EBER in situ hybridization (EBER) and LMP-1 immunohistochemistry (LMP-1) are widely used for identifying Epstein-Barr virus (EBV) within tumor cells of Hodgkin lymphoma (HL), but measurement error has never been formally evaluated. To determine assay reliability, 40 HL tumors with known EBV status were stained for both EBER and LMP-1 by two laboratories and reviewed twice by four hematopathologists. Inter- and intra-observer agreement were good to excellent, with kappas above 0.78 overall and above 0.60 for most subgroup analyses. However, reliability varied by histologic subtype, preparing laboratory, reviewer and EBV status determined on consensus review. For EBER, inter-observer agreement was high for nodular sclerosis HL but somewhat lower for EBV-negative mixed cellularity HL. For LMP-1, agreement was excellent for mixed cellularity HL but somewhat less reliable for EBV-positive nodular sclerosis HL. Agreement was good for EBER and LMP-1 applied to the same specimens but differed by consensus EBV status. The variability in assay interpretation justifies caution in comparing EBV association results across HL studies and underscores the need for interpretation guidelines.