Neighborhood level factors are implicated in many youth risk behaviors--delinquency, drug use, and risky sex--but neighborhood influences on the effectiveness of programs to prevent such behaviors have not received much systematic attention. We examine how neighborhood characteristics affected the efficacy of a school-based SAMHSA model program for youth substance use prevention, the culturally-grounded keepin' it REAL program. Conceptualizing neighborhood influences on youth prevention programs via theories of neighborhood social disorganization, immigrant adaptation, and social isolation, we investigate neighborhood poverty, crime, family structure, and immigrant composition. Using data from a large, randomized trial of the keepin' it REAL program, we tested for whether neighborhood factors moderated the desired program effects. Pre- and post-test questionnaires from a predominantly Mexican heritage sample of 3595 7th grade students were obtained in Phoenix middle schools, 25 treatment and 10 control. Outcomes were last 30 day substance use, future intentions to use drugs and other drug use attitudes. Using GIS techniques to map 2000 Census and local police data onto the neighborhoods surrounding each school, we then calculated an estimate for each neighborhood of the percentage of residents: (a) who were immigrants to the U.S. within the last five years, (b) living on incomes below the U.S. poverty line, and (c) in families headed by a single mother. Police department reports on crimes per 1/4 mile square grid were aggregated to estimate violent crime rates. To address data clustering and attrition, multilevel models and multiple imputation missing data techniques were employed. Models tested program participation impacts (treatment vs. control) on post-test substance use outcomes controlling for pre-test levels, age, gender, academic performance, and SES. RESULTS: Neighborhood level effects were confined to less acculturated Latino students, who comprised over half the sample. Neighborhood factors moderated the effectiveness of the program in reducing alcohol use but were not significant factors in the program's effectiveness in combating cigarette and marijuana use or pro-drug attitudes. Among less acculturated Latinos, living in poorer neighborhoods and those with many single mother families decreased program effectiveness in combating alcohol use, but a high neighborhood immigrant composition increased program effectiveness for them. Unexpectedly, the program was also more effective against alcohol use in neighborhoods with higher rates of crime. No significant neighborhood influences on the program's effectiveness were found for more acculturated Mexican Americans or for non-Latino White students. Our interpretation of the results considers how these neighborhood factors, which are typically described as contributing to neighborhood disorganization and ineffective social control, might also tap community resiliency against drug use. We address the distinctive social environment of southwestern cities, with both large, established Mexican-Americans communities and many recent Mexican immigrants, relatively low unemployment coupled with high poverty rates, and where households headed by single mothers are relatively less common. We consider neighborhood factors on prevention programs by contrasting individual versus neighborhood level conceptions of risk, and a dual perspective on neighborhood risks as presenting both greater obstacles to prevention and opportunities for it.