It strikes me that, in addition to the face-value interpretations given by the researchers, the subjects of some of these experiments could also be seen as rationally responding to incentives not to reveal their desires. The face attractiveness subjects might be afraid of embarrassing an authority figure or "messing up" the experiment. The split-brain patient might (rightly) think a truthful "I don't know" would be interpreted as evasive or hostile. The children might reason that being seen doing a rewarded activity "for free" would remove the basis for any future rewards.

The priming results don't seem to fit this pattern, though.