This paper argues for the use of conversation analysis (CA) and membership categorisation analysis (MCA) (Sacks, 1992) to investigate children’s social conduct. A majority of prior research in this area has tended to focus on limited theoretical perspectives situated in developmental psychology, resulting in a dichotomous presentation of either prosocial or antisocial behaviour (see Bateman & Church, 2008 for an overview). Although the use of predefined categories “antisocial” and “prosocial” may be helpful for the organisation of data, there is a concern that these pre-defined classifications lead to children themselves being categorised as either consistently prosocial or antisocial (for example Nelson & Crick, 1999). This view encourages stigma and the labeling of children rather than offering further insight into children’s social worlds (Bateman & Church, 2008). This problem represents a shortfall in information regarding the complexity of peer interactions and how they are locally managed by the children themselves, disregarding the range of social competencies engaged in by the participants. Therefore a shift in theoretical approach is argued for here as this informs of how social order is produced through verbal and non-verbal communications between the participants themselves (Butler, Fitzgerald & Gardner, 2009; Sacks, 1992a; 1992b;). Analyzing children’s social conduct through observing their communication offers an innovative, theoretical shift which is becoming more valued in many different areas of early childhood and particularly for the study of social relationships in education. This paper will outline the concept of communication as perceived from an ethnomethodological (EM) perspective, provide a background to EM and conversation analysis (CA), discuss some findings from research and then discuss the practical application of these findings for practice.