You are correct, that study (and it was a study, not a "study", by some of the most respected experts in the field) only included 18 brown & 7 black bear encounters where the bear was exhibiting aggressive behaviour.

These were the only known encounters in Alaska over the 2 decades the study covered where bear spray was employed.

However, I'm having a hard time figuring what percentage of actual attacks these 25 encounters would represent. It looks like there were around 12 fatal attacks in Alaska during the course of the study, but I can't seem to find anything on non-fatal attacks.

Do you know how many attacks happened during the study's time frame ('85- '06)? Given only 12 fatal attacks occurred I'm not sure how insignificant 25 attacks would be in the overall pool of attacks.

That said, these results are consistant with other, similar research performed. The US Fish & Wildlife fact sheet on this topic mentions 2 of these other sources:

"Law enforcement agents for the U.S. Fish and Wildlife Service have experience that supports this reality -- based on their investigations of human-bear encounters since 1992, persons encountering grizzlies and defending themselves with firearms suffer injury about 50% of the time. During the same period, persons defending themselves with pepper spray escaped injury most of the time, and those that were injured experienced shorter duration attacks and less severe injuries. Canadian bear biologist Dr. Stephen Herrero reached similar conclusions based on his own research -- a person’s chance of incurring serious injury from a charging grizzly doubles when bullets are fired versus when bear spray is used."

On the other side, I haven't found anything suggesting the conclusions drawn by these various studies are incorrect.

As a side note, I haven't found the underlying studies mentioned in that fact sheet yet, but I have emailed the USFWS to see if they can point me in the right direction so I can see the results myself.