Abstract
This paper considers the skills needed for data representation. A framework of transnumerative techniques that facilitate data representation is proposed and applied to the responses of 73 year 7 students to two tasks involving association. Students? responses were classified according to their levels of success in representing the association, and the types of techniques used. It was found that while students had techniques for representing data, their choices of graph type were not always suitable, and they overlooked simple techniques such as ordering and grouping data that could have made their representations clearer. Implications for doing and teaching data analysis are discussed.