The spatiotemporally uneven sampling of paleomagnetic observations has remained a vexing problem in paleomagnetism. In analyses of Precambrian inclination data, a traditional method of binning the data has been based on the present-day geographic latitude-longitude grid and fixed-length intervals of geologic time. In this paper, using a simulation of synthetic and real paleomagnetic observations, we show that this method produces flawed estimates of averaged (binned) inclinations, leading to false inclination frequency distributions.