All radar power interpretations require a correction for attenuative losses. Moreover, radar attenuation is a proxy for ice-column properties, such as temperature and chemistry. Prior studies use either paired thermodynamic and conductivity models or the radar data themselves to calculate attenuation, but there is no standard method to do so; and, before now, there has been no robust methodological comparison. Here, we develop a framework meant to guide the implementation of empirical attenuation methods based on survey design and regional glaciological conditions. We divide the methods into the three main groups: (1) those that infer attenuation from a single reflector across many traces; (2) those that infer attenuation from multiple reflectors within one trace; and (3) those that infer attenuation by contrasting the measured power from primary and secondary reflections. To assess our framework, we introduce a new ground-based radar survey from South Pole Lake, comparing selected empirical methods to the expected attenuation from a temperature- and chemistry-dependent Arrhenius model. Based on the small surveyed area, lack of a sufficient calibration surface and low reflector relief, the attenuation methods that use multiple reflectors are most suitable at South Pole Lake.