The consensus standard for this sort of test is IEEE 1528, and you are absolutely correct. The test uses a gel phantom (the standard provides a few formulations

The consensus standard for this sort of test is IEEE 1528, and you are absolutely correct. The test uses a gel phantom (the standard provides a few formulations for different frequencies but most are a mixture of water, salt, sugar, and a thickener) and it is poked at with a sensitive probe to measure local temperature rise.