Research suggests that spaced learning, compared with massed learning, results in superior long-term retention (the spacing effect). Son (2010) identified a potentially important moderator of the spacing effect: metacognitive control. Specifically, when participants chose massed restudy but were instead forced to space the restudy, the spacing effect disappeared in adults (or was reduced in children). This suggests spacing is less effective (or possibly ineffective) if implemented against the wishes of the learner. A closer examination of this paradigm, however, reveals that item-selection issues might alternatively explain the disappearance of the spacing effect. In the current experiments, we replicated the original design demonstrating that an item-selection confound is operating. Furthermore, relative to a more appropriate baseline, the spacing effect was significant and of the same size whether participants' restudy choices were honored or violated. In this paradigm, metacognitive control does not appear to moderate the spacing effect.