Crude ballistic testing is usually too crude.

For small handguns, I see tests into gelatin, wet newsprint,
and water jugs covered with denim.  {I have not seen much
on ballistic clay.}     But these tests only compare loads,
and are not realistic simulations for self defense.

I use poplar wood 2" cylinders as an easy compromise, but ideally,
there should be a layer of leather, followed by something like a human
rib/sternum, followed by gel.
A layer of auto door glass would be a good addition at times.

The leather is easy, and 1/4 inch plywood might do for the bone.
Gelatin is good for soft tissue, but to avoid the fuss, wet newsprint
might do, or even saturated basswood (which might be difficult).

This really matters for pocket guns, which are marginal weapons.
I won't routinely carry a loaded gun that weighs a full pound, or is
much larger than a KT p-3at/Ruger LCP/Taurus TCP.
And If a hollow/soft point breaks up in leather/bone, the perp effect is poor.

Gel and newsprint tests indicate expansion for 32acp and 380 soft/hollow points.
But I would only trust full copper or steel jackets, with rounded-point, in these loads,
since they are more likely to survive the trip to damage vulnerable soft tissues.

A broken rib or arm does not stop a perp.
A broken femur/tibia might knock them down.
But punctured bronchi, nerve tissues, or large blood vessels ruin their plans.

The same might go for 22mag, but that's a better penetrator whatever the bullet.
Unless you carry 9x19+p, 357, 40, 45, or similar-power loads, fmj is more reliable.

If a test result firing at 90 degrees is marginal, then a 45 degree test will fail.
.