Forty toddlers aged 20 to 24 months were presented with 32 pairs of images with the auditory stimulus Look followed by the name of the target image (e.g. Look . . . tree) in an intermodal preferential looking (IPL) paradigm. The same series of 16 items was presented first with one image as target and then with the other member of the pair as target. Half the children were given feedback, in the form of movement of the target image at the end of the trial, while the other half were presented with static images. IPL performance was quantified in terms of number of words showing at least 15% increase in proportion of looking time in the post-naming interval. Looking preference for the named item was correlated with parental report of vocabulary, this effect being stronger for those receiving feedback. The correlation with parental report of vocabulary comprehension was .65 for those receiving feedback, but only .37 for those with no feedback. It is concluded that the preferential looking task, which has been widely used in group studies, has the potential to act as a reliable index of comprehension level in individual children, especially when movement feedback is used to maintain attention.