The principal rationale for the almost universal use of electronic fetal monitoring (EFM) in labor is that its use can effectively prevent the full expression of brain-damaging birth asphyxia by timely intervention in labor. This central hypothesis of EFM has never been adequately tested, in part because of the difficulty of obtaining a large enough sample. However, for this hypothesis to be true, 3 simpler conditions must be met. The first is that EFM be reliable, i.e., that substantial inter-observer agreement exist as to the identity and meaning of EFM patterns. The second is that EFM be valid, i.e., that one or more EFM patterns be statistically associated with the adverse neurological outcome to be prevented. The third condition is that if EFM and adverse outcome are related, the association is causal. A review of the literature indicates that none of these 3 conditions has been met thus far. Future research should concentrate on establishing the reliability and validity of specific EFM patterns; only those patterns truly predictive of adverse outcome and detectable by most practitioners are worth considering as indicators for intervention in labor.