Which is more problematic, random sampling or non-sampling error and why?

Non-sampling error is more problematic because it is more difficult to detect and identify. These errors occur when data are collected improperly, recorded incorrectly, and/or reported inaccurately. Sampling errors, on the other hand, can be detected, identified, and corrected; they are only caused by a sample that is not representative of the population being studied. As such, non-sampling errors are more dangerous and can lead to significant inaccuracies that skew the results of research.