American Indian dropouts and academically at-risk students: an exploratory comparison

Abstract

Low levels of school retention among American Indian adolescents has resulted in calls for more broad based and systematic research on the correlates of dropping out. American Indian dropouts have been compared to their peers who are in good academic standing, but determining the influence of a variety of hypothesized correlates has been clouded by differences in academic abilities between these two groups. A comparison of American Indian students who have academic abilities similar to those of dropouts was predicted to allow the assessment of various correlates of dropping without the moderating influence of academic abilities. Data for this study was obtained from a NIDA and NIAAA funded study of American Indian adolescents. Participants were drawn from schools located in five different cities (four on reservations and one in a mid-sized city with a high percentage of American Indian students). The initial sample consisted of 827 self-identified American Indian adolescents who were categorized in to three school groups: 1) retained students; 2) academically at risk-students; and, 3) school dropouts. As students dropped out of school they were matched with to other youth: both of similar gender and grade. The academically at risk students were further matched on the grade point average of the dropouts. Portions of a multiscale survey completed by the participants were used in this exploratory study. Missing data, equating participant ages across school status groups and the use of only at-risk students and dropouts reduced the sample size to 206 American Indian females and 140 American Indian Males. Scales assessing perceptions of teachers, alcohol involvement, deviant behavior, school misbehavior, cultural identification, and use of an Indian language were subjected to a confirmatory factor analysis to determine their suitability for use with this population. All scales demonstrated measurement coefficients for this sample with the exception of the alcohol involvement scale and the cultural identification measures for females only. A series of ANOVA's with school status and gender as grouping variables were conducted. No interactions between school-status and gender were observed. At-risk students reported higher levels of perceiving teachers to care about the youth and Indians in general compared to dropouts. At-risk students were also less likely to report engaging in misbehavior at school. Correlates were subjected to a discriminant function analysis that correctly classified 77% of the male at-risk students. A similar analysis for females performed poorly. Results are discussed in relationship to gender differences, methodological issues, and prevention strategies.