"10 million words recorded during just under 200,000 trials"

So an average of 50 words per trial? I would have expected a higher number.