If one wants to stick to purely statistical parlance, then it is perfectly acceptable to refer to single element of a sample as un individu, even if it is of inanimate nature. Another possibility is unité statistique (see here and here.)
However, it is necessary to keep in mind that sample / échantillon has ambiguous meaning both in English and French:
- From statistical viewpoint it is a collection of items/individuals sampled from a population
- In specific fields (like physics, biology, medicine, etc.) it often means a single item / device that is an object of study: rock sample, urine sample, etc. If analyzed statistically, many such samples form what a statistician would call a sample from population.
As a more specific example, I could attest the use in computational biology, where the data taken from a single person are referred to as échantillon, and statistical analysis is performed on an ensemble/collection des échantillons (see also statistical ensemble.)
Thus, referring to a single item as échantillon and the collection as ensemble des échantillons is understandable in most fields - perhaps better than the proper statistical use of terms.
Update
Wikipedia article on Unité statistique says:
Les unités statistiques sont les éléments des populations, dans le langage des statisticiens. On parle aussi plus communément d'individus (les premières statistiques étant des études démographiques).
Selon le cas, l'unité statistique peut être un individu, un ménage, une entreprise, un établissement, une commune, un département, une région ou encore un pays.
Another place to look is the definition by Insee, which is the authority in France when it comes to collecting and analyzing statistical data (Insee = Institut national de la statistique et des études économiques.)