Zentall (2008) challenges Arantes and Grace’s (2008) failure to replicate Clement, Feltus, Kaiser, and Zentall (2000) by suggesting that our results may have been due to insufficient training or to subjects’ experimental histories, and that our results are actually consistent with those of Clement et al. when examined closely. On the contrary, our pigeons received more training than Clement et al.’s did, and when an overall measure of preference on test trials was calculated, independent of the effect of initiating event that we reported, there was no evidence of preference for the stimulus preceded by the greater response requirement in our data. Although there have now been two unsuccessful attempts to replicate Clement et al.’s work ethic effect, there is evidence that in some situations, the value of a stimulus, as assessed by transfer tests, varies inversely with the context of reinforcement. It is important to look for convergent results using other procedures and transfer tests, such as resistance to change, to identify the principles that determine when contrast manipulations affect value and when they do not.