Studies were designed to compare the effects of encoding strategies on two types of interference, sensory and semantic. Paired-associate lists were learned under instructional sets encouraging interactive imagery or repetitious rehearsal. Sensory interference was created by mispairing rhymes and semantic interference was produced by mispairing associatively related words. In control lists word pairs were unrelated. The results indicated that encoding strategy did not influence amounts of sensory and semantic interference. Mispairing rhymes produced the same amount of confusion as mispairing associates whether pairs were encoded through interaction imagery or through repetitious rehearsal. These effects were found with both short and long encoding intervals. The findings are inconsistent with the idea that encoding operations or strategies applied to word pairs completely determine the type of code activated for those pairs. Both sensory and semantic codes appear to be activated under each kind of encoding strategy.