One major concern with value of study noted above was that lack of propellant might dramatically alter results from what would occur in "real" loads. Typical boiler-room-filling charges occupy about 60% of available space in case, air fills remainder of boiler room (within interstitial pore spaces). At first consideration, it seem obvious that adding charge will increase compression within boiler room, in response to primer gases, hence, it should increase pre-ignition bullet movement.
However, reality is not so easy as this simplification would suggest. First, granules offer a relatively large amount of cool surface, which interacts with hot primer gases to cool those. Besides the immediate and direct pressure loss associated with cooling, cooling also results in condensation of primer gases, with a resultant significant pressure reduction within boiler room of case. Hence, if granules provide a sufficient heat sink, adding charge might not increase maximum primer-induced pressure within boiler room and could even reduce primer-induced pressure!
With regard to the above noted study, I had the foresight to save the remaining primers in all those boxes, I even have more of the bullets; one wonders why I did not have the foresight to save that test case in a place where I could find it as easily! Because I did not, I cannot readily repeat that test, using added simulant with all other characteristics held constant.