Exploring the Potential of Matching-to-Sample-Based Training Protocols to Enhance Language Abilities in Individuals with Language Impairments: A Scoping Review.

Every study was small. Sixteen studies shared 81 participants, and several were single-case designs with two, three or four people. No study was randomised and none was compared against another approach such as ordinary explicit teaching, so there is no way to tell whether matching-to-sample adds anything. The authors point out that every included study reported success, which raises the possibility that studies with null or negative results were never published. Few studies followed participants past the immediate post-training test, so whether any gains held up or transferred to everyday talking is unknown. Participants' baseline language and thinking skills were low, so training reached only foundational targets such as single words and naming, not grammar or conversation. The studies varied so widely in who took part, what was trained, how prompts and feedback were given and which relations were tested that the reviewers could not say which set-up works best. No participant had developmental language disorder, which is the population the review set out to inform. In most studies the paper does not say who delivered the training.

The authors declare no conflicts of interest. No funding source is reported. Ethical approval was not required because the work reviews published literature.

Do not read this review as evidence that matching-to-sample training works for children with developmental language disorder. Not one of the 16 included studies recruited anyone with DLD, none was a randomised trial with a control group, and the only skills trained were foundational ones such as single-word naming, so anyone with a DLD caseload should treat this as a description of a method awaiting testing, not a treatment to adopt.