Great question! If participants were expressing gratitude to the researchers for the massage, then decison-maker 1's behavior should also have been affected, but it was not. Only decision-maker 2s showed a behavioral effect of massage. In addition, the change in oxytocin strongly predicted the extent of monetary sacrifice in decision-maker 2s. If this was "gratitude" why would oxytocin predict behavior? Lastly, we took many precautions like shielding identities to mitigate the experimenter effect. Indeed,our analysis showed that we changed the oxytocin release patterns for those who were massaged, suggested a true physiologic cause for the resultant change in behavior. See the original article for details at