Dean is the writer of the article in the link from the initial post.

In that article he says of Herraro-

"The often cited Efficacy of firearms for bear deterrence in Alaska by Tom S. Smith, Stephen Herrero, and others, included 37 instances of a handgun being present when a bear attacked a human. The instances collected were from 1883 to 2009. They recorded 6 failures to stop the attack out of the 37 instances. That is an 84% success rate. Pistol and ammunition technology have greatly improved since 1883.

The authors of the Efficacy of firearms have not released their data. There could be as many as six instances of overlap between the Efficacy of firearms data set and our collection, so a combination of the data is not useful unless the Effficacy of firearms data set is released. We cannot know how many of the six “failures” of the efficacy study might be because the handgun was never attempted to be used, was unable to be accessed because it was buried in a pack, or for other reasons."