Minimizing conflicts with humans is a necessary component of the management of American black bears (Ursus americanus) across most of their range. The number of complaints about conflicts with black bears is commonly used to infer trends in the actual frequency or severity of human–bear conflict, and even trends in bear population size. However, the number of complaints received by management agencies is a function of both the frequency of and the reporting rate for conflicts, and the reporting rate may change over time. We tested for effects of food availability, numbers of bears harvested, and management regime changes on 3 measures of human–bear conflict: (1) public complaints, (2) traps set to capture bears involved in conflicts, and (3) bears killed in defense of property in Parry Sound, Ontario, Canada, 1992–2008. All measures of human–bear conflict were inversely related to food availability. Complaints increased following a controversial change in management (cancellation of the spring hunting season), but numbers of traps set and bears killed were not affected. We suggest that an increase in the reporting rate was largely responsible for the increase in complaints following the spring hunt cancellation because (1) an effect on the actual frequency or severity of human–bear conflict should also have been detected in data for traps set but was not, and (2) neither the number nor the sex ratio of harvested bears changed when the spring hunt was cancelled, so the effect of harvest on population size and sex ratio was not altered by the management regime change. Trends in the actual frequency and severity of human–bear conflict should not be inferred from trends in complaint data unless factors that could affect the reporting rate for conflicts are accounted for.