Two groups of children, one with good attention and low hyperactivity and the other with poor attention and high hyperactivity, were formed on the basis of teachers' ratings. The groups were tested on a large number of tests of attentional performance, covering selective attention, divided attention, sustained attention and executive function. Analysis of group differences, with verbal and non-verbal intelligence controlled statistically, indicated that the groups demonstrated significant differences on a variety of measures. Principal components analysis of the tests of attention isolated two components, one of which significantly distinguished the two groups. It is suggested that this component reflected aspects of executive control of attention. The measures which discriminated the groups most clearly were efficiency of visual search involving two targets in alternation, the number of correct responses on a specially devised version of the Wisconsin Card Sorting Task and a task measuring the ability to inhibit a dominant response. It is suggested that all these tasks required a combination of selective attention and executive control. It was also noted that the poor attention group produced bigger individual differences on all measures, indicating that this group may include a variety of weaknesses which require more precise analysis to distinguish them.