ContextBiospecimen analysis may enhance confidence in the accuracy of self-reported substance use among adolescents and transitional age youth (TAY). Associations between biospecimen types and self-reported use, however, are poorly characterized in the existing literature. Objective: We performed a systematic review of associations between biospecimen-confirmed and self-reported substance use. Data sources: PubMed, Embase, and Web of Science. Study selection: We included studies documenting associations between self-reported and biospecimen-confirmed substance use among adolescents (12-18 years) and TAY (19-26 years) published 1990-2020. Data extraction: Three authors extracted relevant data using a template and assessed bias risk using a modified JBI Critical Appraisal Tool. Results: We screened 1523 titles and abstracts, evaluated 73 full texts for eligibility, and included 28 studies. Most studies examined urine (71.4%) and hair (32.1%) samples. Self-report retrospective recall period varied from past 24 h to lifetime use. Agreement between self-report and biospecimen results were low to moderate and were higher with rapidly metabolized substances (e.g., amphetamines) and when shorter retrospective recall periods were applied. Frequently encountered sources of potential bias included use of non-validated self-report measures and failure to account for confounding factors in the association between self-reported and biospecimen-confirmed use. Limitations: Study heterogeneity prevented a quantitative meta-analysis. Studies varied in retrospective recall periods, biospecimen processing, and use of validated self-report measures. Conclusions: Associations between self-reported and biospecimen-confirmed substance use are low to moderate and are higher for shorter recall periods and for substances with rapid metabolism. Future studies should employ validated self-report measures and include demographically diverse samples.