It's been quite awhile since I last read the Larson's work; but if my memory is correct, their results showed pretty wide variations in pressure even with motors they had cleaned the ejection out of and replaced with precisely measured amounts of ejection. It's a very interesting and frustrating problem trying to come up with an amount of ejection that is suitable for all sizes of rockets that the motor might be used in.
Regards,
Ed

Thanks, I'd overlooked that point. To my eye the variations in those tests were smaller, but still a factor of 2 or so.