One of the fundamental purposes of educational research is to provide evidence to facilitate effective practice. However, the evidence itself does not have much value for practitioners unless key information about the context from which the evidence was generated is also provided. In this paper, we use the word 'context' to refer to factors that are relevant for learning, including the interactions that learners experience with multiple people, artefacts, and environments. Unfortunately, in many educational research studies, either these factors do not get the required attention or information about them is presented in an incoherent structure. The resultant lack of information leads to two significant drawbacks. First, it creates confusion among practitioners who want to apply research evidence in their practice. Second, it leads to research studies that on the face of it are similar, but that in reality have resulted from evidence that has been collected in significantly different contexts being included under the same categories in reviews, meta-reviews, and best-evidence syntheses. In this paper, we draw on the concept of 'relatability' of evidence and present taxonomy for collaborative problem-solving (CPS) that can be used to provide the valuable information against which research evidence can be indexed. By addressing the need for more detailed information about the contextual factors from which the evidence is generated to bridge the gap between research and practice in CPS research, we aim to exemplify the approach that is needed in educational research more generally.