The three viruses most studied as possible causes of human breast cancer are mouse mammary tumor virus-like sequences (MMTV-LS), Epstein-Barr virus (EBV), and oncogenic (high risk) types of human papilloma virus (HPV). The first step in fulfilling traditional criteria for inferring that a cancer is caused by a virus is to demonstrate the virus in the affected tissue. Molecular techniques, compared to host antibody assessment and immunohistochemistry, are the most definitive in establishing viral presence. Results of 85 original molecular research investigations to detect one or more of the three viruses have been extremely divergent with no consensus reached. We evaluated the methodology of these studies for the following: type of molecular assay, DNA/RNA quality control, positive and negative assay controls, type of fixation, genome targets, methods for preventing and detecting molecular contamination, pathology of specimens processed, sample size, and proportion of specimens positive for the viral genome region targeted. Only seven of the studies convincingly demonstrated the presence of an oncogenic virus biomarker (EBV: 4/30 studies (13%); HPV 3/29 studies (10%), whereas 25 convincingly demonstrated absence of the virus studied (MMTV-LS: 4/25 (16%); EBV: 15/30 (50%); 6/29 (21%). The remainder of the studies suffered shortcomings, which, in our opinion, prevented a definitive conclusion. Only one of the studies compared frequency of the virus in breast tissue of breast cancer patients versus appropriate normal control subjects with no history of breast cancer. None of the studies were designed as epidemiologic studies to determine if the presence of the virus was significantly associated with breast cancer. Based on our evaluation, the data in the publications reviewed here remain preliminary, and do not justify a conclusion that MMTV-LS, HPV, or EBV are causally associated with breast cancer. However, they form a valuable basis for redirecting future studies.