Diverse measurement contexts require estimates of time-varying quantities. Ideally the measurement device responds to signal variations more rapidly than the modulation of the signal itself. If so, then well-developed techniques may be used for the calibration and analysis of the measurement system. By contrast, as the characteristic timescales for signal modulation and measurement response become commensurate, the situation becomes more complicated; directly measured quantities may require correction for the finite bandwidth of the measurement system response. Toward this goal heuristic estimation rules have evolved over time and are now widely used. We rederive these common rules of thumb, and present sufficient conditions for their validity. Furthermore, we investigate their quantitative performance in cases for which these conditions are violated and encounter surprisingly poor results. As an alternative, we demonstrate that regularized deconvolution analysis exhibits more general quantitative utility at the expense of increased measurement burden and analytical complexity.