Recent research invokes preference imprecision to explain violations of individual decision theory. While these inquiries are suggestive, the nature and significance of such imprecision remain poorly understood. We explore three questions using a new measurement tool in an experimental investigation of imprecision in lottery valuations: Does such preference imprecision vary coherently with lottery structure? Is it stable on repeat measurement? Does it have explanatory value for economic behaviour? We find that imprecision behaves coherently, shows no tendency to change systematically with experience, is related to choice variability, but is not a main driver of the violations of standard decision theory that we consider.