I don't agree with this. On the surface, a study of 107 incidents seems small but in reality that is the nature of bear attacks. They are a very low-probability event so we're never going to get a sample of multiple thousands to study (at least not for a very long time).

To put the numbers in context, in Smith & Herrrero's study on firearms in Alaska, they only had 269 incidents to look at ... and that study covered bear-human conflicts involving firearms from 1883 - 2009. The bear spray study spanned 1985 - 2006.

I understand the Bear Examiner article says this limited amount of a data is due to cherry picking by citing the 2 thousand odd cases in the Miller & Tutterrow study (interesting read by the way, worth checking out). This claim seems off-base, if not simply disingenuous.

The Miller & Tutterow's study was looking at something quite different than Smith & Herrero were. They were comparing bears that were killed in defence of life or property in Alaska to sport kills, to try and better understand the differences. This makes a pretty large difference because the 2 most prevalent reasons for the killing of the bear were "Bear was thought to be dangerous" and "To protect property." Secondly, the study notes that for "both species of bear, the person shooting the bear most frequently described their activity as at home or dwelling." So, this study seems to encompass a large number of incidents that seem to fall outside of the scope of what Smith & Herrero were looking at in their firearm study.

Finally, the biggest difference I think, is that Miller & Tutterow were only studying cases where the bear died ... these are by default successful uses of firearms.

In the full context of their study, the less than 2% injury rate of humans makes sense ... these were all successful uses of firearms by people, often at home, and who, for the most part, weren't under imminent threat. The authors recognized and included in their report, stating: "However, when injuries or deaths from bears occurred, the bear frequently was not found, so these data under represent the frequency of injury to people from bears."