Effective dose was introduced by the ICRP for the single, over-arching purpose of setting limits for radiation protection. Effective dose is a derived quantity or mathematical construct and not a physical, measurable quantity. The formula for calculating effective dose to a reference model incorporates terms to account for all radiation types, organ and tissue radiosensitivities, population groups, and multiple biological endpoints. The properties and appropriate applications of effective dose are not well understood by many within and outside the health physics profession; no other quantity in radiation protection has been more confusing or misunderstood. According to ICRP Publication 103, effective dose is to be used for “prospective dose assessment for planning and optimization in radiological protection, and retrospective demonstration of compliance for regulatory purposes.” In practice, effective dose has been applied incorrectly to predict cancer risk among exposed persons. The concept of effective dose applies generally to reference models only and not to individual subjects. While conceived to represent a measure of cancer risk or heritable detrimental effects, effective dose is not predictive of future cancer risk. The formula for calculating effective dose incorporates committee-selected weighting factors for radiation quality and organ sensitivity; however, the organ weighting factors are averaged across all ages and both genders and thus do not apply to any specific individual or radiosensitive subpopulations such as children and young women. Further, it is not appropriate to apply effective dose to individual medical patients because patient-specific parameters may vary substantially from the assumptions used in generalized models. Also, effective dose is not applicable to therapeutic uses of radiation, as its mathematical underpinnings pertain only to observed late (stochastic) effects of radiation exposure and do not account for short-term adverse tissue reactions. The weighting factors incorporate substantial uncertainties, and linearity of the dose-response function at low dose is uncertain and highly disputed. Since effective dose is not predictive of future cancer incidence, it follows that effective dose should never be used to estimate future cancer risk from specific sources of radiation exposure. Instead, individual assessments of potential detriment should only be based on organ or tissue radiation absorbed dose, together with best scientific understanding of the corresponding dose-response relationships.