NotPegasus
Given a population of only about 500 objects, why examine carefully only 5% of them? Why take a chance that the sample does not represent the population? Why examine in a way that undermines confidence in your findings? Do they lack the resources to do the job?
They are not performing an exercise in an undergraduate stats class; they are interested in what caused a single instance (namely the downing of an airliner over Ukraine 4 months ago. The composition and statistical properties of a population of recovered fragments is not particularly relevant or interesting to the investigators (one would presume that most originated on the aircraft). Their objective is to find and identify at least one fragment that originated from outside the aircraft. Thus, a sampling strategy that maximises the probability of analysing such fragments (e.g. picking iron fragments and ignoring Alu ones) is an appropriate one in this case.