I think you are mistaking the concept of "scientific" with some sort of formal report.

The "rocketing" of the can is important only in the sense that a can may puncture a fire wall. It is the ability of the can to toss liquid across a room either as a stream or as a slug of particles that is important. The liquid goes in the opposite direction from that of the can.

The evidence in the references I posted, coupled with what can be calculated from the Bernoulli equation, known rupture pressures and known mechanisms of rupture make it self-evident that spray cans are capable of creating scenarios that look like multiple origins.

In the work of Michael Fox, pay particular attention to the video of the wasp spray can which is subjected to heat in the absence of an ignition source. Note that the liquid is shotgunned almost instantly over the space between the can and the target blanket. The liquid strikes the target blanket as a saturating blob. If you were to add a fire in the vicinity of the can, you would wind up with two apparent origins, one near the can and one on the blanket.

Cans today are probably more dangerous on average than they were in earlier years because of the replacement of freon by LPG in many products.

Manufacturers associations exist for the purpose of convincing the world that their products are safer than they really are and to minimize the effect of government regulation. The cigarette, plastics and smokeless powder industries have taught us this lesson.

One other point: The difference between "flammable liquid" and "combustible liquid" becomes less important in aerosol cans. When a can ruptures, the high velocity of the liquid creates a mixture of liquid and mist. The mist, will have the ignition properties of vapor from a flammable material even if it originates from a combustible liquid.