We previously demonstrated mental rubbernecking during the simple cognitive act of refreshing a just activated representation. Participants saw two neutral and one negative word presented simultaneously and, 425 msec later, were cued to mentally refresh (i.e., think of) one of the no-longer-present words. They were slower to refresh a neutral word than the negative word (Johnson et al., 2005, Experiment 6A). The present experiments extended that work by showing mental rubbernecking when negative items were sometimes the target of refreshing, but not when negative items were present but never the target of refreshing, indicating that expectations influence mental rubbernecking. How expectations might modulate the impact of emotional distraction is discussed.