Over the past 30 years successive governments in the UK have endeavoured to make the statutory framework suitable for children with special educational needs (SEN). More recently, efforts have been made to personalise children's learning, making educational experience more innovative and responsive to the diversity of needs in schools. A drive is emerging in health and education to develop and evaluate intervention strategies for children with language impairments (LIs), which is both methodologically challenging and rewarding. The current review demonstrates difficulties encountered with using evidence‐based (evidence‐related) outcome research involving children with LIs. Many studies have inherent methodological problems such as small sample sizes, ill‐matched groups and designs that are difficult to replicate or compare. Such approaches are unlikely to yield significant results, or if they do, it is difficult to devise clear guidance regarding choice of intervention strategies. In the light of these difficulties, theoretical, methodological and practical issues are discussed herein and a model is proposed to assist in enabling interventions to be identified, evaluated in a robust manner, and the results shared with educators. We suggest that the use of a process‐driven model ensures a more rigorous approach when undertaking large‐scale systematic, evidence‐based research into the effective approaches to teaching children not only with LIs but across the field of special needs education.