If 1% of the total population is the expected value, the range of likely outcomes for the number of stutterers you meet out of 1000 people is approx. 4-16. Meeting (3 or fewer) or (17 or more) stutterers in a group of 1000 is highly unlikely to occur by random chance (less than 5% probability). And so it's highly likely that the expected value of 1% is inaccurate or there's some other factor at play.