In the context of language pathologies detection, linguistic skills are mostly assessed through standardized language tests used in formal communicative situations to assess patients' language and make a diagnosis. Yet the relevance of these subtests assumed to assess a linguistic level could be questioned. For instance, to assess lexical-semantic skills in oral production, most of the time the patient has to take a test in which he/she has to label a number of items. But only little information about the characteristics of these items and the way they have been selected are mentioned in the handbooks provided with the tests. However, questions can be raised concerning their representativeness and their relevance. The exploratory study presented here is organized around three main points: an inventory of lexical items used in different language tests for children aged 4;06 to 8 years old; a frequency list of lexical forms produced and heard by children in spontaneous situations; a comparison of the results gathered in the first two phases of the study, aiming at observing the degree of matching between the lexical forms used in standardized language tests and interaction corpora as well as the matching of their frequency, in order to carry out a reflection on the choices of lexical items and the relevance of these choices.