We describe an approach to aiding the design and interpretation of experiments involving biological effects of weakly interacting electromagnetic fields that range from steady (dc) to microwave frequencies. We propose that if known biophysical mechanisms cannot account for an inferred, underlying molecular change signal-to-noise ratio, (S/N)gen, of a observed result, then there are two interpretation choices: (1) there is an unknown biophysical mechanism with stronger coupling between the field exposure and the ongoing biochemical process, or (2) the experiment is responding to something other than the field exposure. Our approach is based on classical detection theory, the recognition that weakly interacting fields cannot break chemical bonds, and the consequence that such fields can only alter rates of ongoing, metabolically driven biochemical reactions, and transport processes. The approach includes both fundamental chemical noise (molecular shot noise) and other sources of competing chemical change, to be compared quantitatively to the field induced change for the basic case that the field alters a single step in a biochemical network. Consistent with pharmacology and toxicology, we estimate the molecular dose (mass associated with field induced molecular change per mass tissue) resulting from illustrative low frequency field exposures for the biophysical mechanism of voltage gated channels. For perspective, we then consider electric field-mediated delivery of small molecules across human skin and into individual cells. Specifically, we consider the examples of iontophoretic and electroporative delivery of fentanyl through skin and electroporative delivery of bleomycin into individual cells. The total delivered amount corresponds to a molecular change signal and the delivery variability corresponds to generalized chemical noise. Viewed broadly, biological effects due to nonionizing fields may include animal navigation, medical applications, and environmental hazards. Understanding necessary conditions for such effects can be based on a unified approach: quantitative comparison of the estimated chemical change due to a particular electromagnetic field exposure to that due to competing influences, with both estimates based on a biophysical mechanism model within the context of a model of a biological system.