Although some bright students in primary school are able to organise numerical data into classes, most attend to the characteristics of individuals rather than the group, and "see the trees rather than the forest". How can teachers in upper primary and early high school teach students to organise large sets of data with widely varying values into groups, and to represent the resulting grouped data appropriately in a histogram? How can they choose data sets that allow integration of different topics in the mathematics curriculum? What involvement can students have in the data collection? In a "teaching experiment" with two year 7 classes in two different Queensland schools over a three-week period, the authors report on their teaching investigation in which students explore the organisation and grouping of data. They highlight the type of thinking needed to represent grouped data in graphs. (Contains 3 figures.)