This would have eliminated the backlash. However, I like that PG seems to live by what he wrote about in "What you can't say" and strives for perfect accuracy in his observations. While problems in communication might be the root cause (I'm almost sure they are, why is it surprising that if you can't communicate you'll do worse? Change "accents" to "can't speak English - the language of the country they are in" and no one would argue), PG was perfectly accurate in describing his observations in the data. Isn't it better to be open about this, along with him noticing his selection bias? Being aware of biases is better than ignoring them.