Purpose: Although mentoring is increasingly hailed as one of the most critical components of US teacher induction programs, the corresponding research base has failed to provide conclusive support for the effectiveness of teacher mentoring. Design/methodology/approach: Cross-sectional data from the 2015 to 2016 National Teacher and Principal Survey (NTPS) was used to empirically evaluate the relationship between school-based mentoring programs and US first-year teachers' occupational stress risk. NTPS items assessing teachers' workplace demands, resources and mentoring experiences were used to examine associations between stress risk and mentoring supports received. Multiple regression analysis and chi-square tests of independence were conducted to answer the study's three research questions. Findings: Mentor status and study covariates statistically significantly predicted teacher stress risk, with mentor assignment being associated with decreased stress risk. Findings also revealed teachers who were not assigned a mentor were significantly more likely to experience greater stress risk and when assigned mentors, teachers with greater stress risk reported significantly different mentoring experiences than did teachers with less stress risk. Originality/value: The current study addresses common limitations found in research on teacher mentoring and occupational stress, including the use of local, small samples and the absence of a comparison group and control variables. The associations between first-year teacher mentoring experiences and stress risk have not been firmly established and the current study provides needed evidence to support that mentored first-year teachers are more likely to report lower stress risk than their unmentored colleagues.