Abstract  Using the Hebb Digits task, an incidental sequential learning paradigm, the effects of chunking of both the presentation and response phases of performance were examined. In the first experiment, consistent stimulus chunking increased learning, and performance was at an equivalent level to this when consistent chunking of both stimuli and responses was present. Consistent chunking of the responses alone did not significantly improve learning over a baseline condition where neither stimuli nor responses were chunked. The disruption of response organization in a second experiment, through a random response condition, failed to impact learning in non-chunked and stimulus chunked conditions. A third experiment found that response chunking did benefit learning in a condition where stimuli were presented in random chunks. A final experiment suggested extended processing of the digits could not account for performance gains in the stimulus chunking condition. Overall, the results suggest that the enhanced effects of chunking on learning were stimulus-driven rather than response-driven, except under conditions that constrained a consistent pattern of stimulus organization.