Testing with children is very common
Testing with children is very common. If you can do tests with animals and babies, obviously you can test with children. However, as PhilipW points out in his comment, not all ages are the same. A common way to differentiate them is 4-6 (pre-reading), 7-10 (early schooling), 11-14 (preadolescence), 14 and over. At least that's the way it is in my country, based on learning cycles. There may be some minor differences in other countries, but the important thing is to consider these age ranges (unless you consider Vygotsky theories of learning which put more emphasis on sociocultural aspects, but that would mean a lot of work to analyze, a lot of work at screening time and you'll need really complex statistical methods).
Is it possible to do low fidelity tests with children?
I think the answer is NO. At least not until 11-14 years old. Low fidelity tests require a very complex degree of abstraction for children in early childhood, who acquire this level of abstraction from 10-11 years. Therefore, doing low fidelity tests will generate false results (or no results at all).
I want to clarify that I have never done such tests and maybe I'm completely off base, I simply base the answer on known theories and previous test we amde with children (just never low fidelity test). For example, for Piaget abstract thinking is only achieved after having passed the sensorimotor stage and concrete thought, and this occurs at approximately 12 years, where it progressively develops until it reaches its consolidation at around 15 years.
the sensorimotor stage, from birth to about age 2; the preoperational stage, from age 2 to about age 7; the concrete
operational stage, from age 7 to 11 and the formal operational stage,
which begins in adolescence and continues into adulthood.
That is why when teaching children they are shown concrete objects or graphic representations, which is the exact opposite of a low fidelity test.