An easily implemented extension of the standard response method of tidal analysis is outlined. The modification improves the extraction of both the steady and the tidal components from problematic time series by calculating tidal response weights uncontaminated by missing or anomalous data. Examples of time series containing data gaps and anomalous events are analyzed to demonstrate the applicability and advantage of the proposed method.