This article explores tile concepts of risk and resilience applied to
children and youth. I suggest advantages in considering a definition o
f risk grounded in actual school performance and behavior, as opposed
to common conceptions of risk tied to various group-level probabilitie
s of failure. The study focuses on students doing poorly or lacking co
nfidence in finishing school as of the eighth grade who turn themselve
s around by the tenth grade. Using data from the National Education Lo
ngitudinal Study of 1988 (NELS:88) survey, analyses examine the nature
of performance-based risks and the conditions under which students im
prove their performance levels and outlooks bf grade 10. Tile estimate
d models of resilience shown influences resembling those found in stud
ies of individual resilience in other domains. The importance of famil
y supports, school responsiveness to students, and student involvement
in school and community activities stand out as predictors of recover
y from low performance. A few patterns within subgroups emerge: for in
stance, the noninfluence of socioeconomic status (SES) on resilience w
ithin the Hispanic and African American cohorts, and the finding that
Hispanic youth are less resilient in schools where there are perceived
problems with youth gangs.