The study examined the role of risk and protective influences among high school Latino/a students by using a sample from the Education Longitudinal Study (ELS: 2002). A clustering approach was used to create profiles comprising a variety of variables, including students' involvement in extracurricular, home-based parental involvement, and social cognitive factors that are typically thought of as risk factors for school completion, mathematics achievement, and enrollment in postsecondary education. Latent class analyses (LCA) was used to identify the academic risk groups, assess academic risk group differences with respect to key covariates including gender, SES, immigrant status, whether English is a student's native language, preschool attendance, and 10th grade mathematics. This study assessed academic risk group differences with respect to 12th grade mathematics achievement, dropout rates, and enrollment in postsecondary education.

Results indicated four distinct academic risk classes. The classes revealed complex patterns of extracurricular involvement, academic discussions with parents, and social cognitive factors. The profiles were labeled as follows: Highly Engaged, Low School Engagement, Low Individual Engagement , and Low Parental Engagement. The Highly Engaged class had the lowest risk across all three academic outcomes. That is the Highly Engaged class had the highest twelfth grade mean mathematics achievement score, was the class most likely to attend postsecondary, and least likely to dropout. In contrast the Low Parental Engagement class had the highest risk across all three academic outcomes.

Results indicate that using the Highly Engaged class as the reference group, students who did not attend preschool were significantly more to be represented in the Low Parental Engagement or Low School Engagement classes. Male students were significantly more likely to be in the Low Parental Engagement class. Students who were not low-income were less likely to be represented in the Low Parental Engagement class and Low School Engagement classes. Also, 10th grade mathematics were significant predictors for student in the Low Parental Engagement class, Low School Engagement, and Low Individual Engagement classes. Students whose native language was not English were less likely to be in the Low Individual Engagement class. Implications of the findings and limitations are discussed.