The Role of Parent Social Capital and College-aligned Actions in Explaining Differences in Intergenerational Resource Transfer Among Hispanic and White Youth on the Path to College


by Sarah Ryan - 2017

Background: A growing number of scholars argue that Hispanic–White disparities in college pathways are at least partly attributable to the fact that Hispanic students experience the college preparation and enrollment process differently from their White peers. Recent research suggests that one way in which the process differs between the two groups is in how parent resources, and particularly forms of parent social capital, facilitate college preparation and enrollment.

Purpose: This study examines whether group-level variability in the utility of parent social capital can help explain the recent finding that parent income and education confer greater benefits among White youth, relative to similar Hispanic youth, when it comes to 4-year college enrollment. The study also asks whether engagement in college-aligned actions might mediate the association between parent social capital and college enrollment.

Participants: The study focuses on a nationally representative sample of Hispanic and White 10th graders in 2002 who completed high school expecting to complete a bachelor’s degree.

Research Design: The study uses multiple group structural equation modeling techniques and three waves of data from the Education Longitudinal Study of 2002 (ELS:2002) to test whether forms of parent social capital indirectly influence 4-year college enrollment through an association with college-aligned actions. Multiple group invariance testing is used to test whether associations of interest are measurably different across the two groups.

Conclusions: The results suggest that Hispanic–White differences in the intergenerational transmission of parents’ educational and economic resources can be partly attributed to differences in the functioning of parent social capital. This finding is consistent with the assertion that social capital often serves as a hidden form of capital—and mechanism of stratification—by facilitating the conversion of tangible resources into advantageous outcomes.



As the nation’s two largest racial/ethnic groups, the U.S. Hispanic and White populations will feature prominently in unfolding economic and social trends over the coming decades. At the same time, Hispanic–White differences in 4-year college enrollment (56% versus 72%) and completion (11% versus 22%; Fry & Taylor, 2013) constitute two of the greatest racial/ethnic disparities in educational outcomes among U.S. students (Carnevale & Strohl, 2014). These differences have increasingly captured the attention of researchers and policymakers, who frequently attribute divergent postsecondary pathways among Hispanic and White youth largely to gaps in parent resources (O’Connor, Hammack, & Scott, 2010).


In contrast, a growing number of scholars instead argue that Hispanic–White disparities in college pathways are not driven simply by different levels of family resources but by the fact that Hispanic students experience the college preparation and enrollment process differently from their White counterparts (Pérez & Ceja, 2015; Pérez & McDonough, 2008; Ryan, 2016; Welton & Martinez, 2014). Acknowledging this evolution in the Hispanic college access literature, Pérez and Ceja (2015) nonetheless observe that less is known about how and why college access, choice, and enrollment processes transpire differently for Hispanic students relative to other groups.


Several recent studies (e.g., Alon, Domina, & Tienda, 2010; O’Connor et al., 2010) suggest that one way in which the college preparation and enrollment process differs among Hispanic and White youth is in how parent resources, and particularly parent social capital, facilitate college preparation and enrollment. In response, the current study empirically tests for differences in the process by which parent resources appear to influence college enrollment among similar Hispanic and White youth.


In this study I examine whether group-level differences in the operation of parent social capital help explain the finding (Alon et al., 2010; O’Connor et al., 2010) that parent income and education do not function the same way among Hispanic and White youth, particularly when it comes to 4-year college enrollment. Using a national longitudinal dataset and structural equation modeling techniques, I also address the absence of research about factors that may mediate associations between parent resources and student college enrollment. Specifically, I ask whether forms of parent social capital, along with parent income and education, are differently associated with enrollment due to group-level differences in how these resources facilitate college-aligned actions among White and Hispanic youth. Informed by the scholarship of Schneider and Stevenson (1999), college-aligned actions refer to the set of key steps that students must take during high school in order to align their college preparation with their educational expectations. I focus on the distinct subset of Hispanic and White students who complete high school expecting to complete at least a bachelor’s degree.


While Alon et al. (2010) demonstrated that Hispanic youth did not experience the same enrollment boost as did similar White youth given the same increase in parent education, these researchers did not account for students’ educational expectations. Thus, it is difficult to know whether Hispanic youth were less likely than White youth to enroll in a 4-year institution because their parents’ socioeconomic resources were not as useful or because fewer Hispanic students had expectations of doing so. On the eve of high school graduation, youth are more objectively aware of whether they are likely to attain their expectations than at any earlier point (Mickelson, 1990; Park, Wells, & Bills, 2015). By focusing on bachelor’s-expecting high school graduates, I am able to study the process by which parent resources are converted into 4-year college enrollment among those Hispanic and White youth most likely to have concluded that achieving this level of education is a realistic possibility (Park et al., 2015).


Moreover, the focus on Hispanic and White students who complete high school with bachelor’s degree expectations reflects a critical quantitative approach to the study of college enrollment. According to a critical quantitative perspective, applying widely accepted models and assumptions to all students often provides a myopic perspective on the actual experiences of particular groups and especially the experiences of historically marginalized groups (Stage, 2007). In particular, relegating the potential for differences among racial/ethnic groups to a dummy indicator in statistical models yields little information about how aspects of the process under study vary across groups. Here, the groups of interest include Hispanic and White students who complete high school with the expectation of completing a bachelor’s or advanced degree. Focusing on these students acknowledges the fact that youth who do and do not expect to complete a bachelor’s degree as they finish high school represent two distinct cohorts of students that merit further study (Wells, Seifert, & Saunders, 2013).


Few studies have empirically tested how the process of preparing for and enrolling in college differs among Hispanic and White youth (Pérez & Ceja, 2015). Specifically, there is a need to better understand Hispanic–White differences in how parents’ economic, human, and social capital resources operate as well as factors that may mediate the association between parent resources and college enrollment. Focusing on the distinct subset of Hispanic and White youth who finish high school expecting to complete a 4-year degree, this research addresses these gaps. The findings lend empirical support to Tienda’s (2011) suggestion that differences in the operation of parent resources, and especially forms of parent social capital, between Hispanic and White students may contribute to their different patterns of college entry. The findings also confirm the critical advantage for both groups of engaging in actions that align with their educational plans.


PARENT SOCIAL CAPITAL AND HISPANIC–WHITE DIFFERENCES IN COLLEGE ENROLLMENT

Following an extensive status attainment literature documenting how family resources and students’ educational expectations shape the educational attainment process (Sewell, Hauser, Springer, & Hauser, 2003), previous research demonstrates how disparate levels of parent income and education contribute to divergent educational outcomes for Hispanic and White students. Among both groups, family income shares a positive relationship with students’ educational expectations (Bohon, Johnson, & Gorman, 2006), their college preparation (Cabrera & LaNasa, 2001), and the rate at which they enroll in college (Hurtado, Inkelas, Briggs, & Rhee, 1997; Perna & Titus, 2005). There is also a strong relationship between parent education and students’ postsecondary plans and outcomes (Cabrera & LaNasa, 2001; Choy, 2001; Horn & Nuñez, 2000).


However, recent scholarship indicates that even when Hispanic and White parents have similar levels of income and education, these resources do not operate the same way when it comes to their children’s college enrollment. Considering enrollment in a bachelor’s-granting institution, O’Connor et al. (2010) demonstrated that parent socioeconomic status generated higher returns for White students than for Hispanic students. Similarly, Alon et al. (2010) found that Hispanic students benefited less from increases in parental education relative to similar White students. The latter finding led Tienda (2011) to postulate a more nuanced process involving parents’ social capital resources. Tienda surmised that it may be more difficult for Hispanic parents to leverage their economic and human capital on behalf of their children if they experience lower returns to forms of parent social capital, and particularly forms reflecting access to social networks in which critical information about 4-year college enrollment is exchanged.


Social capital, or access to resources available through one’s social ties, has been asserted as one manifestation of socioeconomic status (Kerckhoff & Campbell, 1977; Stanton-Salazar & Dornbusch, 1995). This perspective suggests that attempting to investigate the relationship between parent resources and children’s college enrollment without attending to the ways in which advantage is transferred to children via their own or their parents’ social networks ignores the role of social capital in the production of human capital (Coleman, 1988; Desmond & Turley, 2009). Social capital theory posits that it is often through relationships with others that individuals gain access to information, advice, and assistance (Bourdieu, 1986; Coleman, 1988; Lin, 2001; Portes, 1998), making the theory a natural lens through which to understand and explain access to college-relevant information among students and parents. In this study, social capital is defined as individuals’ capacity to gain access to valued resources—including information, advice, and social support—by virtue of their relationships with others. This definition captures elements of both Coleman’s (1988) as well as Bourdieu’s (1986) work on social capital.


Coleman’s (1988) functionalist interpretation of social capital is most frequently deployed in the educational literature and is typically used to study the positive effects of parent social capital on school-related outcomes (Dika & Singh, 2002; Ream, 2005). Coleman sees parents’ roles as predominant in promoting their children’s status attainment, making his work highly applicable to research on student college-going (O’Connor et al., 2010). Especially relevant, Coleman describes the information and support channels that operate among parents, and the ties between parents and their children, as important forms of social capital.


In contrast to Coleman, Bourdieu’s (1986) conceptualization of social capital attends to social realities and the ways in which some individuals are advantaged because of their membership in particular groups while other individuals and groups repeatedly come up against the restrictions imposed by structural barriers (Dika & Singh, 2002; Portes, 1998). Bourdieu proposes that the volume and quality of social capital possessed by a person depends on both the size of his or her network of connections and the volume of capital—economic, human, cultural, and symbolic—possessed by each connection or social tie. According to Bourdieu, social capital often functions as a hidden form of capital, augmenting its capacity to serve as a mechanism of stratification (Ream, 2005; Stanton-Salazar, 2001). Similarly, Stanton-Salazar’s (2001) critical network analytic perspective views individuals as embedded within multiple webs interwoven across individual, community, and societal levels and intersected by class, race, and gender stratification. Stanton-Salazar emphasizes inequitable access to particular social and institutional contexts among Hispanic youth and parents, as well as the uneven distribution of opportunities to form relationships with those who wield control over institutional resources. The social capital perspectives of both Bourdieu and Stanton-Salazar direct attention to the fact that while Hispanic parents draw upon various forms of cultural wealth to support the academic goals of their children (Yosso, 2005), societal stratification along the dimensions of race/ethnicity and class may limit their access to social networks in which key resources, including information about how the educational system works, are embedded (Pérez & McDonough, 2008; Perna & Titus, 2005; Ryan, 2016; Tienda, 2011).


Within the family, parent–child relationships provide the foundation for social capital development (Coleman, 1990). Numerous studies document the educational benefits of parents’ supportive interactions with their own children (Hart & Risley, 1995), their informal relationships with other parents (Carbonaro, 1998; Horvat, Weininger, & Lareau, 2003), and their more formal relationships with school personnel (Delgado-Gaitan, 1992). Many of the ways in which parents are involved in the education of their children fit nicely within a social capital framework given that parents’ relationships with their children, other parents, and school personnel are all important means through which parents accrue and transfer human capital (e.g., college information and know-how) to their children (Perna & Titus, 2005). It is unsurprising, then, that education researchers frequently use the concept of social capital explicitly in their consideration of parent involvement, both at home and in the formal school setting (Dika & Singh, 2002).


Among Hispanic students, relationship dynamics within the family play an especially central role in postsecondary transitions (Alvarez, 2015; Kao, Vaquera, & Goyette, 2013). While many Hispanic and White adolescents look to institutional agents such as school counselors and teachers for academic guidance and college information, Hispanic students also tend to rely more on their parents and other family members (Kao et al., 2013; Stanton-Salazar, 2001; Tornatzky, Cutler, & Lee, 2002). Sometimes family support occurs through direct intervention in children’s schooling (Delgado-Gaitan, 1992) or via the norms and expectations that parents and other family members communicate to youth. Further, when parents participate in extrafamilial social networks that include, for example, the parents of similar-aged children or school staff, they may gain access to crucial information about how to help their offspring successfully navigate educational processes and experience academic success (Carbonaro, 1998; Horvat et al., 2003). Indeed, when it comes to preparing for and enrolling in college, Hispanic parents’ information channels and the ability to connect students to educational resources, including college knowledge, appear especially important (Alvarez, 2015; Kiyama, 2011).


Research in higher education suggests that Hispanic families in particular may face unique barriers that inhibit access to, and activation of, college-relevant network-bound information, especially when it comes to preparing for entry into a 4-year college or university (Marsico & Getch, 2009; Rosenbaum & Person, 2006; Tornatzky et al., 2002). Numerous researchers have argued that, relative to other groups including their non-Hispanic White counterparts, parents of Hispanic youth are less likely to be situated in social networks where accurate and timely college information is regularly exchanged (Gándara & Contreras, 2009; Nora & Crisp, 2009; O’Connor et al., 2010; Pérez & McDonough, 2008; Rosenbaum & Person, 2006). Differential access to college information and support via social ties may result in Hispanic–White differences in the utility of parent social capital for college preparation and enrollment. This is likely because without ready access to complete and accurate information about college preparation, some Hispanic parents may struggle to use their social capital resources to promote the types of college-aligned behaviors likely to facilitate children’s 4-year college enrollment (Ryan, 2016; Ryan & Ream, 2016).


COLLEGE-ALIGNED ACTIONS AS A MECHANISM FOR PARENT SOCIAL CAPITAL

It has been demonstrated elsewhere that enrollment differences persist even among Hispanic and White students with consistently high educational expectations (Klasik, 2012; O’Connor et al., 2010). This may be because even among students with bachelor’s degree expectations, Hispanic and White youth are not equally likely to engage in college-aligned actions by preparing for college in ways that match their degree goals. Schneider and Stevenson (1999) showed that many American youth report high educational expectations but are underinformed about how to actually achieve specific educational goals. Schneider and Stevenson characterized these youth as demonstrating misaligned ambitions. Their work has been succeeded by numerous studies demonstrating that while college ambitions are a necessary precursor to bachelor’s degree attainment, expectations alone are not enough (Avery & Kane, 2004; Bozick, Alexander, Entwistle, Dauber, & Kerr, 2010; Grodsky & Riegle-Crumb, 2010; Klasik, 2012; Morgan, 2005; Roderick, Coca, & Nagaoka, 2011). Students must also engage in the college-aligned actions that will best prepare them to fulfill their goals. For instance, youth must become academically prepared for college coursework, which means enrolling in rigorous courses and maintaining good grades (Adelman, 2002; Berkner & Chavez, 1997). They must also complete various detailed tasks—submitting college applications, taking national entrance exams, applying for financial aid—in time to meet specific deadlines (Roderick et al., 2011).


Schneider and Stevenson asserted the importance of parent social capital for helping youth align their college ambitions and preparation. D. H. Kim and Schneider (2005) further hypothesized that when parents connect their offspring to resources that support students’ ambitions, these efforts become social capital in action. Building from these earlier studies, I refer to students who engage in preparatory actions that are aligned with their educational expectations as demonstrating college-aligned actions (Ryan, 2016; Ryan & Ream, 2016). Parent social capital may play an important role in the process by which youth become college-aligned (Schneider, Judy, Ebmeyer, & Broda, 2014). However, racial/ethnic stratification may constrain access to social networks in which college-relevant resources tend to be concentrated, raising the possibility that some forms of parent social capital may be differently associated with college-aligned actions among similar Hispanic and White youth. Moreover, Hispanic–White differences in the utility of parent social capital for promoting college-aligned actions may affect parents’ ability to leverage other socioeconomic resources as their children navigate college enrollment (Tienda, 2011).


RESEARCH QUESTIONS

By applying multiple group structural equation modeling techniques to a national longitudinal dataset, I test whether group-level differences in the functioning of parent social capital predict differences in 4-year college enrollment between bachelor’s-expecting Hispanic and White youth, at least in part through an influence on college-aligned actions. The conceptual model that serves as the theoretical framework for this study is presented in Figure 1. As depicted in the conceptual model, it is expected that forms of parent social capital will be associated with student engagement in college-aligned actions, which will in turn be related to the level at which youth are enrolled in college 2 years beyond high school. I predict that differences in the utility of parents’ social resources may help explain previously documented differences in returns to parents’ economic and human capital resources among Hispanic and White youth. Therefore, parent income and education are also included as focal predictors of student engagement in college-aligned actions and subsequent college enrollment. A number of important student background characteristics are also accounted for. The conceptual model guides the investigation of the following research questions:


To what extent are parent social capital and other parent resources indirectly associated with college enrollment via an influence on student engagement in college-aligned actions among bachelor’s-expecting Hispanic and White youth?

To what extent are the associations among parent social capital and other parent resources, college-aligned actions, and college enrollment measurably different among bachelor’s-expecting Hispanic and White youth?


Figure 1. Conceptual framework

[39_21849.htm_g/00002.jpg]


Note: While control covariates were employed in estimation of the full model, for ease of readability these associations are not depicted in the conceptual framework. In the conceptual framework, ellipses depict latent measures while rectangles represent observed measures. Dashed lines indicate indirect effects between parent resources and college enrollment via college-aligned actions. A baseline model (the configural invariance model) established the same pattern of fixed and free factor loadings across the two groups and invariance testing procedures were used to evaluate the extent to which structural coefficients varied across the three groups (see Appendix for further details).


METHODS

DATA SOURCE AND PARTICIPANTS

This research used data from the Education Longitudinal Study of 2002 (ELS:2002), which were collected by the National Center for Education Statistics (NCES). The study was designed to explore students’ transitions from high school into college and the workforce. The ELS data (see http://nces.ed.gov/surveys/els2002/) were collected through a two-stage stratified random sampling strategy and include a nationally representative cohort of students in the 10th grade in 2002 (Ingels et al., 2007). Multiple respondent populations were surveyed during the base year ELS data collection in 2002, including students, parents, teachers, and school administrators. In addition to the base year interview, student data were also collected during interviews in 2004 and 2006 and from high school transcripts.


The research sample was drawn from the base year panel of 2002 Hispanic and White 10th graders who participated in all three waves of data collection (n=9,470). In order to achieve this sample, all statistics were weighted by the first and second follow-up panel weight (F2F1WT) and the Grade 10 cohort flag (G10COHRT). All sample sizes were rounded to the nearest ten in accordance with NCES guidelines. The sample was further limited to students who had obtained a high school diploma or an alternative credential by June of 2005 and who, as high school seniors, reported ultimately expecting to complete a bachelor’s (or advanced) degree (n=6,440). Approximately 62% and 74%, respectively, of Hispanic and White students who participated in all three waves of ELS data collection finished high school intending to complete a bachelor’s degree or higher. The final research sample included 1,020 Hispanic students and 5,420 non-Hispanic White students.


Dependent Variable


The key outcome of interest in this investigation was student college enrollment status in 2006. The five-level ELS variable F2PS0601 was recoded as a three-level outcome reflecting whether a student was enrolled in no postsecondary institution (coded as 0), in a less-than-4-year institution (coded as 1), or in a 4-year institution (coded as 2).


Parent Social Capital


A subset of items from the ELS parent survey, administered in 2002, was used to measure three latent constructs reflecting parents’ access to college-relevant information and support via relationships with their children, relationships with the parents of their children’s friends, and interaction with school personnel. This approach facilitated the measurement of parent social capital across familial and extrafamilial domains, accounting for both quantity (i.e., presence of a relationship) and quality (i.e., nature of that relationship). These measures are similar in construction to measures of parent social capital shown to be associated with college enrollment in prior research (Kao & Rutherford, 2007; Nuñez & Kim, 2012; Perna & Titus, 2005; Ryan, 2016) and demonstrated high internal consistency (ρ). The three latent constructs representing various forms of college-relevant parent social capital included:


College-Relevant School Social Capital (or SSC; ρ=.85): a three-item construct including the frequency with which the parent contacted the school during the prior year about the student’s course selection, academic program, and post-high school plans.

College-Relevant Family Social Capital (or FSC; ρ=.81): a three-item construct including the frequency with which the parent provided the student with advice during the prior year about selecting high school courses, taking entrance exams, and applying to college.

Intergenerational Closure (ρ=.87): a five-item construct including whether the parent reported knowing the mother and/or the father of the student’s closest friend, the number of times a parent of one of the student’s friends gave the parent advice about the school’s courses or teachers, how often the parent provided a favor to the student’s friend’s parent, and how often the parent received a favor from the student’s friend’s parent.


Confirmatory factor analysis (CFA) was used to analyze the measurement model specifying the associations between latent constructs and their observed indicators. All models were assessed for fit according to the chi-square (χ2) statistic, the comparative fit index (CFI) and the root mean square error of approximation (RMSEA). Departures from normality or sample sizes above 200 nearly always lead to rejection of the model according to the chi-square statistic, even when the model is properly specified (McIntosh, 2006). However, the CFI accounts for sample size, with a value above 0.90 indicating a good fit between the model and the data. The suggested upper-limit cutoff for the RMSEA is 0.06 (Hu & Bentler, 1999). A multiple group CFA indicated that the measurement model held for the combined Hispanic and White sample (χ2=1298.33(84), p=.00, CFI=.98, RMSEA=.05). Separate CFA analyses for each group also indicated good model fit (results available upon request). Convergent validity was evaluated through examination of standardized factor loadings (see Table 1).


Other Parent Resources


In addition to the forms of parent social capital described above, the other parent resources included in the models were family income, as represented by an indicator of parents’ combined income from all sources in 2001, and parent education, as represented by an indicator of whether either of the student’s parents completed a bachelor’s or advanced degree. The income measure, taken during base year data collection, was rescaled by NCES as a 13-level continuous variable prior to data release (no income=0, $200,000 or more=13). While information about parental wealth may provide a more complete picture of the family financial situation, the ELS parent questionnaire did not collect information about the kinds of financial assets used to construct measures of wealth.


College-Aligned Actions


It is expected that the extent to which bachelor’s-expecting students engage in instrumental college-going actions during high school will be highly associated with their probability of enrolling in college, and particularly in a 4-year institution. The latent college-aligned actions measure captures the shared variation among indicators of these key actions. A student’s level on the aligned actions construct reflects the extent to which s/he engaged in actions during high school that align with preparation for entry into a 4-year college. The work of Berkner and Chavez (1997) and others (Klasik, 2012; Roderick et al., 2011; Schneider & Stevenson, 1999) informed the selection of indicators for this construct. These researchers have documented a sequence of steps required to master the process of 4-year enrollment. The aligned actions construct was measured using indicators of these steps available in the ELS (see Table 1), including: 1) a six-level indicator of the student’s cumulative high school grade point average; 2) a three-level indicator reflecting the highest math course s/he completed; 3) a binary indicator of national college entrance exam completion; and 4) a three-level indicator of the number of 4-year colleges to which s/he submitted an application. The data used to create the college-aligned actions construct were based on student report during data collection in 2004; postsecondary application was verified by NCES using the Integrated Postsecondary Education Data System (IPEDS). The reliability coefficient for this construct was .84.


Background Variables


This study focuses primarily on how parent social capital and other parent resources, as well as student college-aligned actions, were associated with level of college enrollment among bachelor’s-expecting Hispanic and White adolescents. However, other key characteristics associated with 4-year college enrollment among Hispanic and White youth were also accounted for. For example, females tend to be more likely to enroll than males (Riegle-Crumb, 2010), as are students with higher prior achievement levels (Deil-Amen & Turley, 2007). Further, while researchers have demonstrated that high educational expectations are an important predictor of college enrollment, sustained plans for educational attainment over time appear to be especially predictive of college enrollment (Kao & Tienda, 1998), possibly by increasing the likelihood that youth will align their college preparatory behaviors to their educational goals (Bozick et al., 2010; Klasik, 2012). Beyond various parent resources, whether or not youth live in the same home with both biological parents is another family characteristic that has been demonstrated as important for postsecondary outcomes (Deil-Amen & Turley, 2007). Finally, students attending a Catholic or other private high school are much more likely to enroll in college, particularly at the 4-year level (Coleman & Hoffer, 1987; Gándara et al., 2013).


A number of control covariates, all collected in 2002, were used in the analytic models to help account for these background characteristics. An indicator of whether the student also expected to complete a bachelor’s degree in the 10th grade was included in all models (consistent expectations=1). The base year standardized test score composite, which reflects a student’s achievement in math and language arts in 10th grade, was included to control for achievement differences prior to data collection. Several other indicators were included to control for gender (female=1), the student’s family structure in the 10th grade (student who does not live with both biological parents=1), and high school type (Catholic or other private=1).


Table 1. Measurement Model Descriptions and Standardized Factor Loadings


Latent Construct and ELS:2002 Indicator



Item Description

Factor Loading

Hispanic

White

College-Relevant SSC

  

 BYP53B

Parent contacted school about student school program for year

0.67

0.69

 BYP53C

Parent contacted school about student plans after high school

0.87

0.80

 BYP53D

Parent contacted school about student course selection

0.88

0.76

   

College-Relevant FSC

  

 BYP56A

Parent provided advice about selecting courses or programs

0.75

0.74

 BYP56B

Parent provided advice about plans for college entrance exams

0.85

0.81

 BYP56C

Parent provided advice about applying to college/school after high school

0.73

0.61

   

Intergenerational Closure

  

 BYP59DA

Parent knows mother of 10th grader’s 1st friend

0.57

0.59

 BYP59EA

Parent knows father of 10th grader’s 1st friend

0.33

0.47

 BYP60A

Times child’s friend’s parent gave advice about teachers and courses at the school

0.71

0.72

 BYP60B

Times child’s friend’s parent gave a favor (to parent)

0.72

0.70

 BYP60C

Number of times friend’s parent received a favor (from parent)

0.67

0.67

Aligned Actions

   

 F1RGPP2

GPA for all courses taken in the 9th–12th grades

0.73

0.76

 F1HIMATH

Highest math course completed by student

0.72

0.75

 F2PSEEXM

Whether student took college entrance exams

0.81

0.82

 F2NAPP2P

Non-open enrollment schools student applied to

0.67

0.60

Note: Hispanic n=1,020; White n=5,420. All factor loadings significant at p≤.001.


ANALYTIC STRATEGY

In order to examine the possibility that forms of parent social capital indirectly influence children’s 4-year college enrollment through an association with students’ college-aligned actions, this study relied on the process outlined by Baron and Kenny (1986) and others (Mackinnon, Lockwood, Hoffman, West, & Sheets, 2002). The Mplus “MODEL CONSTRAINT” command was used to obtain three types of mediation output that parallel the three conditions Baron and Kenny described as necessary for mediation: 1) total effects, or the associations between parents’ social capital and other resources and 4-year college enrollment prior to accounting for the hypothesized college-aligned actions mediator; 2) indirect effects, or the extent to which parent social capital and other parent resources were associated with 4-year college enrollment through an association with college-aligned actions, and; 3) direct effects, or the associations between parents’ social capital and other resources and 4-year college enrollment after accounting for college-aligned actions. Direct effects provide information about the extent of mediation given a significant indirect effect. Standard errors for the indirect effects were estimated using the delta method, and any additional bias in the standard errors was corrected for through the use of the bootstrap procedure in Mplus.


Group-level differences in the statistical significance of parameter estimates do not definitively indicate whether the magnitudes of parameter estimates differ measurably across groups. In other words, differences in the statistical significance of estimates do not indicate whether the operation of parent resources differs among Hispanic and White youth. For this reason, model difference testing was conducted in order to statistically evaluate whether the magnitudes of the associations between parents’ social and other resources, college-aligned actions, and college enrollment differed in strength among Hispanic and White youth.


The robust weighted least squares (WLSMV) estimator was used to estimate the models described here (see Appendix for model equations). The WLSMV estimator is most appropriate for this research given that most of the observed variables, including the outcome, were categorical in nature. This estimator uses the probit link and provides standard errors and a chi-square test statistic that are robust to non-normality (L. K. Muthén & Muthén, 2015). Because the sample consisted of children nested in schools, violating the assumption of independence, the Mplus setting TYPE=COMPLEX was used to adjust standard errors for nonindependence of observations, sample stratification and selection bias (L. K. Muthén & Muthén, 2015).


RESULTS

The results are presented in two sections. The first section summarizes descriptive statistics on the observed variables included in the study. In the second section, the hypothesized model is tested by simultaneously estimating a series of regression equations.


DESCRIPTIVE ANALYSIS

Students


Eighty-two percent of Hispanic and 91% of White students also held bachelor’s degree expectations in the 10th grade, suggesting that most high school seniors with bachelor’s degree plans have maintained this goal over time. About half of each group was comprised of females, while the average composite test score was almost seven points higher among White students relative to Hispanic students. A smaller share of Hispanic youth lived with both parents and attended a private or Catholic school than was the case among their White peers. White youth were more likely than Hispanic youth to have engaged in each of the behaviors comprising college-aligned actions and to be enrolled in a 4-year institution in 2006. Further descriptive details for all observed variables are provided in Table 2.


Parents


Group-level differences were apparent across several of the indicators of the three forms of parent social capital, yet latent mean difference testing (results available upon request) indicated that the parents of Hispanic and White students only differed significantly in average levels of intergenerational closure. The average family income earned by parents of Hispanics in 2001 was just over $30,000 and the average level of education was some college. Parents of the average White sample member had a combined family income of just over $50,000 and an average education level of some college.


Table 2. Descriptive Statistics on Primary Study Measures

 

Hispanic

White

 
 

M(SD)

M(SD)

Range

Parent Social Capital

   

College-Relevant SSC

How often parent contacted school

about…

…academic program**

…course selection**

…plans after high school




1.36(0.61)

1.27(0.53)

1.23(0.50)




1.48(0.67)

1.34(0.54)

1.24(0.50)




1(never)–4(5+ times)a

1(never)–4(5+ times)

1(never)–4(5+ times)

College-Relevant FSC

How often parent provided advice about…

…course selection**

…taking college entrance exams

…applying to college




2.36(0.68)

2.15(0.79)

2.16(0.80)




2.45(0.61)

2.18(0.72)

2.15(0.75)




1(never)–3(often)b

1(never)–3(often)

1(never)–3(often)

Intergenerational Closure

Knows mother of child’s 1st friend**

Knows father of child’s 1st friend**

How often parent of child’s friend…

…gave advice about teacher/course**

…provided a favor to parent**

…received a favor from parent**


0.79(0.41)

0.57(0.50)


1.33(0.66)

2.07(1.13)

2.17(1.11)


0.88(0.33)

0.77(0.42)


1.56(0.80)

2.44(1.05)

2.54(1.05)


0(no)–1(yes)

0(no)–1(yes)


1(never)–4(5+ times)a

1(never)–4(5+ times)

1(never)–4(5+ times)

Other Parent Resources

   

Income**

8.31(2.41)

10.02(1.85)

1($0)–13($200,000+)c

Education**

0.91(0.80)

1.39(0.73)

0(none)–2(≥ 4-year)

College-Aligned Actions

   

High school GPA**

Highest math course completed**

Has taken a college entrance exam**

Four-year institutions applied to**

4.06(1.37)

1.23(0.76)

0.68(0.47)

1.40(1.21)

4.74(1.19)

1.53(0.68)

0.90(0.30)

1.76(1.12)

0(0–1.00)–6(3.51–4.0)

0(< AlgII)–2(> AlgII)

0(no)–1(yes)

0(none)–3 (> two)

College Enrollment

   

Level of enrollment in 2006**

2.00(0.85)

2.37(0.82)

0(none)–2(4-year)


Background Covariates

   

Consistent Expectations**

0.82(0.39)

0.91(0.29)

0(no)–1(yes)

Test scores**

49.18(8.50)

56.47(8.03)

20.9–81.04

Female

0.58(0.49)

0.55(0.50)

0(male)–1(female)

Family composition**

0.40(0.49)

0.27(0.44)

0(both parents)–1(other)

Private school**

0.23(0.42)

0.35(0.48)

0(public)–1(private)

Note: Hispanic n=1,020; White n= 5,420.

a 1=never; 2=1–2 times; 3=3–4 times; 4=5 or more times (since the beginning of the school year)

b 1=never; 2=sometimes; 3=often

c The 13 income levels used by NCES are as follows: (1) no income; (2) $1,000 or less; (3) $1,001–$5,000; (4) $5,001–$10,000; (5) $10,001–$15,000; (6) $15,001–$20,000; (7) $20,001–$25,000; (8) $25,001–$35,000; (9) $35,001–$50,000; (10) $50,001–$75,000; (11) $75,001–$100,000; (12) $100,001–$200,000; (13) $200,001 or more.

* p≤.05, ** p≤.01 (group mean differences)


STRUCTURAL EQUATION MODELING

In this section I investigate how parent resources were associated with college enrollment both directly and indirectly via their relationship with college-aligned actions across the two groups. I then evaluate the possibility that the magnitudes of the hypothesized associations differed between Hispanics and Whites.


The Indirect Associations Among Parent Social Capital and Other Resources and College Enrollment Via College-Aligned Actions


Table 3 presents the results from the full structural model, which accounted for the associations that parent social capital and other parent resources shared with both college-aligned actions and college enrollment while controlling for background characteristics. The full structural model, which fit the data well across groups (χ2=928.96(384), p=.00, CFI=.99, RMSEA=.02), explained about 61% of the variability in 2006 enrollment status among Hispanic students and about 67% among White students.


For both groups, the results from this model clearly demonstrated the strong positive relationship between students’ engagement in college-aligned actions and their level of college enrollment. After accounting for associations between forms of parent social capital and students’ aligned action-taking, none of the three forms of parent social capital shared a significant direct relationship with college enrollment in either group. However, among White youth, over and above a relationship with student college enrollment via college-aligned actions, both parent income and parent education in the form of a bachelor’s degree also shared a significant direct association with student college enrollment (p=.02 and p=.05 respectively).


Table 3. The Associations Between Parent Social Capital and Other Resources and College Enrollment After Accounting for College-Aligned Actions (Unstandardized and Standardized Coefficients)


College Enrollment Level Hispanic White

 

b

SE

B

b

SE

B

Parent Social Capital







College-Relevant SSC

-0.07

 0.09

-0.05

 0.08

0.06

0.04

College-Relevant FSC

-0.04

-0.08

-0.03

0.01

0.08

0.00

Intergenerational Closure

0.10

0.10

0.07

 0.05

0.10

0.01

 







Other Parent Resources







Income

0.01

0.02

0.02

0.03**

0.01

0.05

Education: BA or above

0.17

0.12

0.13

0.13*

0.07

0.11

 







College-Aligned Actions

0.97***

0.08

0.81

1.20***

0.13

0.90

 







Background Covariates







Consistent expectations

-0.17

0.11

-0.16

-0.14**

0.06

-0.13

Test scores

-0.11

0.08

-0.11

-0.25**

0.04

-0.08

Female

-0.08

0.09

-0.07

-0.02

0.04

-0.02

Family composition

-0.05

0.10

-0.08

 0.11*

0.05

0.09

Private school

 0.02

0.12

-0.04

 0.01

0.06

0.00

Fit Statistics

χ 2

CFI

RMSEA

R2 Hispanic

R2 White

928.96(384), p=.00

0.99

0.02 (CI: 0.01–0.02)

.61

.67

Note: Hispanic n=1,020; White n=5,420. CFI=Comparative Fit Index; RMSEA=Root Mean Square Error of Approximation; CI=90% confidence interval.

* p≤.05, ** p≤.01, *** p≤.001


In order to evaluate whether college-aligned actions served to mediate any of the associations between parent resources and college enrollment for either group, the total effect of each form of parent social capital, as well as of parent economic and human capital, was broken down into its direct and indirect effects. Relative to the total effect, if the remaining direct effect of any form of capital on college enrollment decreased after accounting for its indirect effect via college-aligned actions, this would offer evidence of mediation. More specifically, if the remaining direct effect (total effect minus indirect effect) of any parent resource on college enrollment became nonsignificant relative to its total effect, this would suggest full, as opposed to partial, mediation (Baron & Kenny, 1986). The effect decompositions are provided in Table 4.


Among Hispanic students, parent human capital in the form of a bachelor’s or advanced degree had a significant total (direct plus indirect) effect on college enrollment. However, none of the indirect effects between forms of parent social capital, parent income, or parent education and college enrollment were statistically significant. This result provides evidence that associations between parent resources and college enrollment generally do not travel through an influence on student engagement in college-aligned actions among Hispanic youth. While not statistically significant, it is worth noting that the indirect and total effects of college-relevant SSC were positive among Hispanic youth, yet were negative for their White peers. I return to this somewhat counterintuitive finding in the discussion.


Among White students, all forms of parent social capital as well as parent income and education shared significant indirect associations with enrollment via college-aligned actions (see Table 4). Engagement in college-aligned actions appeared to fully mediate the relationships that both college-relevant FSC and intergenerational closure shared with enrollment; the direct effects of these two forms of parent social capital on college enrollment were indistinguishable from zero in the presence of the mediator. Engagement in college-aligned actions also partially mediated the associations that both parent income and parent education shared with student college enrollment among White youth; the direct effects for these resources were reduced in size and significance in the presence of the mediator.


Table 4. Decomposition of the Total Effects of Parent Social Capital and Other Resources on College Enrollment (Unstandardized Coefficients)

 


Direct Effect



Indirect Effect via

College-Aligned Actions


Total Effect


 

Hispanic

White

Hispanic

White

Hispanic

White

 

b

SE

b

SE

b

SE

b

SE

b

SE

b

SE

Parent Social Capital













College-Relevant SSC

-0.07

0.09

0.08

0.06

0.12

0.07

-0.16**

0.05

0.05

0.10

-0.09

0.06

College-Relevant FSC

-0.04

0.08

0.01

0.08

0.04

0.06

 0.30**

0.11

0.00

0.08

0.31**

 0.11

Intergenerational Closure

0.10

0.10

0.05

0.10

0.05

0.08

0.33**

0.12

0.15

0.13

0.37**

0.14

 













Other Parent Resources













Income

0.01

0.02

0.02*

0.01

0.01

0.02

0.03**

0.01

0.02

0.02

0.06***

0.01

Education: BA or above

0.17

0.12

0.13*

0.07

0.10

0.11

0.25**

0.06

0.27*

0.12

0.38**

0.07

Note: Hispanic n=1,020; White n=5,420.

* p≤.05, ** p≤.01, ***p≤.001


Converting the probit coefficients from the model to predicted probabilities, the results imply that for the average student in each sample, a standard deviation increase in college-aligned actions (holding all else constant) would raise the predicted probability of enrollment in a 4-year institution from 27% to 76% among Hispanic youth and from 58% to 97% among Whites (see Table 5). Holding all else constant, shifting parent education to a bachelor’s or advanced degree would lead to the second largest change in the probability of 4-year enrollment across groups, or about a 13 percentage point increase for both Hispanic and White youth. The results also suggest that standardized changes in parent income and most forms of parent social capital would be associated with anywhere from a 6 to 7 percentage point increase in the probability of 4-year enrollment for White students, and with a 2 to 5 percentage point increase among Hispanic youth.


Table 5. Predicted Probabilities of College Enrollment

 

Predicted Probability

Hispanic

White

None

< 4-year

4-year

None

< 4-year

4-year

Baseline (average student)

0.29

0.44

0.27

0.12

0.30

0.58

       

College-relevant SSC + 1SD

0.27

0.44

0.29

0.14

0.31

0.54

College-relevant FSC + 1SD

0.29

0.44

0.27

0.09

0.26

0.65

Intergenerational closure + 1SD

0.24

0.44

0.32

0.09

0.26

0.65

Parent income + 1SD

0.27

0.44

0.29

0.09

0.27

0.64

At least one parent with BA or above

0.18

0.42

0.40

0.07

0.23

0.70

       

College-Aligned Actions

      

Aligned actions + 1SD

0.03

0.21

0.76

0.00

0.03

0.97

 Note: Hispanic n=1,020; White n=5,420. Baseline probabilities reflect a female student with consistent expectations and sample mean standardized test scores who attended a public school and lived with both parents, who together held average stocks of all resources included in this research. Estimates in Table 5 reflect the expected change in the predicted probability of enrollment at each level (none, < 4-year, 4-year) for the indicated change in a predictor while holding all else constant.

 

Group-Level Differences in Model Functioning


Group-level differences in the statistical significance of parameter estimates do not definitively indicate whether the magnitudes of parameter estimates differ measurably across groups. For this reason, model difference testing was conducted in order to statistically evaluate whether parent resources were indeed more strongly associated with college-aligned actions and college enrollment among White youth relative to similar Hispanic youth. Details regarding the measurement and structural invariance testing procedures are provided in the Appendix.


The results (Table 6) from invariance testing indicated that both the indirect and total associations between both parent income and education, on the one hand, and student college enrollment, on the other, did not differ significantly across groups. However, the indirect effects of SSC (Χ2(1)=9.99, p=.00), of FSC (Χ2(1)=5.05, p=.02), and of intergenerational closure (Χ2(1)=4.46, p=.03) were of greater magnitude among White students relative to similar Hispanic students. Further, the total associations between both college-relevant SSC and FSC and college enrollment also differed between the two groups (Wald∆(3)=15.11, p=.00 and Wald∆(3)=9.22, p=.03, respectively). In other words, the previously observed pattern by which similar increases in parent education and income offered greater benefits for White youth than for Hispanic youth (Alon et al., 2010; O’Connor et al., 2010) was not observed in the current study, which accounted for forms of parent social capital. Instead, differences were observed in the functioning of all three forms of parent social capital, which were more strongly associated with both college-aligned actions and level of enrollment among White youth. These results are consistent with Tienda’s (2011) hypothesis that differences in the operation of college-relevant forms of parent social capital among Hispanic and White youth might help explain previously documented differences in the operation of parent income and education. Finally, structural invariance testing suggested that a similar level of college-aligned actions offered a greater enrollment boost among White students relative to similar Hispanic peers (Χ2(1)=3.88, p=.05).


Table 6. Differences Between Hispanic and White Youth in the Magnitudes of the Associations Between Parent Social Capital and Other Resources and College Enrollment

 

 Via College-Aligned

Actions

College

Enrollment

 

df

Χ2

df

Wald∆

Parent Social Capital

    

College-Relevant SSC

1

9.99**

3

15.11**

College-Relevant FSC

1

5.05*

3

9.22*

Intergenerational Closure

1

4.46*

3

6.77

Other Parent Resources

    

Income

1

0.21

3

5.15

BA or above

1

1.13

3

4.50

Note: Hispanic Sample n=1,020; White Sample n=5,420.

a In Mplus the “MODEL TEST” command is used to evaluate parameter constraints for indirect effects or total effects when the model includes a mediator; the Wald test statistic is provided when using the “MODEL TEST” command to evaluate total effects.

* p≤.05, ** p≤.01


LIMITATIONS

This study contributes to knowledge about differences in the process by which parent social capital and other parent resources are associated with college enrollment among bachelor’s-expecting Hispanic and White youth. However, it is important to remember that this research excludes students who did not expect to complete a bachelor’s degree in their senior year of high school. Reflecting a critical quantitative approach (Stage, 2007), limiting the sample to bachelor’s-expecting youth facilitated a more careful examination of the experiences of a distinct subset of students. However, conclusions drawn from this research may not be applicable to Hispanic and White youth who leave high school with other plans. Supporting the notion that high school seniors with different educational goals differ in how they experience postsecondary preparation and enrollment (Wells et al., 2013), separate analyses revealed that the hypothesized model tested here proved to be a poor fit with the experiences of students who did not finish high school with bachelor’s degree expectations. Further, while 34% and 44%, respectively, of Hispanic and White students in the research sample were enrolled in a 4-year institution in 2006, only 6% and 9% of their peers with non-bachelor’s degree plans were similarly enrolled.


In addition, while this study considered college preparation and enrollment among Hispanic and White students generally, this approach is not intended as a dismissal of the diversity within the U.S. Hispanic population (for example, according to factors such as national origin and immigrant generation). Prior research demonstrates that national origin and immigrant generation status appear to help explain differences in educational processes and outcomes among Hispanic high school-age youth (for example, Conchas, Oseguera, & Vigil, 2012; Nuñez & Crisp, 2012; Ryan & Ream, 2016). However, recent research indicates that these factors do little to explain differences between Hispanic and White youth when it comes to college enrollment, above and beyond differences in socioeconomic background (Morgan & Gelbgiser, 2014; Ovink & Kalogrides, 2015). Similarly, while indicators of both national origin and immigrant generation were included in preliminary analyses, neither was a significant predictor or helped to explain differences observed between bachelor’s-expecting Hispanic and White youth in this research. As others have noted (E. Kim & Díaz, 2013; Pérez & Ceja, 2015; Ryan & Ream, 2016), researchers should continue seeking to better understand how, and under what circumstances, diversity within the U.S. Hispanic population contributes to educational processes and outcomes—ideally in conjunction with the development of datasets specifically designed to facilitate research of this nature.


Finally, the intent of this study was not to account for all possible influences on college enrollment, nor was it designed to isolate causal effects. Instead, this study was specifically designed to test whether group-level variability in the utility of parent social capital might help explain the finding that parent income and education do not operate the same way among Hispanic and White youth when it comes to 4-year college enrollment. The findings speak directly to previous research (Alon et al., 2010; O’Connor et al., 2010) while highlighting directions for future research. For instance, there is growing recognition of the ways in which school and geographical contexts may contribute to divergent patterns of college entry and persistence among Hispanic and White youth (G. F. Martinez & Deil-Amen, 2015). A key finding here was that college-relevant forms of parent social capital were more readily converted into college-aligned actions and college enrollment among bachelor’s-expecting White youth relative to similar Hispanic youth. In response, future research might examine whether and how school and community contexts differentially shape access to social networks in which college-relevant information circulates, even among Hispanic and White parents with similar levels of socioeconomic resources. Especially relevant to inquiry of this nature, Stanton-Salazar’s (2001) critical network analytic perspective provides a useful framework for investigating the mobilization of resources for instrumental purposes against the backdrop of institutional power relations and the unequal distribution of access to socially valued resources.


DISCUSSION

Gaps between Hispanic and White young adults in 4-year college enrollment and degree completion remain, even as rates of Hispanic postsecondary participation have increased (Fry & Taylor, 2013). This pattern has been referred to as the “Hispanic college puzzle” (Alon et al., 2010). Alon et al. documented one piece of this puzzle: namely, that Hispanic parents do not transmit their educational advantages to their children at the same rate as similarly resourced White parents. Tienda (2011) subsequently surmised that rates of intergenerational resource transmission from Hispanic parents to their children might reflect reduced access to forms of social capital that facilitate the exchange of college-relevant information and support. Addressing the need for research about how the college preparation and enrollment process unfolds in particular ways for Hispanic students relative to other groups (Pérez & Ceja, 2015), this study explored Tienda’s hypothesis.


The results suggest that differences in the intergenerational transmission of parents’ educational and economic resources among bachelor’s-expecting Hispanic and White youth can be partly attributed to differences in the functioning of parent social capital. This finding is noteworthy given that few studies have examined how the educational utility of network-bound parental resources varies across racial/ethnic groups (Ream, 2005; Ream & Palardy, 2008). The results also confirm that when students with plans to complete a 4-year degree take steps during high school that are commensurate with their goals, they are much more likely to be enrolled in a 4-year institution 2 years beyond graduation. However, the relationship between aligned actions and college enrollment appears to be stronger among White youth. These findings warrant further discussion.


This research sheds light on one potential reason behind earlier evidence (Alon et al., 2010; O’Connor et al., 2010) suggesting that parents’ educational and financial resources are not as readily translated into 4-year college enrollment among Hispanic youth relative to White youth. Both the exchange of college-related information between parents and children (FSC) and the exchange of information and support among parents at the school (intergenerational closure) were more strongly associated with college-aligned actions among White youth. Accounting for these differential returns to parent social capital across the two groups, parent income and education were similarly associated with college enrollment among Hispanic and White youth. These results suggest that the lower rates of intergenerational transmission of income and education among Hispanic parents and children observed in previous research (Alon et al., 2010; O’Connor et al., 2010) can be at least partly attributed to differences in the utility of parents’ social resources. This finding is consistent with Bourdieu’s (1986) assertion that social capital acts as a catalyst, facilitating the conversion of more tangible resources into advantageous outcomes. Hispanic families draw upon numerous strengths to support the educational trajectories of children (Auerbach, 2004; Delgado-Gaitan, 1992; Yosso, 2005), effectively creating a “culture of possibility” as youth contemplate educational goals and options (M. A. Martinez, 2013). However, when parents of Hispanic youth do not have access to the same kinds of information about navigating the higher education system through their relationships with others as do the parents of White youth, it may be more difficult for them to fully leverage other assets as their children prepare for college.


An unanticipated result with respect to parent social capital was the negative association between college-relevant SSC and aligned actions among White youth. One plausible interpretation is that Hispanic and White parents sometimes have different motivations for contacting the school about their child’s academic progress and future plans (Auerbach, 2004). Hispanic parents may reach out to school personnel in pursuit of information about their child’s education that they find difficult to obtain through more informal channels. On the other hand, parents of many White youth may have access to this information through other sources and might instead approach school agents to advocate for their child when, for example, he or she is not demonstrating the expected level of preparation for college or is not performing well academically (Auerbach, 2004). The fact that the association between SSC and college-aligned actions was positive among Hispanic youth aligns with the observed need for programs and policies that leverage a deep commitment to education and the desire for information specifically among families of color (Tierney & Auerbach, 2005).


In accordance with the theory of aligned ambitions (Schneider & Stevenson, 1999), this research also demonstrates the importance of enabling students to understand and complete the actions that correspond to their specific educational goals. Hispanic students, on average, are more likely than other groups to enter higher education through a community college despite the fact that 2-year colleges may not provide effective transfer mechanisms to 4-year institutions (Arbona & Nora, 2007; Fry & Taylor, 2013; Pérez & McDonough, 2008). When bachelor’s-expecting Hispanic youth engage in actions that are aligned with their educational goals, they substantially reduce their probability of following a college pathway that is misaligned with their degree expectations. For the average Hispanic student, a standard deviation unit increase in aligned actions shifted the probability of being enrolled in a 4-year institution in 2006 from 27% to 76% (see Table 5).


An analysis of student scores on the latent aligned actions construct revealed that, on average, among Hispanic students with a level of aligned actions falling between the mean and one standard deviation above the mean, 99% had taken an entrance exam, 56% had completed a math course higher than Algebra II, 38% had applied to two or more 4-year institutions, and their average cumulative high school GPA was between 2.5 and 3.0. Among Hispanic students more than one standard deviation above the aligned actions mean, all had taken an entrance exam, 92% had completed a math course higher than Algebra II, 66% had applied to two or more 4-year institutions, and their average cumulative high school GPA was between 3.0 and 3.5.


Increasing the extent to which bachelor’s-expecting Hispanic youth match their expectations with adequate preparation has the potential to improve their 4-year college-going rates. However, the fact that a similar level of college-aligned actions was less predictive of enrollment in a 4-year institution among Hispanic youth relative to similar White students may partly reflect other factors that can lead students to begin the pursuit of a bachelor’s degree at a 2-year institution. For instance, Hispanic youth tend to be concentrated in states with large and comprehensive community college systems (O’Connor et al., 2010), they are more likely to say that living at home while attending college is important (Desmond & Turley, 2009), and their college goals and decisions may be more deeply affected by perceptions of college costs and financial aid than is the case among White students (Gándara & Contreras, 2009). Further, Hispanic youth may more often be advised by teachers and family members to start at the community college level in order to improve their grades and save money (González, Stoner, & Jovel, 2003; Pérez & McDonough, 2008). Future research might investigate potential sources of group-level differences in the advantages that accrue to students with well-aligned postsecondary expectations and preparation.


DIRECTIONS FOR POLICY AND PRACTICE

It has been asserted that the single most important obstacle to 4-year college access among Hispanic students is a lack of instrumental knowledge about the sequence of actions necessary to go on to college (Gándara, 2002). When schools are not prepared to guide students on the path to college, the task may fall even more squarely on the shoulders of parents. This fact may have unique significance for Hispanic parents, as their social and economic realities often make it especially difficult to access concrete college information (Ceja, 2006). Relative to similar White parents, the information and support that Hispanic parents were able to secure from other school parents and to share with their children constituted less useful forms of social capital with respect to children’s college preparation and enrollment. This finding lends support to the notion that Hispanic parents and youth have less access to information about the college choice process via their informal social networks (Gándara & Contreras, 2009; Pérez & McDonough, 2008; Rosenbaum & Person, 2006). The implementation of policies and practices described in this section may help facilitate access to college knowledge for Hispanic families and may also promote access to more equitable preparation for college among Hispanic students.


Research has suggested that parent outreach around college at many high schools, which often consists primarily of college and financial aid nights and newsletters, tends to draw parents from higher socioeconomic backgrounds and parents whose students have higher achievement levels (Rowan-Kenyon, Bell, & Perna, 2008). School personnel tend to have clear expectations for parent involvement in college preparation, particularly in terms of attendance at school events. Yet many Hispanic parents are unable to meet these expectations given work obligations, unease with school staff, language differences, and a lack of trust (Rowan-Kenyon et al., 2008).


Certainly, as a central institution in the lives of youth and their families, schools can serve to reproduce existing status hierarchies that continue to locate power and control over valued resources with certain, often White and middle- or upper-class, segments of society. But equally as certainly, schools and the communities in which they are embedded do not have to operate this way. Schools can also serve as sites where historically marginalized individuals and groups are able to access institutional resources through empowering relationships with school agents—in other words, as sites where counterstratification forces flourish (Stanton-Salazar, 2001). Counterstratification forces in ethnic and kinship networks prevail when grounded in natural support systems characterized by a help-seeking orientation, or the inclination to resolve challenges through the mobilization of resources and relationships (Stanton-Salazar, 2001).


Said another way, while schools can function as a stratifying structure within society, schools can also play an integral role in facilitating the type of support systems Stanton-Salazar describes. This includes networks through which Hispanic students and parents can mobilize resources and relationships via institutional agents who often serve as gatekeepers to college knowledge (Stanton-Salazar, 2001). For instance, instead of viewing college preparation as an individual student activity with an “add-on” parent component (Yamamura, Martinez, & Saenz, 2010), school-based college readiness efforts can consciously seek to create a college-going culture characterized by relational trust among students, parents, teachers, and counselors (Schneider et al., 2014). Further, high schools and area colleges can work with community-based nonprofit organizations to bring college information to parents where they are instead of expecting parents to surmount the often substantial cultural, linguistic, and logistical barriers required to attend college and financial aid informational events at their child’s school (Rowan-Kenyon et al., 2008). Researchers, for their part, might seek to better understand how particular aspects of school contexts or college readiness interventions do (or do not) foster access to resources and relationships not only among students, but also among their families. This type of research seems especially important given that college-going among Latino students has been asserted as a family process as opposed to an individual one (Alvarez, 2015).


In addition to policies and practices that can address the need for equitable access to information about college among Hispanic parents and youth, students also deserve access to strong preparation for college. Preparation must begin early on and continue through the high school years, requiring a focus on students’ experiences in both elementary and secondary education. The cumulative effects of inadequate preparation, beginning with a lack of access to high-quality preschool programs and continuing through high school, forcefully impact rates of college enrollment and attainment among Hispanic students (Sólorzano, Villalpando, & Oseguera, 2005). Comparatively lower outcomes among Hispanic students as measured by state and national standardized tests scores, high school graduation rates, and rates of transition to college all directly reflect early disparities in achievement and school readiness (Gándara & Contreras, 2009). Research evidence also suggests that many Hispanic students may find themselves ineligible for admissions to a 4-year college because they have not been placed into the requisite college preparatory courses they need in order to be admitted (Calderón, Slavin, & Sánchez, 2011; De La Rosa & Tierney, 2006; González et al., 2003).


In response to calls for more equitable college preparation among Hispanic youth, it has been argued that one of the most critical steps is to improve college counseling, particularly in the underresourced, high-minority, and high-poverty schools attended by a substantial proportion of Hispanic youth (Contreras, 2005; Gándara & Contreras, 2009). Unquestionably, counselors and other institutional agents can play a critical role in leveling access to college knowledge (Robinson & Roksa, 2016; Stanton-Salazar, 2001). At the same time, however, effective college-promoting practices must be embedded in a school and district context in which educators, students, and parents, through a system of shared social relationships and values, establish and consistently reinforce norms for college attendance across a student’s pre-k–12 educational experience (Schneider, 2007). Thus, as previously asserted, future research might investigate how aspects of school and community contexts structure student and parent access to the norms, relationships, and resources necessary for students to successfully align college preparation with college goals. Such investigation should include a focus on how contextual factors may differentially support or deter Hispanic and White youth and families.


CONCLUSION

A limited view of the assets that Hispanic youth and families bring to the college choice process has too often led researchers and policymakers to conclude that Hispanic students are less likely than their White peers to enroll in a 4-year college because they tend to come from poorer and less well-educated families (Kao et al., 2013; Rodriguez & Morrobel, 2004). In contrast, recent studies suggest that even when parents of Hispanic and White youth have similar levels of income and education, these resources do not operate the same way as their children navigate college preparation and enrollment. Tienda (2011) hypothesized lower returns to college-relevant forms of parent social capital among Hispanic parents as one potential explanation. In response, the current study empirically investigated differences in the process by which parent resources appear to influence college enrollment among Hispanic and White youth by focusing on the role of parent social capital.


The finding that lower rates of intergenerational resource transmission among Hispanic parents and children (Alon et al., 2010; O’Connor et al., 2010) can be at least partly attributed to differences in the utility of parents’ social resources reflects Bourdieu’s (1986) assertion that social capital facilitates the conversion of more tangible resources into advantageous outcomes. When parents of Hispanic youth do not have access to the same kinds of information about navigating postsecondary transitions as do parents of White youth, fully leveraging their many other assets to support their children’s college preparation and enrollment will be more difficult.


Policymakers, practitioners and researchers must continue working to ensure that college preparation programs and practices are meeting the specific needs of Hispanic families. Findings from this study suggest the need to better connect Hispanic parents to the college knowledge and support that can help them leverage their numerous strengths and assets to the benefit of their bachelor’s-seeking children. At the same time, there remains a need for a more complete understanding of factors, including those beyond the student and the family, that help explain how and why processes of preparing for and enrolling in college among Hispanic and White youth differ—particularly when those differences serve to further existing inequalities.


References


Adelman, C. (2002). The relationship between urbanicity and educational outcomes. In W. G. Tierney & L. S. Hagedorn (Eds.), Increasing access to college: Extending possibilities for all students (pp. 35–64). Albany: State University of New York Press.


Alon, S., Domina, T., & Tienda, M. (2010). Stymied mobility or temporary lull? The puzzle of lagging Hispanic college degree attainment. Social Forces, 88(4), 1807–1832.


Alvarez, C. (2015). A model for understanding the Latina/o student and parent college-going navigation process. In P. Pérez & M. Ceja (Eds.), Higher education access and choice for Latino students (pp. 55–68). New York, NY: Routledge.


Arbona, C., & Nora, A. (2007). The influence of academic and environmental factors on Hispanic college degree attainment. Review of Higher Education, 30(3), 247–269.


Auerbach, S. (2004). Engaging Latino parents in supporting college pathways: Lessons from a college access program. Journal of Hispanic Higher Education, 3(2), 125–145.


Avery, C., & Kane, T. J. (2004). Student perceptions of college opportunities: The Boston COACH program. In C. Hoxby (Ed.), College choices: The economics of where to go, when to go, and how to pay for it (pp. 355–394). Chicago, IL: University of Chicago Press.


Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51(6), 1173–1182.


Berkner, L., & Chavez, L. (1997). Access to postsecondary education for the 1992 high school graduates. Washington, DC: National Center for Education Statistics.


Bohon, S., Johnson, J., & Gorman, B. (2006). College aspirations and expectations among Latino adolescents in the United States. Social Problems, 53, 207–225.


Bourdieu, P. (1986). The forms of capital. In J. G. Richardson (Ed.), Handbook of theory and research for the sociology of education (pp. 241–258). New York, NY: Greenwood Press.


Bozick, R., Alexander, K., Entwistle, D., Dauber, S., & Kerr, K. (2010). Framing the future: Revisiting the place of educational expectations in status attainment. Social Forces, 88(5), 2027–2052.


Cabrera, A. F., & LaNasa, S. M. (2001). On the path to college: Three critical tasks facing America’s disadvantaged. Research in Higher Education, 42, 119–150.


Calderón, M., Slavin, R., & Sánchez, M. (2011). Effective instruction for English learners. Future of Children, 21(1), 103–127.


Carbonaro, W. J. (1998). A little help from my friend’s parents: Intergenerational closure and educational outcomes. Sociology of Education, 71, 295–313.


Carnevale, A. P., & Strohl, J. (2014). How increasing college access is increasing inequality and what to do about it. In R. D. Kahlenberg (Ed.), Rewarding strivers: Helping low-income students succeed in college. New York, NY: Century Foundation Press.


Ceja, M. (2006). Understanding the role of parents and siblings as information sources in the college choice process of Chicana students. Journal of College Student Development, 47(1), 87–104.


Choy, S. (2001). Students whose parents did not go to college: Postsecondary access, persistence, and attainment. Washington, DC: National Center for Education Statistics.


Coleman, J. S. (1988). Social capital in the creation of human capital. American Journal of Sociology, 94, 95–120.


Coleman, J. S. (1990). Foundations of social theory. Cambridge, MA: Harvard University Press.


Coleman, J. S., & Hoffer, T. (1987). Public and private high schools: The impact of communities. New York, NY: Basic Books.


Conchas, G. Q., Oseguera, L., & Vigil, J. D. (2012). Acculturation and school success: Understanding the variability of Mexican American youth adaptation across urban and suburban contexts. Urban Review, 44(4), 401–422.


Contreras, F. (2005). Access, achievement, and social capital: Standardized exams and the Latino college-bound population. Journal of Hispanic Higher Education, 4, 197–214.


De La Rosa, M., & Tierney, W. G. (2006). Breaking through the barriers to college. Los Angeles, CA: Center for Higher Education Policy Analysis.


Deil-Amen, R., & Turley, R. N.L. (2007). A review of the transition to college literature in sociology. Teachers College Record, 109(10), 2324–2366.


Delgado-Gaitan, C. (1992). School matters in the Mexican-American home: Socializing children to education. American Educational Research Journal, 29, 495–513.


Desmond, M., & Turley, R. N. L. (2009). The role of familism in explaining the Hispanic-White college application gap. Social Problems, 56(2), 311–334.


Dika, S. L., & Singh, K. (2002). Applications of social capital in educational literature: A critical synthesis. Review of Educational Research, 72, 31–60.


Fry, R., & Taylor, P. (2013). Hispanic high school graduates pass whites in rate of college enrollment. Washington, DC: Pew Hispanic Center.


Gándara, P. (2002). A study of high school puente: What we have learned about preparing Latino youth for postsecondary education. Educational Policy, 16(4), 474–495.


Gándara, P., & Contreras, F. (2009). The Latino education crisis: The consequences of failed social policies. Cambridge, MA: Harvard University Press.


Gándara, P., Oseguera, L., Huber, L. P., Locks, A., Ee, J., & Molina, D. (2013, December). Making education work for Latinas in the U.S. Los Angeles, CA: The Civil Rights Project at the University of California Los Angeles.


González, K. P., Stoner, C., & Jovel, J. E. (2003). Examining the role of social capital in access to college for Latinas. Journal of Hispanic Higher Education, 2, 146–170.


Grodsky, E., & Riegle-Crumb, C. (2010). Those who choose and those who don’t: Social background and college orientation. ANNALS of the American Academy of Political and Social Science, 627, 14–35.


Hart, B., & Risley, T. R. (1995). Meaningful differences in the everyday experience of young American children. Baltimore, MD: Brookes Publishing.


Horn, L. J., & Nuñez, A. M. (2000, March). Mapping the road to college: First-generation students’ math track, planning strategies, and context of support (NCES 2000–153). Washington, DC: National Center for Education Statistics.


Horvat, E. M., Weininger, E. G., & Lareau, A. (2003). From social ties to social capital: Class differences in the relations between schools and parent networks. American Educational Research Journal, 40(2), 319–351.


Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6(1), 1–55.


Hurtado, S., Inkelas, K. K., Briggs, C., & Rhee, S. (1997). Differences in college access and choice among racial/ethnic groups: Identifying continuing barriers. Research in Higher Education, 38(1), 43–64.


Ingels, S. J., Pratt, D. J., Wilson, D., Burns, L. J., Currivan, D., Rogers, J. E., & Hubbard-Bednasz, S. (2007). Education longitudinal study of 2002: Base-year to second follow-up data file documentation. Washington, DC: National Center for Education Statistics.


Kao, G., & Rutherford, L. T. (2007). Does social capital still matter? Immigrant minority disadvantage in school-specific social capital and its effects on academic achievement. Sociological Perspectives, 50(1), 27–52.


Kao, G., & Tienda, M. (1998). Educational aspirations of minority youth. American Journal of Education, 106, 349–384.


Kao, G., Vaquera, E., & Goyette, K. (2013). Education and immigration. Malden, MA: Polity.


Kerckhoff, A. C. and Campbell, R. T. (1977). Race and social status differences in the explanation of educational ambition. Social Forces, 55, 701–714.


Kim, E., & Díaz, J. (Eds.). (2013). Immigrant students and higher education [Special issue]. ASHE Higher Education Report, 38(6).


Kim, D. H., & Schneider, B. (2005). Social capital in action: Alignment of parental support in adolescents’ transition to postsecondary education. Social Forces, 84(2), 1181–1206.


Kiyama, J. (2011). Family lessons and funds of knowledge: College going paths in Mexican American Families. Journal of Latinos and Education, 10(1), 23–42.


Klasik, D. (2012). The college application gauntlet: A systematic analysis of the steps to four-year college enrollment. Research in Higher Education, 53(5), 506–549.


Lin, N. (2001). Building a network theory of social capital. In N. Lin, K. Cook, & R. S. Burt (Eds.), Social capital: Theory and research (pp. 3–29). New York, NY: Aldine De Gruyter.


Mackinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002). A comparison of methods to test the significance of the mediated effect. Psychological Methods, 7, 83–104.


Marsico, M., & Getch, Y. Q. (2009). Transitioning Hispanic seniors from high school to college. Professional School Counseling, 12(6), 458–462.


Martinez, G. F., & Deil-Amen, R. (2015). College for all Latinos? The role of high school messages in facing college challenges. Teachers College Record, 117, 1–50.


Martinez, M. A. (2013). (Re)considering the role familismo plays in Latina/o high school students’ college choices. High School Journal, 97(1), 21–40.


McIntosh, C. (2007). Rethinking fit assessment in structural equation modeling: A commentary and elaboration on Barrett (2007). Personality and Individual Differences, 42(5), 859–67.


Mickelson, R. A. (1990). The attitude-achievement paradox among black adolescents. Sociology of Education, 63(1), 44–61.


Morgan, S. L. (2005). On the edge of commitment: Educational attainment and race in the United States. Palo Alto, CA: Stanford University Press.


Morgan, S. L., & Gelbgiser, D. (2014). Mexican ancestry, immigrant generation, and educational attainment in the United States. Sociological Science, 1, 397–422


Muthén, B. O., & Asparouhov, T. (2002). Latent variable analysis with categorical outcomes: Multiple-group and growth modeling in Mplus (Web Note No. 4). Los Angeles, CA: Mplus.


Muthén, L. K., & Muthén, B. O. (2015). Mplus User’s Guide (7th ed.). Los Angeles, CA: Muthén & Muthén.


Nora, A., & Crisp, G. (2009). Hispanics and higher education: An overview of research, theory, and practice. In J. C. Smart (Ed.), Higher education: Handbook of theory and research (Vol. 24, pp. 321–358). New York, NY: Springer Publishing.


Nuñez, A. M., & Crisp, G. (2012). Ethnic diversity and Latino/a college access: A comparison of Mexican American and Puerto Rican beginning college students. Journal of Diversity in Higher Education, 5(2), 78–95.


Nuñez, A. M., & Kim, D. (2012). Building a multicontextual model of Latino college enrollment: Student, school, and state-level effects. Review of Higher Education, 35(2), 237–263.


O’Connor, N., Hammack, F. M., & Scott, M. A. (2010). Social capital, financial knowledge, and Hispanic student college choices. Research in Higher Education, 51(1), 195–219.


Ovink, S. M., & Kalogrides, D. (2015). No place like home? Familism and Latino/a-white differences in college pathways. Social Science Research, 52, 219–235.


Park, S., Wells, R., & Bills, D. (2015). Changes in educational expectations between 10th and 12th grades across cohorts. Social Psychology of Education, 18(3), 561–581.


Pérez, P. A., & Ceja, M. (Eds.). (2015). Higher education access and choice for Latino students: Critical findings and theoretical perspectives. New York, NY: Routledge.


Pérez, P. A., & McDonough, P. M. (2008). Understanding Latina and Latino college choice: A social capital and chain migration analysis. Journal of Hispanic Higher Education, 7, 249–265.


Perna, L. W., & Titus, M. A. (2005). The relationship between parental involvement as social capital and college enrollment: An examination of racial/ethnic group differences. Journal of Higher Education, 76(5), 486–518.


Portes, A. (1998). Social capital: Its origins and applications in modern sociology. Annual Review of Sociology, 24, 1–24.


Ream, R. K. (2005). Uprooting children: Mobility, social capital, and Mexican American underachievement. New York, NY: LFB Scholarly Publishing, LLC.


Ream, R. K., & Palardy, G. P. (2008). Student engagement, peer social capital, and school dropout among Mexican American and non-Latino white students. Sociology of Education, 81(2), 109–139.


Riegle-Crumb, C. (2010). More girls go to college: Exploring the social and academic factors behind female postsecondary advantage among Hispanic and White students. Research in Higher Education, 51, 573–593.


Robinson, K. J., & Roksa, J. (2016). Counselors, information, and high school college-going culture: Inequalities in the college application process. Research in Higher Education, 57(7), 845–868.


Roderick, M., Coca, V., & Nagaoka, J. (2011). Potholes on the road to college: High school effects in shaping urban students’ participation in college application, four year college enrollment, and college match. Sociology of Education, 84(3), 178–211.


Rodriguez, M., & Morrobel, D. (2004). A review of Latino youth development research and a call for an asset orientation. Hispanic Journal of Behavioral Sciences, 26, 107–127.


Rosenbaum, J. E., & Person, A. (2006). Chain enrollment and college enclaves: Benefits and drawbacks for Latino college students. In G. Orfield, C. L. Horn, & S. Flores (Eds.), New directions for community colleges: Community colleges and Latino educational opportunity (pp. 51–60). San Francisco, CA: Jossey-Bass.


Rowan-Kenyon, H. T., Bell, A., & Perna, L. W. (2008). Contextual influences on parental involvement and college-going: Variations by socioeconomic class. Journal of Higher Education, 79(5), 564–586.


Ryan, S. (2016). Investigating college enrollment among Hispanic youth who expect to complete a bachelor’s degree. Journal of Hispanic Higher Education, 15(1), 3–21.


Ryan, S., & Ream, R. K. (2016). Variation across Hispanic immigrant generations in parent social capital, college-aligned actions, and four-year college enrollment. American Educational Research Journal, 53(4), 953–986.


Schneider, B. (2007). Forming a college-going community in U.S. public high schools. Lansing, MI: Michigan State University.


Schneider, B., Judy, J., Ebmeyer, C. M., & Broda, M. (2014). Trust in elementary and secondary urban schools: A pathway for student success and college ambition. In D. Van Maele, P. B. Forsyth, & M. Van Houtte (Eds.), Trust and school life (pp. 37–56). New York, NY: Springer.


Schneider, B., & Stevenson, D. (1999). The ambitious generation: America’s teenagers, motivated but directionless. New Haven, CT: Yale University Press.


Sewell, W. H., Hauser, R. M., Springer, K. W., & Hauser, T. A (2003). As we age: A review of the Wisconsin longitudinal study, 1957–2001. Research in Social Stratification and Mobility, 20, 3–111.


Sólorzano, D. G., Villalpando, O., & Oseguera, L. (2005). Educational inequities and Latina/o undergraduate students in the United States: A critical race analysis of their educational progress. Journal of Hispanic Higher Education, 4(3), 272–294.


Stage, F. K. (2007). Answering critical questions using quantitative data. New Directions for Institutional Research, 2007(133), 5–16.


Stanton-Salazar, R. D. (2001). Manufacturing hope and despair: The school and kin support networks of US-Mexican youth. New York, NY: Teachers College Press.


Stanton-Salazar, R. D., & Dornbusch, S. (1995). Social capital and the reproduction of inequality: Information networks among Mexican-origin high-school students. Sociology of Education, 68, 116–135.


Tienda, M. (2011). Hispanics and U.S. schools: Problems, puzzles and possibilities. In M. T. Hallinan (Ed.), Frontiers in Sociology of Education (pp. 303–310). New York, NY: Springer.


Tierney, W. G., & Auerbach, S. (2005). Toward developing an untapped resource: The role of families in college preparation. In W. G. Tierney, Z. B. Corwin, & J. E. Colyar (Eds.), Preparing for college: Nine elements of effective outreach (pp. 29–48). Albany: State University of New York Press.


Tornatzky, L. G., Cutler, R., & Lee, J. (2002). College knowledge: What Latino parents need to know and why they don’t know it. Los Angeles, CA: Tomás Rivera Policy Institute.


Wells, R., Seifert, T., & Saunders, D. (2013). Gender and realized educational expectations: The roles of social origins and significant others. Research in Higher Education, 54(6), 599–626.


Welton, A. D., & Martinez, M. A. (2014). Coloring the college pathway: A more culturally responsive approach to college readiness and access for students of color in secondary schools. Urban Review, 46(2), 197–223.


Yamamura, E. K., Martinez, M. A., & Saenz, V. B. (2010). Moving beyond high school expectations: Examining stakeholders’ responsibility for increasing Latina/o students’ college readiness. High School Journal, 93(3), 126–148.


Yosso, T. J. (2005). Whose culture has capital? A critical race theory discussion of community cultural wealth. Race Ethnicity and Education, 8(1), 69–91.



APPENDIX


MODEL EQUATIONS


The structural equation model can be expressed for typical individual i as:


Measurement model:

Zi * = νz + Λ zηi + ΚzΧi + εzi

(1)


Structural model:

Yi * = νy + Βyηi + ΚyΧi + εyi

(2)


ηi = α + Βηi + ΓΧi + ζi

(3)


Zi* is a (p x 1) vector of dependent variables (latent variables underlying observed indicators)

Yi* is the dependent variable (enrollment in 2006)

ηi (eta) is a (m x 1) vector of latent factors

νz and νy (nu) and α (alpha) are vectors of intercepts

Λz (Lambda) is a (p x m) matrix of latent factor loadings

Κz and Κy (Kappa) and Γ (Gamma) contain slopes for exogenous covariates in the (q x 1) vector Χi (Chi)

Βy (Beta) is a (1 x m) matrix of structural regression slopes

Β (Beta) is a (m x m) matrix of structural regression slopes

εi (epsilon) and ζi (zeta) are vectors of error terms

The conditional probability of being enrolled at level q for student i can be represented by the following equation:

P(y=0 ηiy, xi) = F[(τ1 - βyηi - κjxi)(1/√θ)]

(4)


P(y=1 ηiy, xi) = F[(τ2 - βyηi - κjxi)(1/√θ)] - F[(τ1 - βyηi - κjxi)(1/√θ)]

(5)


P(y=2 ηiy, xi) = F[(-τ2+ βyηi + κjxi)(1/√θ)]

(6)


where η is a vector of latent factor means for individual i with the observed outcome variable y and x is a vector of observed covariates for individual i. The cumulative distribution function of the standard normal distribution ɸ is represented by F, and τ defines a threshold for the observed latent response variable y* that underlies the observed y such that that y = 0 is observed when y* < τ1, the threshold separating the y = 0 from y = 1, y=1 when τ1 y*≤ τ2 and y=2 when τ1 y*. The vector of probit regression coefficients for the regression of y* on the latent factors for individual i is represented by β and the vector of probit regression coefficients for the regression of y on the observed covariates is represented by κ. Finally, θ is residual variance of y*.


MULTIPLE GROUP INVARIANCE TESTING


When working with categorical outcomes using the delta parameterization, the procedure for assessing group-level measurement invariance in Mplus proceeds through the following two steps: 1) In Model 1, thresholds and factor loadings are freed across groups, with scale factors fixed at 1 across groups and factor means fixed at 0 across groups, and; 2) In Model 2, thresholds and factor loadings are constrained to be equal across groups, scale factors are fixed at 1 in one group and freed in the other group(s), and factor means are fixed at 0 in one group and freed in the other group(s).


I began by running the confirmatory four-factor model as specified in Step 1 above. This model showed a good fit to the data (χ2 (157)=354.19, CFI=.99, RMSEA=.02). The second model also fit the data well (data (χ2(187)=427.99, CFI=.99, RMSEA=.02); however, the chi-square difference test revealed a significant difference between the two models (χ2∆(30)=94.27, p=.00). Modification indices indicated that constraining to equality the thresholds for two of the four indicators of the aligned actions construct—cumulative high school GPA and number of non-open enrollment colleges applied to—across the two groups was too restrictive to establish measurement invariance. To address this, first the thresholds for the number of non-open enrollment schools applied to were freed. This model demonstrated a good fit to the data (χ2(184)=390.21, CFI=.99, RMSEA=.02) and resulted in a reduction of the chi-square difference statistic (χ2∆ (27)=47.78, p=.01); however, the difference between this model and the first model remained significant at the p≤.05 level. Next, the thresholds for cumulative high school GPA were freed. This model fit the data well (χ2(178)=373.71, CFI=.99, RMSEA=.02) and the chi-square difference test revealed that the two models no longer differed significantly (χ2∆(21)=27.15, p=.17). B. O. Muthén and Asparouhov (2002) assert that in order to make meaningful comparisons of factor distributions across groups, a majority of the variables serving as latent factor indicators should have both threshold and loading invariance. Given that it was only necessary to free the thresholds for two of the four indicators of one factor to achieve measurement invariance of the four latent factors, the criteria suggested by B. O. Muthén and Asparouhov were satisfied.


With an adequate level of measurement invariance established, I next examined structural invariance and, more specifically, whether the magnitude of parameter estimates including regression coefficients and latent factor means varied across ethnic groups using the “MODEL TEST” command in Mplus. I began with a baseline model in which factor loadings and thresholds were constrained to be equal across groups while intercepts and scale factors were held at 0 and 1, respectively, in the first group and were freely estimated in the other group. This model, which fit the data well (χ2(637) =928.93, CFI=.99, RMSEA=.02), was then used as a baseline against which to compare each subsequent model in which one parameter was freed at a time across groups (see Table 6 for results).




Cite This Article as: Teachers College Record Volume 119 Number 10, 2017, p. 1-39
https://www.tcrecord.org ID Number: 21849, Date Accessed: 10/17/2021 2:37:47 PM

Purchase Reprint Rights for this article or review