age, SES, and baseline grades, all of which had single indicators). The minimum fit function was χ²(308, N = 1,472) = 1100.00, p < .001. The Comparative Fit Index (CFI) was .97 and Root Mean Squared Error of Approximation (RMSEA) was .04, which suggest very good measurement model fit (Little, 2013). Of note, the chi-square statistic is reported here by convention; however, it is usually significant in large samples (Tabachnick & Fidell, 2013). The CFI and RMSEA are less susceptible to sample size. In Table 2, the association of each indicator with its construct is presented, including the LISREL parameter estimates (interpretable as factor loadings), their statistical significance, and standard errors. Also, the indicators R² (or squared multiple correlation), which represents the proportion of variance in the indicators explained by their latent construct, are listed. To decrease measurement bias in each construct with two indicators, factor loadings for each indicator were constrained to be equal to attain balance in measurement (Little, Lindenberger, & Nesselroade, 1999).
Table 2. Factor Loadings, Standard Errors, and R² of Indicators (Observed Variables) in the SJOLY Model (Test of Measurement Model)
Note. ***p < .001; l= Lambda coefficient (factor loading for indicator); R² = Determination coefficient (proportion of variance in the indicator explained by its latent construct). Constructs with single indicators (i.e., SPANISH, COMS, Age, SES, and Baseline grades) have R²s = 1 because, as single indicators, all the variance is explained by their latent construct.
As shown in Table 2, the determination coefficients or percent of variance explained (R²) and factor loadings of most of the indicators on their latent constructs were in the moderate to large range (i.e., R² ranging from .40 to .94). However, two of the indicators of SJO and one indicator of agency presented small factor loadings and small proportion of variance explained by their latent construct (i.e., SJO3 R² = .12; SJO4 R² = .06; and Agency4 = .09). Because of the theoretical centrality of these indicators to SJO (i.e., importance to work to end societal inequalities; importance to work to help in the community) and agency (i.e., when the student makes plans is certain that they will work), they were kept in the model for the subsequent analyses. Overall, the measurement model had a very good fit to the data; therefore, the structural model (associations among the latent constructs) was evaluated.
The model was estimated using the covariance matrix and the maximum likelihood method. The fit statistics revealed an adequately fitting model, χ²(365, N = 1,472) = 2310.58, p < .001, CFI = .92, RMSEA = .06. Model modifications were performed to develop a better fitting model. Thus, all the initially estimated paths (i.e., β and ϕ paths) among the constructs that were not significant were deleted for model parsimony. Wald statistics, which indicate the statistical significance of each parameter or path, were evaluated. Wald statistics smaller than 1.96 (critical value for p < .05) were deemed as not significant and the associated paths were deleted from the model. In addition, modification indices (i.e., Lagrange multiplier tests or LM tests) were inspected to determine if it was necessary to add paths among the constructs to improve model fit based on the data. Only paths with large modification indices that made sense theoretically and practically were added to the model. Also, based on previous research and theory that proposes an influence of Spanish proficiency and English proficiency in personal sense of agency (e.g., Buriel et al., 1998; Martínez et al., 2004), paths from Spanish background and English proficiency predicting agency were also added to the model.
Tests to assess the outcomes of such changes were also conducted (i.e., chi-square difference tests with less than 1 degree of freedom difference/change) to evaluate changes in the overall model fit after each modification was conducted (Tabachnick & Fidell, 2013). All changes were made sequentially (i.e., changing one relational path at a time) because each path modification might have affected the entire model fit and other path coefficients. A total of 13 models were estimated in response to the LISREL estimates, modification indices, and theory. The first six estimated modelsbased on deletion of nonsignificant pathsdid not produce a significant change in model fit. However, the last six models estimated yielded significant changes in the chi-square difference tests, which were also observed in the increase of the CFI by .03 points (from .92 to .95) and decrease of the RMSEA by .01 point (from .06 to .05). The final model yielded a very good fit to the data, χ²(364, N = 1,472) = 1624.48, p < .001, CFI = .95, and RMSEA = .05.
To check for the validity of the final estimates and to avoid Type I error (given the use of stepwise procedures), bivariate correlations among the parameters estimated (i.e., standardized solutions for each estimated path) in the original hypothesized model and the final model were conducted to ensure that the patterns of association among the latent variables were not altered by the model modification procedures. The correlations among the parameters (association paths) estimated in the hypothesized model (i.e., Model 1) and the parameters estimated in the final model (i.e., Model 13) were calculated. Results yielded a correlation coefficient of .97, indicating that the pattern of associations among the latent constructs was kept intact despite the stepwise model modifications.
Table 3. Significant Paths for Model Covariates
Note. ** = p < .01; *** = p < .001; β= Beta coefficient; ϕ = Psi Coefficient, which represents a correlation. Regression paths (β) are predicted directional associations among the variables (i.e., variable on the left having a direct influence on the variable on the right). Correlation paths (ϕ) are predicted associations among the variables.
All significant paths (p < .01 and p < .001) of the modified structural model are presented in Figure 2. A positive relational climate in 8th grade predicted higher SJO in 10th grade. School language climate in 8th grade was associated with language proficiency/background in 10th grade; both were associated with greater sense of personal agency. SJO in 10th grade predicted less school behavioral disengagement in the 12th grade (ß = -.27, p < .001), indicating that students who were more motivated for social justice also presented lower levels of disengagement (i.e., truancy and disciplinary problems) at their school. This direct path between SJO and school behavioral disengagement was not initially hypothesized; however, a modification index suggested that this regression coefficient was significant and its estimation improved the fit of the model. All predictors explained variance in the outcomes in 12th grade after controlling for age, SES, and baseline grades. Baseline grades (8th grade) were associated with academic achievement outcomes (i.e., grades and school dropout) in the 12th grade. In addition, baseline grades predicted personal agency beliefs in the 10th grade (ß = .29, p < .001), which was also predicted by SJO (ß = .10, p < .01). Overall, over a 4-year span, the predictors in the model explained 14% of the variance observed in community engagement, 16% of the variance observed in school behavioral disengagement, 10% of the variance observed in school dropout, and 34% of the variance observed in grades.
Figure 2. Final SJOLY
The sample was taken from the National Education Longitudinal Study of 1988 (NELS:88), a large-scale longitudinal study conducted by the National Center for Education Statistics (NCES) using a nationally representative sample of students who were followed from the 8th grade until adulthood in four data collection waves (Curtin, Ingels, Wu, & Heuer, 2002). Model with structural paths (LISREL standardized solution) is shown above; χ²(363, N = 1,472) = 1611.34, p. < .001; CFI= .95; RMSEA= .05. Covariate paths are presented in Table 3.
Indirect effects were observed among some of the latent constructs in the model. For example, SJO had significant indirect effects on all the outcome variables in the model via its association with personal agency beliefs. Specifically, 10th grade SJO had a small indirect effect on 12th grade community engagement (ß for indirect effect = .01, p < .05), 12th grade school behavioral disengagement (ß for indirect effect = .02, p < .05), 12th grade school dropout (ß for indirect effect = .01, p < .05), and 12th grade grades (ß for indirect effect = .02, p < .05). Of note, the effect of SJO on school behavioral disengagement was only partially mediated by personal agency. The modification indices provided by the LISREL program indicated the need to add a direct association path between SJO and school behavioral disengagement to improve model fit (ß = -.27, p < .001). In addition, 8th grade school relational climate had indirect effects on the models academic outcomes via SJO and personal agency beliefs. Specifically, 8th grade school relational climate had a small indirect effect on 10th personal agency beliefs via its association with 10th grade SJO (ß for indirect effect = .02, p < .05). In addition, 8th grade school relational climate had small indirect effects on 12th grade school behavioral disengagement (ß for indirect effect = -.06, p < .01), 12th grade school dropout (ß for indirect effect = -.002, p < .05), and 12th grade grades (ß for indirect effect = .01, p < .05). On the other hand, school language climate did not have significant indirect effects on any of the outcome variables. With respect to the language variables, 10th grade Spanish background did not have significant indirect effects in the 12th grade outcomes variables, whereas 10th grade English proficiency had significant indirect associations with 12th grade community engagement (ß for indirect effect = .01, p < .05), school behavior disengagement (ß for indirect effect = -.03, p < .01), school dropout (ß for indirect effect = -.02, p < .01) and grades (ß for indirect effect = .03, p < .01).
Prior to making comparisons between girls and boys, the SJOLY measurement model was assessed to ensure that the constructs in the model were equivalent in the two groups. Specifically, it was important to determine that the indicators assessed the same constructs for girls and boys and that measurement equivalence across groups could be assumed. Measurement model equivalence constraints across the groups were tenable (i.e., configural invariance; factorial invariance; and strong factorial invariance). Therefore, measurement equivalence could be assumed, which allowed further comparisons about the associations of the constructs between the groups.
When assessing construct level mean differences between the groups, we found that the constraint equivalence test of the means was tenable when nesting means equivalence across groups with the strong invariance model (see Table 4). Therefore, it can be concluded that there were not significant differences between girls and boys in this sample with respect to the construct levels. In other words, mean equivalence in the latent constructs across the groups could be assumed. Subsequently, an omnibus test of variance-covariance homogeneity across the groups was also conducted to assess whether a multigroup model instead of a single group model would provide further information about the way the constructs were (or were not) associated differently for girls and boys. As shown in Table 4, constraining the variance and covariance to be equal between the groups was tenable; therefore, it was determined that a single group SJOLY model, rather than a multigroup model, was most appropriate for this sample.
Table 4. Model Fit Statistics for Gender Comparisons of the SJOLY Model
Model compared to strong factorial invariance model.
(b) Model compared to the means equivalent model with SJO, disengagement, grades, age, and baseline grades unconstrained.
Note: All models are nested in the previous one.
The purpose of the present study was to develop and test a longitudinal model of early school environment predictors of long-term academic/civic outcomes of Latina/o youths based on personal and cultural strengths. Specifically, a model incorporating language skills, social justice orientation, and personal agency beliefs for Latina/o youths (SJOLY) was developed. The theory of sociopolitical development in youths of Color (Watts et al., 1999), which explains the process by which youths develop critical consciousness about systems of oppression and motivation for social justice, was used as a framework for the SJOLY model. Results indicated an overall good fit of the model to the data. All the hypotheses, with the exception of English and Spanish language skills predicting SJO, were supported.
SOCIAL JUSTICE ORIENTATION, SCHOOL CLIMATE, AND ACADEMIC OUTCOMES
Results indicated that higher levels of SJO predicted higher levels of agency, which in turn were also associated with greater levels of community engagement and higher grades, lower levels of school disengagement, and decreased likelihood of dropping out of school. With respect to the association between SJO and agency, it is noted that this finding is congruent with previous qualitative and quantitative research. For example, Cammarota (2004) found that Latina/o youths who were motivated to end societal inequities in their communities displayed greater sense of agency and empowerment. In addition, quantitative research using a sample of low-income students of Color taken from the NELS:88 provided similar findings with respect to the association of SJO and personal agency (locus of control) obtained in the present study (Diemer, 2009).
It was also determined that community engagement in 12th grade was significantly predicted by personal agency in 10th grade, but its effect size (β = .10) was small. It is worth noting that this sample of students displayed low frequencies of community service. The mean of community service before the variable’s square root transformation was 1.32 (SD = .69) with a median and mode of 1 (possible range of responses for this variable was 1 = never/rarely to 4 = every day/almost every day). However, the self-reported statistics on community engagement found in this study were congruent with national levels of youths community engagement and participation, especially for some youths of Color who may have more limited time for or access to community engagement opportunities (Newberg, 2010). It is important to also note that this is a self-reported measure, which poses measurement limitations. Future studies may more adequately assess this construct by also including behavioral and other non-self-report measures.
Additionally, results of the present study indicated that greater SJO via its association with personal agency in 10th grade were associated with decreased school disengagement in 12th grade. This finding is congruent with theory and previous research findings, whereby self-efficacy beliefs were associated with greater school engagement (i.e., decreased absences) among students of Color, including Latinas/os (Close & Solberg, 2008). The model modification indices in LISREL also suggested the inclusion of a direct path between SJO and school behavioral disengagement. This is an important finding, as greater levels of SJO in 10th grade predicted lower levels of school behavior disengagement in 12th grade (i.e, truancy, disciplinary problems) after controlling for age, SES, and baseline grades. This finding provides empirical support for potential long-term positive effects that SJO may have on the educational outcomes of Latina/o students. The SJOschool engagement relationship is congruent with previous research in which SJO and critical consciousness translated into greater school engagement, especially when education was seen as a way to end oppression (Bingham & Okagaki, 2012; Cammarota, 2004; Solórzano & Bernal, 2001). The direct path between SJO and school disengagement also suggests that other mechanisms linking these constructs should be explored in future studies.
In addition, the finding that a more positive relational climate in middle school was associated with greater levels of students SJO and more positive academic outcomes during their high school years was congruent with previous research. Specifically, previous research findings have highlighted the importance of an early positive relational climate for fostering long-term positive personal, emotional, and academic outcomes among Latinas/os (Conchas, 2001; Rubin, 2007). School relational climate is an important school-related variable to continue address in studies and interventions for Latinas/os. On the other hand, school relational climate should not be studied in isolation, as other variables pertinent to Latina/o students development may also correlate to school relational climate, such as school language climate and school racial climate (Jayakumar, 2008; Jernigan & Daniel, 2011; Mattison & Aber, 2007; Rubin, 2007).
SCHOOL LANGUAGE CLIMATE AND LANGUAGE
School language climate was positively associated with Spanish language background, but negatively associated with English proficiency. That is, more school language resources/instruction and more diverse student-body composition with respect to language (i.e., school language climate) in 8th grade predicted more Spanish background in 10th grade and lower self-reported English proficiency in 10th grade. The negative association between language instructional resources and English language learner (ELL) student-body composition in 8th grade (i.e., school language climate) and English self-reported proficiency in 10th grade might be interpreted as indicating that more language resources are associated with less English proficiency. However, in the present study, language instructional resources in schools were evaluated with student bodies that were mostly proficient in English. Therefore, not having many language instructional sources might make sense. However, without having a baseline for the students actual English proficiency, it is difficult to assess the effects of language instructional resources on English proficiency. The obtained finding with respect to school language resources and Spanish language background is congruent with previous research findings that indicate a positive association between school language instructional resources and Spanish proficiency (e.g., Guglielmi, 2008; Lindholm-Leary & Borsato, 2005; Quintanar-Sarellana, 2004).
Most of the schools included in the present study did not have many language resources, which restricted the variability of this construct. Specifically, there was little variability in the composition of the schools with respect to the percentages of ELL or English as a second language (ESL) students and related language resources. For example, 80% of the students attended schools that had 7% or fewer students enrolled in ESL programs.
Additionally, the NELS:88 student survey was only available in English, and all of the students were selected to participate in the study if they were deemed by a school counselor/staff to be proficient enough in English to answer the survey questions. Therefore, lack of language variability likely affected the results of the present study and suggests that they may not be generalizable to Latinas/os who are not proficient in English.
Overall, the SJOLY model had a good fit with the data and was informative in assessing the longitudinal impact of school cultural climate variables, SJO, language proficiencies, and personal agency in long-term secondary civic and educational outcomes. The SJOLY model provided evidence of cross-sectional and longitudinal associations among the studied variables in a national sample of Latina/o students.
Study Design Limitations
A longitudinal model of the association among variables thought to be related to positive outcomes for Latina/o students was tested. Because some paths were cross-sectional, however, caution should be used in interpreting the direction of these relationships or attributing causation. For example, although the predictive paths between SJO and agency are rooted in theory and previous empirical evidence, given the cross-sectional nature of the data used for this path in the current study, it is not possible to draw definite conclusions about their relationship. Watts and Flanagan (2007) suggest that there could be a bidirectional association between SJO and personal agency. That is, greater SJO could lead to greater sense of personal agency; moreover, greater personal agency might also lead to greater SJO. With the exception of baseline grades, other variables were not observed at Time 1 of the study. Controlling for baseline levels of the constructs or a longitudinal panel design would have provided a more conservative approach to estimating changes in the outcome variables.
Limitations Associated with Measurement and NELS:88 Dataset
Although the use of national longitudinal public datasets allows us to further research complex phenomena, measurement limitations should be considered. It is important to note that the variables assessed in this study were constrained by the way they were measured and conceptualized in the NELS:88. For instance, school disengagement and school dropout may emphasize deficit perspectives of students without taking into account contextual variables associated with students performances. Calling these variables school engagement and/or educational attainment, more congruent with a strength-based perspective, would have not been an accurate reflection of the way these constructs were measured in the NELS:88.
In addition, the variable personal agency beliefs was operationalized by using items that emphasized personal choice and control. Yet this operational definition of personal agency may not account for agency in the context of oppressive environments that many Latina/o youths encounter in their schools. Further, the operational definition of personal agency in the present study emphasized individualistic and Western notions of self-orientation, which may not be applicable to some Latinas/os, given that they are generally described as collectivistic rather than individualistic (La Roche, D'Angelo, Gualdrón, & Leavell, 2006; Triandis, 1994). Therefore, future studies should further investigate the role of agency in the educational experiences of Latina/o students. Moreover, other forms of agency that might be more congruent with collectivistic values (e.g., collective agency) should be researched.
Other important constructs assessed in the present study were school language climate and students language skills in Spanish and English. However, these constructs were poorly conceived in the NELS:88 dataset. For example, because the structure of the NELS:88 student survey advised students to skip items under some circumstances, Spanish self-reported proficiency could not be used as a measure of Spanish skills in the present study. Specifically, half of the sample (50.2%) skipped the Spanish proficiency items in the survey, although some of them may have understood or spoken some Spanish given that 80% of the students in the sample reported that Spanish was spoken in their households and their caregivers had reported that the students spoke Spanish. Therefore, Spanish language background was based on the students indication of whether or not Spanish was spoken at her/his household in the present study. This measure was problematic, as it was not possible to directly assess the students Spanish language skills.
Thus, although the associations between the language variables under question were statistically significant and had moderate effect sizes (.37 and -.42, respectively, for Spanish background and English proficiency as predicted by school language climate), it is problematic to interpret and draw conclusions from the associations between constructs that were poorly measured. Future studies should use more adequate measures of school language climate that also include assessment of students experiences. For example, are school administrators and teachers promoting or limiting development of student language skills? Also, language skills could be assessed objectively (e.g., proficiency tests and assessments) and subjectively (e.g., self-report), to more adequately estimate students skills in different languages as they relate to their subjective assessment of them and other psychological variables.
It is important to note that the measurement of SJO was also limited to items available in the NELS:88 dataset. Additionally, the items measuring SJO had low internal consistency, which may have attenuated this constructs relationship with other variables in the model. As noted, SJO is a complex construct which was assessed with single indicators in the present study. Although similar indicators/items have been used successfully by researchers in previous studies with different samples (e.g., Diemer, 2009; Diemer, 2012; Diemer et al., 2009), a multidimensional/multimethod assessment of this construct would be beneficial to further study. Further, the present studys findings and associations of SJO with school climate and academic outcomes variables supports further study and measurement development of this construct.
In addition, although a large sample was used to conduct a longitudinal study, the sample may be dated, as the data were collected in 1988, 1990, and 1992. Although it was possible to assess the ways in which the variables were associated in the present sample, important contextual or cohort factors may have affected the studied youths and current students in the U.S. in different ways. Nonetheless, this study provides a framework and a model for assessing SJO and educational outcomes with future samples and datasets.
STUDY STRENGTHS AND IMPLICATIONS
The strengths of this study include the use of a longitudinal design, which allowed us to test associations among variables over time while controlling for nuisance variables. In addition, the use of SEM techniques allowed testing of hypotheses and relationships simultaneously, including both direct and indirect effects.
The present study provides evidence for the importance of continuing to develop and investigate strength-based variables associated with Latina/o students educational/civic outcomes. It can be argued that although the theory of sociopolitical development was initially developed and applied to African American boys (Watts et al., 1999), the principles that link SJO with school and civic engagement are also relevant for Latina/o youths. However, in this study, this theory may be further developed for Latinas/os, as some specific Latina/o cultural factors were also identified by testing the SJOLY model.
The present studys results also provide evidence of the relevance of early school climate experiences, systemic factors, and their potential indirect longitudinal effects on the academic achievement of Latina/o youths. Specific student-level mechanisms in association with school-related variables were tested as they are relevant for the academic achievement of Latina/o youths. Thus, the SJOLY model may inform further research as well as preventive and educational practices to enhance the academic experiences and outcomes of Latina/o youths when capitalizing on their personal and cultural strengths. For instance, interventions targeting the educational experiences of Latina/o youths may incorporate SJO as a way to support school engagement and positive outcomes. Further research could evaluate the feasibility and effectiveness of such interventions. Strength-based approaches that capitalize on Latina/o cultural strengths would be highly beneficial.
Future studies should replicate the SJOLY model with other samples. Furthermore, researchers should continue to add culturally congruent constructs for Latina/o students to the SJOLY model (e.g., biculturalism, racial identity, collective agency). In addition, although analyses suggested that there were not differences among Latinas/os from different generational, ethnic, racial backgrounds in this sample, future studies should continue to explore these associations given prior research and the impact of context in Latinas/os experiences (e.g., C. Suárez-Orozco & Suárez-Orozco, 2001). Potential gender differences should also continue to be explored in future studies (Cammarota, 2004).
Based on the findings of the present study, it can be argued that SJO may function as a personal strength and protective factor for Latina/o youths in oppressive/unequal educational systems. If this is so, educators and counselors might be able to support Latina/o students academic achievement by developing programs of prevention and interventions that foster the development of their students SJO and personal agency. Specifically, it was found that SJO and agency were positively associated with academic outcomes. Hence, counseling interventions that focus on social justice in the face of the sociopolitical inequities that Latina/o youths face could be potentially empowering in educational contexts. In addition, counseling interventions intended to target school engagement/disengagement may emphasize the exploration (and the acknowledgment) of the students experiences in their social/academic contexts as embedded in the sociopolitical inequities that many students of Color face.
Also, the active support and recognition of the students roles in their community as agents of change may be key for instilling a sense of agency and hope. According to the SJOLY model in the current study, such recognition could function as a powerful resource for the students personal, community, and educational benefit. Counseling interventions targeting SJO may benefit the students immediate sense of agency, which may also translate in more positive academic outcomes. In addition, long-term civic and academic outcomes may be observed from interventions that emphasize the development of an SJO. Because of the potential personal and educational benefit of civic engagement observed in the present study and previous research (e.g., Cammarota, 2004; Morsillo & Prilleltensky, 2007), it is also important that counselors and educators work in collaboration with community agencies and promote civic engagement opportunities for empowerment.
The SJOLY model may also inform educational practices to benefit Latina/o students. Specifically, the SJOLY model may have concrete implications for school leadership and teaching practices to enhance Latina/o educational experiences and well-being. With respect to school leadership, interventions and practices aimed at enhancing school climate are of particular relevance. Specifically, the SJOLY model provided longitudinal evidence of the importance of cultural factors associated with school climate that benefit Latina/o students personally and academically. School principals, administrators, and counselors should pay close attention to their school environment to foster positive relationships and collaboration among teachers, students, and staff for the benefit of their students. In the present study, it was observed that a positive relational climate in which teachers cared and listened to the students and discipline was fair was associated with greater levels of SJO. In turn, SJO was positively associated with personal agency beliefs, grades, engagement, and lower likelihood of dropping out of school. In the present study, it was clear that environments in which students experienced more positive and caring relationships translated into more positive personal and academic outcomes. Early positive relational climate was beneficial for the students during their high school years and was indirectly associated with high school completion.
With respect to the teaching implications of the SJOLY model, it is important to point out that curricula that emphasize or promote social justice matters may benefit Latinas/os school engagement and their personal agency. That is, teaching that emphasizes critical consciousness and/or provides social justice content may be of particular benefit for the educational engagement of Latina/o students. More specifically, social justice matters that are pertinent to cultural and race-related experiences of students should be purposely integrated. For example, studies have found that when Latina/o students engaged in ethnic/Chicana/o studies courses, students increased their SJO and empowerment, engaged in transformational resistance, and were significantly more engaged in their education (Cammarota, 2004).
Targeting SJO through education may be accomplished not only through social justice-centered curricula, but also through teachers who serve as role models and embrace SJO issues in their classrooms. This has been exemplified by teachers working with Latina/o students who have historically made a difference for their students empowerment and education. This kind of teaching that fosters SJO and empowerment has promising results, not only for Latina/o students, but also for U.S. society. Therefore, implications for teacher education programs that train new generations of teachers for social justice should also be considered. It is important to note that teachers background and experiences (e.g., racial identity schemas) may inform the way that they conceptualize social justice issues and make sense of the sociopolitical realities that impact their students. Hence, it is important to note that teachers awareness, sensitivity, and critical consciousness must be taken into consideration when designing training programs that view teachers and students as agents of change. Future studies may inform the specifics of such training programs. However, based on the SJOLY model, teachers (as central figures of school relational climate) should have a key role in developing or supporting their students SJO and personal agency.
Alfaro, E. C., Umaña-Taylor, A. J., Gonzales-Backen, M. A., Bámaca, M. Y., & Zeiders, K. H. (2009). Latino adolescents academic success: The role of discrimination, academic motivation, and gender. Journal of Adolescence, 32, 941962. doi:10.1016/j.adolescence.2008.08.007
Bandura, A. (2001). Social cognitive theory: An agentic perspective. Annual Review of Psychology, 52, 126. doi:10.1146/annurev.psych.52.1.1
Bingham, G. E., & Okagaki, L. (2012). Ethnicity and student engagement. In S. Christenson, A. Reschly, & C. Wylie (Eds.), Handbook of research on student engagement (pp. 6595). New York, NY: Springer.
Bronfenbrenner, U. (1994). Ecological models of human development. In M. Gauvain & M. Cole (Eds.), Readings on the development of children (2nd ed, pp. 3743). New York, NY: Freeman.
Buriel, R., Perez, W., DeMent, T. L., Chavez, D. V., & Moran, V. R. (1998). The relationship of language brokering to academic performance, biculturalism, and self-efficacy among Latino adolescents. Hispanic Journal of Behavioral Sciences, 20, 283297. doi:10.1177/07399863980203001
Cammarota, J. (2004). The gendered and racialized pathways of Latina and Latino youth: Different struggles, different resistances in the urban context. Anthropology & Education Quarterly, 35, 5374. doi:10.1525/aeq.2004.35.1.53
Cammarota, J. (2007). A social justice approach to achievement: Guiding Latina/o students toward educational attainment with a challenging, socially relevant curriculum. Equity & Excellence in Education, 40(1), 87-96.
Carpenter, D. M., & Ramirez, A. (2007). More than one gap: Dropout rate gaps between and among Black, Hispanic, and White students. Journal of Advanced Academics, 19, 3264.
Christenson, S., Reschly, A., & Wylie, C. (Eds.). (2012). Handbook of research on student engagement. New York, NY: Springer.
Close, W., & Solberg, S. (2008). Predicting achievement, distress, and retention among lower-income Latino youth. Journal of Vocational Behavior, 72, 3142. doi:10.1016/j.jvb.2007.08.007
Cohen, J. (2014, February). School climate policy and practice trends: A paradox. Teachers College Record. Retrieved from https://www.tcrecord.org/Content.asp?ContentID=17445.
Cohen, J., Pickeral, T., & McCloskey, M. (2009). Assessing school climate. Education Digest, 74, 4548.
Conchas, G. Q. (2001). Structuring failure and success: Understanding the variability in Latino school engagement. Harvard Educational Review, 71, 475504.
Curtin, T. R., Ingels, S. J., Wu, S., & Heuer, R. (2002, July). National Education Longitudinal Study of 1988: Base year to fourth follow-up data file users manual (NCES 2002-323): Washington, DC: National Center for Education Statistics. Retrieved from http://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=2002323
De la Piedra, M. T. (2011). Tanto necesitamos de aquí como necesitamos de alla: Leer juntas among Mexican trasnational mothers and daughters. Language and Education, 25, 6578.
Diemer, M. A. (2009). Pathways to occupational attainment among poor youth of color: The role of sociopolitical development. Counseling Psychologist, 37, 635. doi:10.1177/0011000007309858
Diemer, M. A. (2012). Fostering marginalized youths political participation: Longitudinal roles of parental political socialization and youth sociopolitical development. American Journal of Community Psychology, 50, 246256. doi:10.1007/s10464-012-9495-9
Diemer, M. A., Hsieh, C., & Pan, T. (2009). School and parental influences on sociopolitical development among poor adolescents of color. Counseling Psychologist, 37, 317344. doi:10.1177/0011000008315971
Dolan, S. L. (2009). Missing out: Latino students in Americas schools. Retrieved from http://www.nclr.org/content/publications/detail/57016/
Fine, M. (1986). Why urban adolescents drop into and out of public high school. Teachers College Record, 87, 393409.
Fine, M., Burns, A., Payne, Y. A., & Torre, M. E. (2004). Civics lessons: The color and class of betrayal. Teachers College Record, 106, 21932223. doi:10.1111/j.1467-9620.2004.00433.x
Finn, J., D., & Rock, D. A. (1997). Academic success among students at risk of school failure. Journal of Applied Psychology, 82, 221234. doi:0021-9010/97/s3.00
Fredricks, J. A., Blumenfeld, P. C., & Paris, A. H. (2004). School engagement: Potential of the concept, state of the evidence. Review of Educational Research, 74, 59109. doi:10.3102/00346543074001059
Freidlin, B., & Salvucci, S. (1995, March). Empirical evaluation of social, psychological, and educational construct variables used in NCES surveys (Working Paper No. 95-14). Washington, DC: National Center for Education Statistics. Retrieved from http://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=9514
Ginwright, S., & Cammarota, J. (2007). Youth activism in the urban community: Learning critical civic praxis within community organizations. International Journal of Qualitative Studies in Education, 20, 693710. doi:10.1080/09518390701630833
Glanville, J. L., & Wildhagen, T. (2007). The measurement of school engagement: Assessing dimensionality and measurement invariance across race and ethnicity. Educational and Psychological Measurement, 67, 10191041. doi:10.1177/0013164406299126
Gonzalez, R. (2009). Beyond affirmation: How the school context facilitates racial/ethnic identity among Mexican American adolescents. Hispanic Journal of the Behavioral Sciences, 31, 531. doi:10.1177/0739986308328387
Gregoire, M., & Algina, J. (2000, April). Reconceptualizing the debate on school climate and students academic motivation and achievement: A multilevel analysis. Paper presented at the meeting of the American Educational Research Association, New Orleans, LA.
Guglielmi, R. S. (2008). Native language proficiency, English literacy, academic achievement, and occupational attainment in limited-English-proficient students: A latent growth modeling perspective. Journal of Educational Psychology, 100, 322342. doi:10.1037/0022-06126.96.36.1992
Helms, J. E. (2003). Racial identity and racial socialization as aspects of adolescents identity development. In R. Lerner, F. Jacobs, & D. Wertlief (Eds.), Handbook of applied developmental science: Promoting positive child, adolescent, and family development through research, policies, and programs (Vol. 1, pp. 143163). Thousand Oaks, CA: Sage.
Helms, J., & Cook, D. (1999). Using race and culture in counseling and psychotherapy: Theory and process. Needham Heights, MA: Allyn & Bacon.
Ingels, S. J., Dowd, K. L., Baldridge, J. D., Stipe, J. L., Bartot, V. H., & Frankel, M. R. (1994, September). National Education Longitudinal Study of 1988: Second follow-up: Student component data file users manual (NCES 94-374). Washington, DC: National Center for Education Statistics. Retrieved from http://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=94374
Ingels, S. J., Dowd, K. L., Taylor, J. R., Bartot, V. H., Frankel, M. R., & Pullian, P. A. (1995, March). National Education Longitudinal Study of 1988: Second follow-up: Transcript component data file users manual (NCES 95-377). Washington, DC: National Center for Education Statistics. Retrieved from http://nces.ed.gov/pubsearch/pubsinfo.asp?pubid=95377
Jayakumar, U. M. (2008). Can higher education meet the needs of an increasingly diverse and global society? Campus diversity and cross-cultural workforce competencies. Harvard Educational Review, 78, 615651.
Jernigan, M. M., & Daniel, J. H. (2011). Racial trauma in the lives of Black children and adolescents: Challenges and clinical implications. Journal of Child and Adolescent Trauma, 4, 123141.
Jöreskog, K., & Sörbom, D. (2009). PRELIS 2: Users reference guide. Lincolnwood, IL: Scientific Software International.
La Roche, M., D'Angelo, G., Gualdrón, L. G., & Leavell, J. (2006). Culturally sensitive guided imagery for allocentric Latinos: A pilot study. Psychotherapy: Theory, Research, Practice, Training, 34, 555560.
La Roche, M. J., & Shriberg, D. (2004). High stakes exams and Latino students: Toward a culturally sensitive education for Latino children in the United States. Journal of Educational and Psychological Consultation, 15, 205223. doi:10.1207/s1532768xjepc1502_8
Lindholm-Leary, K., & Borsato, G. (2005). Hispanic high schoolers and mathematics: Follow-up of students who had participated in two-way bilingual elementary programs. Bilingual Research Journal, 29, 641652.
Little, T. D. (2013). Longitudinal structural equation modeling. New York, NY: Guilford Press.
Little, T. D., Cunningham, W. A., Shahar, G., & Widaman, K. F. (2002). To parcel or not to parcel: Exploring the question, weighting the merits. Structural Equation Modeling, 9, 151173. doi:10.1207/S15328007SEM0902_1
Little, T. D., Lindenberger, U., & Nesselroade, J. R. (1999). On selecting indicators for multivariate measurement and modeling with latent variables: When "good" indicators are bad and "bad" indicators are good. Psychological Methods, 4, 192211. doi:10.1037/1082-989X.4.2.192
Locke, T. F., Newcomb, M. D., Duclos, A., & Goodyear, R. K. (2007). Psychosocial predictors and correlates of dysphoria in adolescent and young Latinas. Journal of Community Psychology, 35, 135149. doi:10.1002/jcop.20139
Marín, G., & Marín, B. V. (1991). Research with Hispanic populations, Applied Social Research Methods (Vol. 23). Newbury Park, CA: Sage.
Martínez, C. R., Jr., DeGarmo, D. S., & Eddy, J. M. (2004). Promoting academic success among Latino youths. Hispanic Journal of Behavioral Sciences, 26, 128151. doi:10.1177/0739986304264573
Mattison, E., & Aber, M. S. (2007). Closing the achievement gap: The association of racial climate with achievement and behavioral outcomes. American Journal of Community Psychology, 40, 112. doi:10.1007/s10464-007-9128-x
Miller, J., & Garran, J. M. (2008). Racism in the United States: Implications for the helping professions. Belmont, CA: Thompson Brooks/Cole.
Morsillo, J., & Prilleltensky, I. (2007). Social action with youth: Interventions, evaluation, and psychopolitical validity. Journal of Community Psychology, 35, 725740. doi:10.1002/jcop.20175
National Center for Education Statistics (2015, May). The condition of education: Racial/ethnic enrollment in public schools. Retrieved from http://nces.ed.gov/programs/coe/indicator_cge.asp
National School Climate Council (2007). The school climate challenge: Narrowing the gap between school climate research and school climate policy, practice guidelines and teacher education policy. Retrieved from http://www.schoolclimate.org/climate/documents/policy/school-climate-challenge-web.pdf
Newberg, N. A. (2010). Rebirth: Civic engagement from adolescence to adulthood. In D. P. Swanson, M. C. Edwards, & M. B. Spencer (Eds.), Adolescence: Development during a global era (pp. 367387). Burlington, MA: Academic Press.
North, C. (2008). What is all this talk about social justice"? Mapping the terrain of education's latest catchphrase. Teachers College Record, 110, 11821206.
Oakes, J., Rogers, J., & Lipton, M. (2006). Learning power: Organizing for education and justice. New York, NY: Teachers College Press.
OConnor, C. (1997). Dispositions toward (collective) struggle and educational resilience in the inner city: A case analysis of six African-American high school students. American Educational Research Journal, 34, 593629. doi:10.2307/1163351
Portes, A., & Hao, L. (2002). The price of uniformity: Language, family and personality adjustment in the immigrant second generation. Ethnic and Racial Studies, 25, 889912. doi:10.1080/0141987022000009368
Quintanar-Sarellana, R. (2004). ¡Si se puede! Academic excellence and bilingual competency in a K8 two-way dual immersion program. Journal of Latinos and Education, 3, 87102. doi:10.1207/s1532771xjle0302_3
Ream, R. K., & Rumberger, R. W. (2008). Student engagement, peer social capital, and school dropout among Mexican American and non-Latino White students. Sociology of Education, 81, 109139. doi:10.1177/003804070808100201
Rubin, B. C. (2007). Theres still not justice: Youth civic identity development amid distinct school and community contexts. Teachers College Record, 109, 449481.
Rumberger, R. W., & Rotermund, S. (2012). The relationship between engagement and high school dropout. In S. Christenson, A. Reschly, & C. Wylie (Eds.). Handbook of research on student engagement (pp. 491513). New York, NY: Springer.
Schneider, S. H., & Duran, L. (2010). School climate in middle schools: A cultural perspective. Journal of Research in Character Education, 8, 2537.
Solórzano, D. G., & Bernal, D. D. (2001). Examining transformational resistance through a critical race and LatCrit theory framework: Chicana and Chicano Students in an urban context. Urban Education, 36, 308342. doi:10.1177/0042085901363002
Stearns, E., & Glennie, E. J. (2006). When and why dropouts leave high school. Youth and Society, 38, 2957. doi:10.1177/0044118X05282764
Suárez-Orozco, C., & Suárez-Orozco, M. M. (2001). Children of immigration. Cambridge, MA: Harvard University Press.
Suárez-Orozco, C., Suárez-Orozco, M. M., & Todorova, I. (2008). Learning a new land: Immigrant students in American society. Cambridge, MA: Harvard University Press.
Suárez-Orozco, M. M., & Páez, M. M. (Eds.). (2008). Latinos: Remaking America. Berkeley: University of California Press.
Tabachnick, B. G., & Fidell, L. S. (2013). Using multivariate statistics (6th ed.). Boston, MA: Allyn & Bacon.
Thapa, A. (2013, February). School climate research. Retrieved from https://www.schoolclimate.org/publications/documents/sc-brief-research.pdf
Thapa, A., Cohen, J., Guffey, S., & Higgins-DAlessandro, A. (2013). A review of school climate research. Review of Educational Research, 83, 357385. doi:10.2102/0034654313483907
Thomas, W. P., & Collier, V. P. (2002). A national study of school effectiveness for language minority students long-term academic achievement. Retrieved from http://www.usc.edu/dept/education/CMMR/CollierThomasExReport.pdf
Triandis, H. (1994). Theoretical and methodological approaches to the study of collectivism and individualism. In U. Kim, H. Triandis, C. Kagitcibasi, S. Choi, & G. Yoon (Eds.), Individualism and collectivism (pp. 4151). Thousand Oaks, CA: Sage.
Watts, R. J., & Flanagan, C. (2007). Pushing the envelope on youth civic engagement: A developmental and liberation psychology perspective. Journal of Community Psychology, 35, 779792. doi:10.1002/jcop.20178
Watts, R. J., Griffith, D. M., & Abdul-Adil, J. (1999). Sociopolitical development as an antidote for oppression: Theory and action. American Journal of Community Psychology, 27, 225271. doi:10.1023/A:1022839818873
Watts, R. J., Williams, N. C., & Jagers, R. J. (2003). Sociopolitical development. American Journal of Community Psychology, 31, 185194. doi:10.1023/A:1023091024140
Weisman de Mamani, A., Rosales, G. A., & Navarro, M. (2007). Personal efficacy, attitudes towards immigrants, and voting behavior among Latino and White university students. Revista Iberoamericana de Psicología/Iberoamerican Journal of Psychology, 41, 341348.