Predictive factors in postsecondary educational attainment among Latinos.

The study in this article investigated factors that distinguish the increasing number of Latino students who continue their education beyond high school from the small and stable number who complete a baccalaureate degree. The sample included a cohort of 866 Latino men and women who participated in the National Educational Longitudinal Study (1988-2000) and had varying amounts of postsecondary education by the year 2000. The study employed educational, psychological, and familial predictor variables from 1990 when participants were sophomores in high school. Data were analyzed using a multinomial logistic regression and results indicated that several factors were significant in distinguishing those with some postsecondary education, no degree, from bachelor's completers. Parent support and locus of control were the two most significant predictors. Implications for school counselors are discussed.

**********

In July 2002, the U.S. Census Bureau announced that Latinos had become the nation's largest non-dominant group with a population of 38.8 million, surpassing the African American population now numbered at 38.3 million (U.S. Bureau of the Census, 2002b). With a high birthrate, their numbers are expected to continue to grow rapidly. Latinos are overrepresented among the poor with about a third of families living below the poverty line. They are a relatively young group with 35% under the age of 18 compared to less than 25% of non-Latino Whites. Because of their low-income status and the high school-aged population, the education of Latinos has become a major concern in this country (Llagas & Snyder, 2003). This concern was manifested most notably in the creation on October 12, 2001, of the President's Advisory Commission on Educational Excellence for Latino Americans.

EDUCATIONAL ATTAINMENT AMONG LATINO AMERICANS

Educational attainment data for Latinos beyond high school reveal two very different scenarios. Large numbers of Latinos are enrolled in postsecondary education; yet the majority are either older than the 18-to-24-year-old cohort, enrolled in community colleges, or attending part-time (Fry, 2002). There is a substantial enrollment gap between Latinos and all other groups among 18-to-24-year olds--the traditional age group for college attendance and the cohort that reaps the greatest economic benefit from a college degree. Only 35% of Latino high school graduates in that age group are enrolled in college compared to 46% of Whites. Latinos are far more likely to be enrolled in 2-year colleges than any other group. About 40% of Latino 18-to-24-year-old college students attend 2-year institutions compared to about 25% of White and Black students in that age group. Latinos are more likely to be part-time students. Nearly 85% of White 18-to-24-year-old college students are enrolled fulltime compared to 75% of Latino students in that age group (Fry). Similar to high school graduation, college enrollment is more likely among U.S.-born Latinos, especially the second generation, yet their attainment of a baccalaureate degree still ranks far below that of Whites. Of every 100 Latino kindergarteners, 11 will obtain at least a bachelor's degree compared to 33 for every 100 White kindergarteners (U.S. Bureau of the Census, 2002a).

FACTORS IN LATINO EDUCATIONAL ATTAINMENT

Explanations of Latino educational achievement and attainment have focused on familial, societal, educational, and psychological factors. The level of family involvement has been shown to be a significant factor in the high school achievement of Latino students (Bamaca-Gomez & Plunkett, 2003; Quezada, Diaz, & Sanchez, 2003). Latino parents have consistently emphasized the importance of education in the lives of their children and cite education as an important reason for immigration to the United States (Behnke, Diversi, & Piercy, 2004). While a number of studies have examined parental influence and high school completion among Latinos, few studies have examined such influence on postsecondary educational attainment.

Latino academic underachievement also has been linked to social influences and lack of community support. Certain negative stereotypes and biases can play a role in lowering academic expectations for Latinos (Aviles, Guerrero, Howarth, & Thomas, 1999; McWhirter, 2004). The absence of positive role models within the community also has been linked to higher school dropout rates (Aviles et al.). Too few positive role models prevent Latino students from obtaining the necessary motivation and information on how, for example, to apply to college (Immerwahr, 2003). High delinquency rates in poor Latino neighborhoods also make it difficult for students and their families to rely on resources within their neighborhoods to further their education (Pabon, 1998). While socioeconomic status (SES) is presumed to be a significant factor in distinguishing those who attend 4-year colleges, few studies have examined this variable in conjunction with others to differentiate the levels of postsecondary educational attainment among Latinos.

Regarding educational factors, previous research has examined the high school curriculum effects upon postsecondary educational attainment. When Latino students follow a rigorous math curriculum and enroll in college prep courses, their reading and math scores improve to the degree that the achievement gap is significantly narrowed (Adelman, 1999; National Center for Educational Statistics, 1994; Trusty & Niles, 2003). This improvement is unrelated to previous educational background. In other words, when students are held to higher standards, their scores improve in spite of not necessarily having prior preparation to meet those standards. Gamoran and Hannigan (2000) found that all students who take algebra score higher, yet Latinos often are excluded from such courses.

In general, Latino high school students are less likely to be enrolled in a college preparatory track-only 47% compared to 54% of Whites (Gonzalez, 2002; National Center for Educational Statistics, 2004). Latino students are overrepresented in special education and very underrepresented in Advanced Placement (AP) courses. Of the students in 2000 who took AP exams, only 9% were Latino (President's Advisory Commission on Educational Excellence for Latino Americans, 2003). Besides following a less rigorous curriculum, the second major variable contributing to low achievement among Latinos is quality of instruction. Underqualified teachers (i.e., teachers who lack a major or minor in the field they are teaching) more often teach classes in high-poverty schools (Ingersoll, 2003). In schools with a high (i.e., top quartile) population of non-White students, 41% of math courses are taught by teachers lacking a major in the field compared to 21% of math courses in schools with a low (i.e., bottom quartile) population of non-White students (Ingersoll).

These data about curricula and quality of instruction raise questions about a school's support for Latino students going to college. Because many Latino students are unaware and uninformed about the processes involved in pursuing a college education, it is easy for school personnel to assume that such students are not interested and do not want to go to college (Martinez, 2003). Counselors may assume they need not meet to talk with these students about college planning nor encourage enrollment in more college-prep courses and may reserve their time only for students who are very decided about going to college (Falbo & Romo, 1996; Gonzalez & Ortiz, 2000). Like parental influence, few studies, to date, have examined the influence of school personnel (teachers, counselors, etc.) upon Latino students and their postsecondary education status.

Finally, in regards to psychological variables, the cultural characteristic of fatalismo (being resigned to one's fate) has been thought to play a significant role in preventing Latinos from overcoming obstacles to development (Sue & Sue, 2003) and from establishing a more internal locus of control. Fighting for educational opportunity and advancement may require a more internal locus of control. Thus, it would seem important to consider psychological variables such as locus of control and self-efficacy when investigating postsecondary education among Latinos. In addition, the construct of self-esteem has been considered to play an important role in school bonding and educational advancement especially among ethnic/racial minorities (Erkut & Tracy, 2002; Way & Robinson, 2003). Whether these psychological variables are significant predictors for postsecondary educational attainment among Latinos has yet to be investigated.

PURPOSE OF THIS STUDY

Given the increasing number of Latino students who go on to postsecondary education (PSE) but who fall to achieve a baccalaureate degree, it is vitally important, especially in the absence of previous research, to study factors that distinguish levels of PSE among Latinos. Toward this purpose, the present study employed a PSE outcome variable with four categories of students: some PSE, no degree; certificate or license; associate's degree; and bachelor's degree or higher by the year 2000. For investigating factors in PSE status, this study utilized a combination of family, societal, educational, and psychological variables derived from previous literature on Latino educational attainment.

This study sought to answer the following question: What are the significant factors that distinguish the ever increasing number of Latino students who go on to postsecondary education and do not complete a degree from the much smaller percentage who complete a bachelor's degree?

METHOD

Our data came from the 1988-2000 National Educational Longitudinal Study (NELS; U.S. Department of Education, 2002). NELS began with eighth graders in 1988 and collected subsequently four waves of data in 1990, 1992, 1994, and 2000. For the independent variables, data came from the 1990 wave of NELS when students were in their sophomore year of high school. This decision was based on the critical nature of sophomore year in the determination of future educational achievement because those doing poorly in sophomore year are at greater risk for dropping out at the end of that year (Clarke, Haney, & Madaus, 2000; Egemba & Crawford, 2003). Chronologically, sophomore year is also crucial for making decisions about courses in view of postsecondary educational plans as students are given more opportunities and choices to plan their curriculum (Savickas, 1999). For the dependent variable (PSE status), this study utilized the fourth-wave 2000 data, 12 years after the base-year data and 8 years after the students were scheduled to graduate from high school.

Participants

Participants were 866 Latino young people who by the year 2000 had attended a postsecondary institution and were part of the 1988-2000 NELS. Data came from two waves of the study: 1990, when participants were sophomores in high school; and 2000, 8 years after scheduled graduation from high school. Data were weighted accordingly (see Curtin, Ingels, Wu, & Heuer, 2002). Of the participants, 51% had some PSE with no degree, 11% had a certificate or license, 12% had an associate's degree, and 26% had a bachelor's degree or higher. Males made up 45% of the sample, females 55%. Sixty-three percent identified themselves as Mexican or Chicano, 5% as Cuban, 9% as Puerto Rican, and 23% as other Hispanic. The majority were born in the United States (87%).

Variables

Background variables. Background variables included gender, SES, language minority status, and cognitive ability (reading and math). SES is a composite variable derived from father's education level, mother's education level, father's occupation, mother's occupation, and family income. Language minority status is defined as coming from a home in which a language other than English (in this case, Spanish) is typically spoken. In the sample for this study, 58% of students were classified as language minority. The National Center for Educational Statistics measured cognitive ability through the use of reading and math standardized tests with item response theory (IRT) scores. IRT scores use a pattern of right, wrong, and omitted responses to account for the difficulty, discriminating ability, and guess-ability of each item (Ingels et al., 1994). This study made use of reading and math scores obtained in the sophomore year.

Parent support. The parent support variable was derived from a mean of 11 Likert-scaled items, which included how often parents checked homework, helped with homework, attended school meetings, called teachers or counselors, and attended a school event along with how often the student talked with a parent about school courses, school activities, grades, things studied in class, preparation for ACT/SAT, and going to college. Higher scores indicate a higher degree of parental support. Cronbach's alpha for the parent support composite variable was .81.

Teacher support. The teacher support variable was derived from a mean of six Likert-scaled items according to whether the student agreed or disagreed with the following statements: Teachers are interested in students, teachers praise effort when the student works hard, most teachers listen to the student, students in school get along well with teachers, in general the teaching is good, and the student feels put down by teachers (this last item was reverse scored). Higher scores indicate a higher degree of teacher support. Cronbach's alpha for the teacher support composite variable was .73.

Psychological variables. The two psychological variables included in the model were locus of control and self-esteem, both derived from a mean of Likert-scaled items. For locus of control, students were asked to rate their agreement/disagreement with the following items: does not have enough control over life; good luck is more important than hard work; when getting ahead, someone/something stops student; feels plans hardly ever work out; when makes plans, student is certain they will work; and chance, luck is very important in student's life. Higher scores indicate a higher sense of control over one's future. Cronbach's alpha for the composite locus of control variable was .72.

For self-esteem, students were asked to rate their agreement/disagreement with the following items: feels good about himself or herself; is a person of worth; is able to do things as well as others; on the whole, is satisfied with self; feels useless at times; at times, thinks he or she is no good at all; and does not have much to be proud of. Higher scores indicate a higher degree of positive self-esteem. Cronbach's alpha for the self-esteem composite variable was .82.

Data Analysis

The dependent variable in this analysis was PSE status, which is a categorical variable with four levels: some PSE, no degree; certificate or license; associate's degree; and bachelor's degree or higher by the year 2000. In order to model the relationship between a categorical dependent variable with more than two possible values and a set of independent or predictor variables, a multinomial logistic regression was used (Norusis, 2004). Logistic regression models produce odds ratios for the independent variables. These odds reflect the increase or decrease in the likelihood of an outcome (e.g., PSE status) for every one-unit increase in the independent variables. Because our dependent variable has four possible values, three nonredundant logits are formed. For each group of the dependent variable, the log of the ratio of the probability of being in that group is compared to being in the baseline group. For this analysis, the first category (some PSE, no degree) was the baseline or reference group to which the other three groups were compared based on the independent variables.

RESULTS

The original multinomial logistic regression model had nine predictor variables (gender, SES, language minority status, parent support, teacher support, locus of control, self-esteem, reading ability, and math ability). Likelihood ratio tests indicated that reading ability ([chi square] [3,868] = 2.04, p = .56) and self-esteem ([chi square] [3, 868] = 2.28, p = .52) were not significant in the overall model and therefore were dropped from subsequent analyses. The revised model included the seven remaining variables. Correlations for the variables used in the principal analysis are reported in Table 1.

For the multinomial logistic regression examining the effects of the seven predictor variables, the likelihood ratio test for the overall model revealed that the overall model was significantly better than the intercept-only model ([chi square] [21, 868] = 297.46, p < .01). In other words, the null hypothesis (that the regression coefficients of the independent variables are zero) was rejected. In addition, the likelihood ratio test for individual effects reveals that all of the independent variables are significantly related to the categories of the dependent variable (SES: [chi square] [3] = 38.75, p < .001; parent support: [chi square] [3] = 18.46, p < .001; teacher support: [chi square] [3] = 9.83, p < .05; math ability: [chi square] [3] = 44, p < .001; locus of control: [chi square] [3] = 51.86, p < .001; gender: [chi square] [3] = 38.82, p < .001; language minority status: [chi square] [3] = 12.65, p < .01).

Table 2 reports the parameter estimates from the logistic regression model examining the effects of the independent variables on PSE status. Estimates of the predictor variables are provided for the three different levels compared to some PSE, no degree: (a) certificate or license, (b) associate's degree, and (c) bachelor's degree or higher. According to these results, the parameter estimates for gender ([beta] = 1.12, p < .001) and locus of control ([beta] = .66, p < .05) are significantly different from zero for the first logit (certificate or licensure compared to some PSE, no degree). In other words, gender and locus of control are significantly and positively related to this distinction. For gender, men were coded the value 1 and women were coded the value 2. Therefore, because the logistic regression coefficient is positive, women were more likely to complete a certificate or licensure than were men. Likewise, students who reported a more internal locus of control were more likely to complete a certificate or licensure than were those with an external locus of control.

Table 2 also reports that as the "some PSE, no degree" group is compared to higher and higher levels of educational attainment, more predictor variables appeared to come into play and to be significantly related to whether these students pursued postsecondary education. For example, for the second logit (associate's degree compared to some PSE, no degree), the parameter estimate for gender is not significantly related to the separation of these groups as it was for the first logit ([beta] = .19, ns). However, locus of control ([beta] = 1.31, p < .001) is significantly related, as are parent support ([beta] = .71, p = .01) and teacher support ([beta] = -.89, p < .01). Therefore, the parameter estimates for parent support and teacher support are not significantly different for the first logit but are for the second logit. Importantly, the logistic regression coefficients for locus of control and parent support are positive, indicating that students who reported a more internal locus of control and higher levels of parent support in their sophomore year of high school were more likely to complete an associate's degree compared to students in the some PSE, no degree group. However, the logistic regression coefficient for teacher support was negative, indicating that students who reported higher levels of teacher support were less likely to complete an associate's degree.

Finally, the third logit compares students who completed a bachelor's degree or higher to those who completed some PSE, no degree. In the third logit, even more variables had a strong positive effect on whether students completed a bachelor's degree than for the previous two logits. Parent support ([beta] = .90, p < .001) and locus of control ([beta] = 1.46, p < .001) continue to have strong effects on postsecondary education, as they were significantly related to the separation between the bachelor's degree group and the some PSE, no degree group. In addition, SES ([beta] = .71, p < .001), math ability ([beta] = .06, p < .001), gender ([beta] = .99, p < .001), and language minority status ([beta] = .53, p < .05) were significantly different from zero for this logit, indicating that they are also significantly related to this separation.

An interpretation of the odds ratio in the overall model reveals that students with a more internal locus of control were three times more likely (a 332% increase) to complete a bachelor's degree compared to students with a more external locus of control (odds m 4.32), while controlling for all other variables in the model. SES (odds = 2.03) and parent support (odds = 2.46) also had a strong positive effect on whether or not the students completed a bachelor's degree. With all other variables controlled, a one-unit increase in each of these variables increased the odds of degree completion by 103% and 146%, respectively. In addition, the positive regression coefficient for gender indicates that women were over one and a half times more likely (a 168% increase) to complete a bachelor's degree than were men (odds = 2.68). Students with higher math ability were 6 times more likely than those with lower math ability (odds = 1.06) to complete a bachelor's degree, while controlling for all other variables in the model. Finally, students who were considered a language minority (i.e., came from a home where Spanish was the predominant language) were 70% more likely to complete a bachelor's compared to students who spoke English predominantly (odds = 1.70).

Regarding effect sizes, the Nagelkerke [R.sup.2] (Norusis, 2004) in the overall model was .32. Therefore, the independent variables included in the model explained 32% of the variability in postsecondary educational attainment.

DISCUSSION

The model significantly predicted Latino students' postsecondary education attainment. Moreover, as Latino students increased their PSE status, more factors from the proposed model came to play a significant role. Two of the predictor variables (locus of control and gender) were significant in differentiating between students who had a certificate or license and those who had some PSE but no degree. For the students who received an associate's degree, three variables (parent support, teacher support, and locus of control) were significant predictors. Finally, six predictor variables (SES, parent support, math ability, locus of control, gender, and language minority status) differentiated between students who completed a bachelor's degree or higher and those who had some PSE but no degree.

The results from this study also revealed the relative strength of the predictor variables, with locus of control as the strongest predictor for all three categories. In other words, the students with a more internal locus of control were three times more likely to get a bachelor's degree or higher, 2.7 times more likely to get an associate's degree, and .92 times more likely to get a certificate or license. Gender was also a strong predictor for certificate or licensure and bachelor's degree or higher, with Latinas being more likely to achieve a higher PSE status than their male counterparts. Parent support and language minority status were also strong predictors for associate's degrees and bachelor's degrees or higher. Those students who reported higher levels of parent support were 1.03 times more likely to get an associate's degree and almost one and a half times more likely to get a bachelor's degree or higher. Latino students who came from homes where Spanish was typically spoken were almost one time (.70) more likely than those who came from homes where English was typically spoken to complete a bachelor's degree or higher.

Dropped from the original model were reading ability and self-esteem, because likelihood ratio tests indicated that these two factors did not play a significant role in postsecondary educational status of Latino students. The result of the reading factor is consistent with other studies that show math more than reading ability and achievement is significant in succeeding academically in college, especially at the baccalaureate level (Adehnan, 1999; Trusty, 2002; Trusty & Niles, 2003). The results of the present study show that math ability was a significant factor at the bachelor's degree or higher category of PSE but not at the certificate/license or associate's degree levels. For Latino high school students who plan to graduate from a 4-year college, their math ability and achievement needs to be carefully considered. School counselors can help these students receive extra instruction, ideally from the more effective and experienced teachers in the school, and encourage them to persevere in higher-level math courses in spite of challenges.

Also dropped from the original model was the self-esteem factor, a result that may be somewhat surprising but should be considered in fight of the significance of another factor, locus of control. Locus of control was the most significant factor at both the associate's and bachelor's (or higher) degree levels. The results of this study show that, for Latino students, having a greater sense of control over their future is more important for PSE attainment than how much they like themselves. This finding is interesting from both a cultural and career development perspective. There is some evidence to suggest that Latinos as a group have a more external locus of control (typically referred to as fatalismo) when compared to the White Euro-centric population (Sue & Sue, 2003). Educational attainment and the tasks involved (e.g., studying) may favor a more internal locus of control that allows one to perceive higher education as a contributor to a greater sense of control over one's life. The acquisition of a more internal locus of control by their sophomore year of high school appears, based on the results of this study, to have played a very significant role in enhancing Latino students' PSE status. While the two constructs of internal locus of control and self-efficacy are not to be equated, there is a relationship. Specifically, those who believe in their ability and expect positive outcomes will also possess a greater sense of control over their lives. In regards to the present study, this belief about oneself and one's efforts (i.e., locus of control) was related to higher educational outcomes much more so than how participants felt about themselves (i.e., self-esteem).

Another very significant factor in the overall model was that of parent support. Parent support was a significant factor for the attainment of an associate's and bachelor's or higher degree. It should be noted that the parent support variable did not include financial support. Six of the 11 items used in the parent support composite variable dealt with communication between participants and their parents in regards to school-related matters; two dealt with homework (checking and helping); and three dealt with parent involvement in the school. While some Latino parents might find it difficult to deal with homework either because of language issues or poor educational background and to become involved in school because of demanding and inflexible work schedules (especially if they are from a low-SES background), communication between parents and children in regards to academic matters is always possible. The results of the present study suggest that such communication as a part of overall parent support enhances postsecondary educational attainment.

Socioeconomic status, often thought to be a significant factor in PSE attainment, was, in the present study, significant only at the bachelor's degree or higher category. This finding is not surprising because it costs a good deal of money to graduate from a 4-year institution. More surprising is that locus of control and parent support had a more significant share of the variance than did SES. While no one can minimize the importance of needing a good deal of financial support to attain a bachelor's degree, the results of this study indicate that other variables may be just as, if not more, important as the SES of one's family in distinguishing a Latino student with a bachelor's degree from one with some PSE and no degree.

Teacher support was not a significant variable in the overall model except at the associate's degree level, where it was a negative predictor. Statistically speaking, those Latino students with more teacher support had less of a chance to complete an associate's degree. One explanation for this finding perhaps has to do with the construct itself. The items used from the 1988 NELS database may not be valid items for constructing a teacher support variable. Most of the items were relational in nature and on the face level were not related to support for higher educational attainment. Another explanation could be that Latino students at the bachelor's or higher level have succeeded without teacher support, at least as understood or perceived in their sophomore year of high school. Because Latino students are concentrated in low-income schools that tend to have a greater share of less qualified teachers, it may be that Latino students who achieve higher levels of PSE are doing so in the absence of teacher support in high school. One study (Sciarra, 2004) found that Latino students garnered teacher support only after the teachers discovered they were good students. Future research will need to examine further teacher support as a factor in Latino educational attainment after high school.

Implications for School Counseling

According to the results of the present study, the two most significant variables among Latino high school sophomores in differentiating those who had some PSE and no degree from those who went on to receive a bachelor's degree or higher were parent support and an internal locus of control. This finding suggests a strategic role for school counselors who work with Latino students at the elementary, middle, and high school levels. In terms of parent support, the need for communication and information sharing between parents and their children is crucial. While some Latino parents may have difficulty (perhaps because of language difficulties and loss of income from missing work) establishing a physical presence in the school, counselors should not take this as a sign of disinterest. Rather, they should take every opportunity to meet with Latino parents outside of the school setting (in neighborhood community rooms, local churches, etc.) to explain, via an interpreter if need be, the postsecondary preparation process and the various possibilities that exist for their children, and to encourage parents to talk with their children about school in spite of perhaps not speaking English or having an impoverished educational background. School counselors can encourage parents to check on homework even if they do not fully understand all that's being done.

We know from recent research that the quality and intensity of the high school curriculum play an important role in determining the quality of postsecondary life (Adelman, 1999; Education Trust, 2005). School counselors need to emphasize this to Latino parents whose children may want to avoid the more challenging and demanding courses. Many Latino parents will not understand the consequences of certain curriculum choices and may easily succumb to their children's avoidance of failure. On the other hand, negative stereotypes of Latino students may give way to subtle racism in the form of not encouraging these students to pursue higher-level courses. School counselors, by working collaboratively with Latino parents in helping them under stand and talk with their children about the effects of course-taking variables upon postsecondary achievement, will be operationalizing their roles as advocates and systemic change agents--roles specified by the ASCA National Model[R] (American School Counselor Association, 2005).

In the same vein, school counselors must encourage their Latino students to communicate with their parents about school. A Latino student might often resist, saying "My parents don't speak English" or "My parents grew up in another country and don't understand much about school here." School counselors should not capitulate to such resistance but help to process what can be a student's frustration in communicating with his or her parents, especially if they are immigrants with little knowledge of how the school system functions. Given the results of this study that indicate language minority students were no less likely to achieve a bachelor's degree or higher, school counselors can persevere in helping Latino students initiate conversations with their parents about school. While communication between adolescents and their parents often can be difficult, school counselors can take advantage of the culturally characteristic closeness and loyalty that exist in the Latino family to promote this communication. Leaving home after high school to go to college, for example, may be more difficult emotionally for Latino adolescents and their parents than for a White family. The local community college may appear a more attractive alternative. Rather than dismiss their reluctance about leaving home, school counselors can facilitate a discussion about the pros and cons of attending the local college.

In regards to locus of control, the implications for school counseling are enormous. Counseling has had a long history of helping people move from an external to an internal locus of control through various methods of empowerment. For example, a collaborative approach in which counselor and client work together in generating strategies to achieve particular goals avoids a dependent relationship on the counselor and helps the client take more control of his or her life, build a positive self-identity, and increase self-efficacy. The results of the present study show that those Latino sophomores who possessed an internal locus of control had a much greater chance of attaining a PSE degree, especially a bachelor's degree. While the cultural challenges in helping Latino students achieve a more internal locus of control may be formidable (e.g., low-SES background, ethnic/racial nondominant status), school counselors need to realize the importance of doing so.

School counselors can be more proactive in sharing information with Latino students who are often uninformed about the processes involved in preparing for life after high school. The sharing of information and resources is often the first step in gaining a sense of control over one's life and future. In addition, school counselors can maintain the goal of increasing Latino students' sense of self-efficacy. Beginning in the elementary schools, they can collaborate with teachers and help to provide Latino students with experiences that are the antithesis to learned helplessness. Such experiences can greatly help Latino students see positive results from their efforts.

Limitations and Future Research

The analyses conducted in this study are limited to the data contained in the 1988 NELS. For example, many of the teacher support items were general in nature and related to the school environment, but none dealt specifically with support for postsecondary education that would have been more relevant to this study. Reading ability and math ability were determined solely through the use of standardized tests and therefore suffer from the limitations of all standardized tests. Future studies will need to include items assessing support for educational attainment and other criteria such as reading and math GPA along with course-taking variables--that is, remedial versus grade-equivalent versus advanced.

The significance of parent support, language minority status, and internal locus of control in the present study suggests that future research should examine these variables in more depth. This study, for example, did not differentiate parent support according to SES or family composition. Is parent support greater in families with higher incomes or in those with two-parent versus mother-only families? The fact that students coming from predominantly Spanish-speaking families had a slightly greater chance of completing a bachelor's degree raises the research question as to whether more traditional Latino families have more forms of support available to them. Perhaps there is a relationship with ethnic identity. Future research needs to explore this. Finally, because internal locus of control was a very significant contributor to differentiating levels of PSE status, future research should examine its relationship to other motivating variables such as self-efficacy. For example, rather than examining if a student wants to go to college and graduate, future research might examine whether Latino students believe they have the ability to go to college and succeed.

Conclusion

This study has sought to examine factors that distinguish the large number of Latino students who have some postsecondary education from the relatively small number who complete a degree. Our sample was representative of this discrepancy: Of the 866 Latinos who had PSE in their background, 51% had no degree versus 26% with a bachelor's degree or higher. The multinomial logistic regression used for data analysis found that parent support and locus of control were the strongest factors differentiating those Latino students with some PSE and no degree from those with a bachelor's degree or higher.

Latinos are the largest nondominant group in the United States and growing rapidly. They are also a very young population. While being ethnically and racially diverse, Latinos are concentrated in lower-SES situations. The ability of the United States to draw in the future from a talented pool of young workers will depend, in part, on the higher educational attainment of Latinos. School counselors, whose training requires multicultural sensitivity and a commitment to social justice, are in an excellent position to be proactive with Latino students and their parents to initiate a process that can lead to greater PSE success. By enhancing parent support especially in terms of communication around school-related matters and Latino students' self-efficacy along with the consequential internal locus of control that it fosters, school counselors can contribute significantly to increasing the future pool of highly qualified Latino workers in this country.

References

Adelman, C. (1999). Answers in the tool box: Academic intensity, attendance patterns, and bachelor's degree attainment. Jessup, MD: U.S. Department of Education.

American Institutes for Research. (2003). AM statistical software. Retrieved May 3, 2005, from http://am.air.org

American School Counselor Association. (2005). The ASCA national model: A framework for school counseling programs (2nd ed.). Alexandria, VA: Author.

Aviles, R. M., Guerrero, M. P., Howarth, H. B., & Thomas, G. (1999). Perceptions of Chicano/Latino students who have dropped out of school. Journal of Counseling and Development, 77, 465-473.

Bamaca-Gomez, M., & Plunkett, S. (2003).The relationships between parenting, acculturation, and adolescent academics in Mexican-origin immigrant families in Los Angeles. Hispanic Journal of Behavioral Sciences, 25, 222-239.

Behnke, A., Diversi, M., & Piercy, K. (2004). Educational and occupational aspirations of Latino youth and their parents. Hispanic Journal of Behavioral Sciences, 26,16-35.

Clarke, M., Haney, W., & Madaus, G. (2000). High stakes testing and high school completion. National Board of Educational Testing and Public Policy, 1(3), 1-15.

Curtin, T. R., Ingels, S. J., Wu, S., & Heuer, R. (2002). National Education Longitudinal Study of 1988: Base year to fourth follow-up data file user's manual (NCES 2002-323). Washington, DC: U.S. Department of Education, Office of Educational Research and Improvement.

Education Trust. (2005}. Achievement in America. Retrieved July 10, 2005, from http://www2.edtrust.org/EdTrust/ Product+Catalog/PowerPoint.htm

Egemba, M. O., & Crawford, J. R. (2003, April). An analysis of Hispanic students drop-out rates. Paper presented at the annual meeting of the American Educational Research Association, Chicago.

Erkut, S., & Tracy, A. J. (2002). Predicting adolescent self-esteem from participation in school sports among Latino subgroups. Hispanic Journal of Behavioral Sciences, 24, 409-429.

Falbo, T., & Romo, H. D. (1996). Latino high school graduation. Austin, TX: University of Texas Press.

Fry, R. (2002). Latinos in higher education: Many enroll, too few graduate. Washington, DC: Pew Latino Center.

Gamoran, A., & Hannigan, E. C. (2000). Algebra for everyone? Benefits of college-preparatory mathematics for students with diverse abilities in early secondary school. Educational Evaluation and Policy Analysis, 22, 241-254.

Gonzalez, R. (2002, July).The No Child Left Behind Act: Implications for local educators and advocates for Latino students, families, and communities. National Council of La Raza: Issue Brief, pp. 2-14.

Gonzalez, R., & Ortiz, F. (2000). Latino high school students' pursuit of higher education. Aztlan, 25, 67-107.

Immerwahr, J. (2003). With diploma in hand: Hispanic high school seniors talk about their future. San Jose, CA: National Center for Public Policy and Higher Education, and the Public Agenda.

Ingels, S. J., Dowd, K. L., Baldridge, J. D., Stipe, J. L., Bartot, V. H., & Frankel, M. R. (1994). National Educational Longitudinal Study of 1988, second follow-up: Student component data file user's manual (NCES 1994-374). Washington, DC: U.S. Department of Education, Office of Educational Research and Improvement.

Ingersoll, R. M. (2003). Education Trust of 1999-2000 school and staffing survey. Unpublished data, University of Pennsylvania.

Llagas, C., & Snyder, T. D. (2003). Status and trends in the education of Hispanics. Washington, DC: U.S. Department of Education, National Center for Educational Statistics.

Martinez, M. (2003). Missing in action: Reconstructing hope and possibility among Latino students placed at risk. Journal of Latinos and Education, 2, 13-21.

McWhirter, M. J. (2004). School dropouts. In J. J. McWhirter, B.T. McWhirter, E. H. McWhirter, & R. J. McWhirter (Eds.), At-risk youth (3rd. ed., pp. 95-113). Pacific Grove, CA: Thomson/Brook-Cole.

National Center for Educational Statistics. (1994). National assessment of educational progress 1992 trends in academic progress. Washington, DC: U.S. Department of Education.

National Center for Educational Statistics. (2004). Educational Longitudinal Study: 2002 data files and electronic codebook system. Washington, DC: Author.

Norusis, M. J. (2004). SPSS 13.0 advanced statistical procedures companion. Upper Saddle River, N J: Prentice Hall.

Pabon, E. (1998). Hispanic adolescent delinquency and the family: A discussion of sociocultural influences. Adolescence, 33, 941-956.

President's Advisory Commission on Educational Excellence for Latino Americans. (2003). Road to a college diploma: The complex reality of raising educational achievement for Latinos in the Unites States (interim report). Washington, DC: Government Printing Office.

Quezada, R. L., Diaz, D. M., & Sanchez, M. (2003). Involving Latino parents. Leadership, 33, 32-34.

Savickas, M. L. (1999).The transition from school to work: A developmental perspective. Career Development Quarterly, 47, 326-336.

Sciarra, D.T. (2004).A qualitative investigation into Latino academic resiliency. Unpublished manuscript.

Sue, D. W., & Sue, D. (2003). Counseling the culturally diverse: Theory and practice (4th ed.). New York: Wiley & Sons.

Trusty, J. (2002). Effects of high-school course-taking and other variables on choice of science and mathematics majors. Journal of Counseling and Development, 80, 464-474.

Trusty, J., & Niles, S. G. (2003). High-school math courses and completion of the bachelor's degree. Professional School Counseling, 7, 99-107.

U.S. Bureau of the Census. (2002a). The condition of education, 2002: March current population survey, 1971-2000. Washington, DC: Author.

U.S. Bureau of the Census. (2002b). Current population reports. Washington, DC: U.S. Government Printing Office.

U.S. Department of Education, Office of Educational Research and Improvement. (2002). National Education Longitudinal Study: 1988-2000 data files and electronic codebook system, base year through fourth follow-up (NCES 2002-322). Washington, DC: Author.

Way, N., & Robinson, M. G. (2003). A longitudinal study of the effects of family, friends, and school experiences on the psychological adjustment of ethnic minority, low-SES adolescents. Journal of Adolescent Research, 18, 324-346.

Daniel T. Sciarra is an associate professor of counselor education at Hofstra University, Hempstead, NY. E-mail: [e-mail] Melissa L. Whitson is with the Counseling Psychology Program at Teachers College, Columbia University, New York.

Table 1. Correlations for SES, Parent Support, Teacher Support, Math
Ability, Locus of Control, Gender, and Language Minority Status

                      SES     Parent    Teacher     Math    Locus of
                              Support   Support   Ability   Control
SES                 1.00
Parent support       .18 **   1.00
Teacher support     -.20 **    .28 **   1.00
Math ability         .35 **    .07 *     .02      1.00
Locus of control     .22 **    .40 **    .25 **    .30 **    1.00
Gender               .03      -.04      -.00       .09 *      .06
Language minority   -.37 **   -.00       .15 **   -.20 **     .04 *
status

                             Language
                    Gender   Minority
                              Status
SES
Parent support
Teacher support
Math ability
Locus of control
Gender              1.00
Language minority   -.17 **   1.00
status

Note. N = 866. * p < [less than or equal to] .05.
** p [less than or equal to] .01.

Table 2. Parameter Estimates from Logistic Regression Model Examining
Effects of Predictor Variables on Postsecondary Education Status

                            Postsecondary Education Status

                      Certificate/      Associate's       Bachelor's
                        License           Degree          or Higher

Variable            [beta]     Odds   [beta]    Odds     [beta]    Odds

SES                 -.30        .74   -.09       .91     .71 ***   2.03
Parent support       .37       1.45    .71 **   2.03     .90 ***   2.46
Teacher support      .09       1.10   -.89 **    .41    -.36        .70
Math ability        -.03        .96   -.03       .97     .06 ***   1.06
Locus of control     .66 *     1.93   3.71              1.46 ***   4.32
Gender              1.12 ***   3.07    .19      1.21     .99 ***   2.68
Language minority   -.04        .96   -.41       .67     .53 *     1.70

Note. The reference category is PSE, no degree. Standard errors for
sampling design effects calculated with the use of AM software
(American Institutes for Research, 2003) were adjusted. Nagelkerke
[R.sup.2] = .330.

N = 1,660. * p [less than or equal to] .05. ** p [less than or equal to]
.01. *** p [less than or equal to] .001.