Vaccines are approved by regulatory agencies based on randomized, placebo-controlled vaccine efficacy (VE) trials that demonstrate vaccination is safe and reduces the risk of acquisition of an infectious disease clinical endpoint. The resource-intensity of VE trials make it difficult to base vaccine approval on this gold-standard approach for every indication of interest (defined by host population, virus strain population, and/or the specific vaccine regimen). Therefore, an important objective in vaccinology is to develop ‘immune marker surrogate endpoints’ that can be used as primary endpoints in smaller and faster studies for provisional or traditional approval of vaccines. After summarizing why immune correlates are needed, the talk summarizes approaches to the evaluation of the quality of an immune marker as a surrogate endpoint for a clinical endpoint based on VE trials and on other evidence sources (lab development of immunoassays/biomarkers, natural history studies, post-approval epidemiological studies, vaccine mechanism studies, meta-analysis of multiple VE trials). In VE trials, data analysis evaluates ‘correlates of risk’ (CoRs) and ‘correlates of protection’ (CoPs) as distinct objectives. CoR analysis assesses the association of an immune marker with acquisition of the clinical endpoint in vaccine recipients. CoR analysis estimates association parameters and is limited in that it does not assess a causal effect of vaccination, such that an immune marker may be a CoR but fail to be a CoP. CoP analysis, on the other hand, studies how immune markers predict, cause, or mediate VE, with goal to generate evidence of whether and how an immune marker can be used to reliably predict VE against the clinical endpoint. An immune marker with established robust evidence for providing this reliable prediction can be designated a ‘non-validated surrogate endpoint’ or ‘validated surrogate endpoint’ (depending on evidence level) and thus be used for provisional or traditional vaccine approval, respectively. The talk summarizes statistical approaches for evaluating CoRs and CoPs in VE trials, illustrated by the Moderna COVID-19 VE trial. Approaches to evaluating CoPs include controlled direct effect analysis that assesses the causal effect of the marker on reducing risk and mediation analysis that assesses how much of the vaccine’s overall efficacy is mediated through the marker. Meta-analysis quantifies how well VE can be predicted by an immune marker.