I think it might be a reference to the NYPD Firearms Discharge Report for the year 2000, which showed, for gunfights (an incident with shots fired by both officers and perps), officer accuracy rates of 9% (16 hits for 185 shots fired) and perp accuracy rates of 17% (7 hits for 42 shots fired).
That year may have been particularly bad, and an aberration for the NYPD, though. There was a 2008 RAND study that said “Between 1998 and 2006, the average hit rate was 18% for gun fights.” But that was for the NYPD, the RAND study doesn’t appear to give accuracy stats for the other side of the gunfight (the perp).
The NYPD seems to be the main agency studied — they are the largest police force in the US and seem to keep copious statistics.
The NYPD also fits their issued sidearms with triggers that are designed to take additional pressure to pull, which was described to me as a "lawyer's trigger" ie. the only way that gun could have possibly fired is because you wanted it to.
The RCMP S&W 5946 has a trigger weight of between 10-12lbs. I can't speak for the NYPD, but they're probably in a similar range.
I have about 10 years of shooting experience, and I find it difficult to shoot with CONSISTENT accuracy even under ideal conditions at a range. When relaxed and focused, it is a pleasure to shoot. Any time I've taken it out to play for keeps? I've never been relaxed, nor focused on the principles of marksmanship.
EDIT: NYPD Trigger pull
Because I know sources are better than hearsay. I can't find any source on the RCMP trigger besides my own experience and training.