This study examines J. H. Flavell, S. G. Shipstead, and K. Croft's (1978) finding that 2 1/2-year-old children can hide an object behind a screen but cannot achieve the same result by placing the screen in front of the object. Experiment 1 replicated this finding alongside a task in which children judged what a person in a picture was looking at. Performance on the move-object task approached ceiling; performances on the move-screen and looking-where tasks were highly correlated even after age and control task performance were partialed out (r = .54, p <.01). Experiment 2 examined whether the finding resulted because the object was more interesting to manipulate than the screen. The move-object task remained easier than the move-screen task with an interesting screen and a dull object. The move-screen task again correlated specifically with the looking-where task. Results are explained in terms of engagement, a precursor to a mature understanding of attention.