The list length effect in recognition memory has been the subject of recent debate. Many studies have identified the effect, however Dennis and Humphreys (2001) argued that previous list length effect findings were the result of a failure to control for four potential confounds. The list length effect can be used to discriminate between item and context noise models of recognition memory. Item noise models predict the effect, while context noise models do not. In this paper, the role of attention on the detection of the list length effect is explored. The attention task at study was manipulated; participants either rated the pleasantness of study items or read the words only. In addition, the design was either retroactive or proactive. The results suggest that it is the proactive design in which the list length effect is evident. When the retroactive design is used in conjunction with the pleasantness rating task, there is the most even performance across list lengths and a nonsignificant effect of list length. This is consistent with context noise models of recognition.