Seventy-two males received €10 (USD 13) to participate in a study on medication and decision-making. Exclusion criteria were medical or psychiatric illness, medication, smoking, and drug or alcohol abuse. Participants came in groups of six, and were seated in soundproof cubicles, randomly assigned to the oxytocin or placebo group (double-blind, placebo-controlled study design) and tested individually.
Participants self-administered, under experimenter supervision, a single intranasal dose of 24 IU oxytocin (Syntocinon-Spray Novartis; three puffs per nostril) or placebo. To avoid any subjective effects (e.g. olfactory effects), other than those caused by oxytocin, the placebo contained all the active ingredients except for the neuropeptide. The placebo was manufactured by Stichting Apothekers Haarlemse Ziekenhuizen in coordination with the pharmacy at the Amsterdam Medical Centre, adhering to the European Union guidelines on Good Manufacturing Practice and Good Clinical Practice. The placebo was produced using the exact recipes and procedures used by Novartis Inc. to produce the carrier of Syntocinon—the synthetic analogue of oxytocin. Placebos were delivered in the same bottles as Syntocinon. In short, the only difference between the placebo and the treatment was the absence versus presence of the active neuropeptide.
The experimenter left and participants completed computer-guided unrelated tests. Because effects of oxytocin plateau 40 min after administration [20
], the computer switched to the instructions for the main tasks after 30 min. To prime intergroup competition, we followed the procedures and tasks used in an earlier study (exp. 3 of [25
]). Participants were organized into two three-person groups and informed that they would engage in decision-making affecting the financial earnings of their own and the other group. Specifically, each member of each group would make a decision that affected their personal income, that of their fellow in-group members and that of the out-group.
The decision situation was structured as a prisoner's dilemma [27
]. Each participant had to choose between A (reflecting cooperation) and B (reflecting non-cooperation), and was informed that when he and the out-group representative chose A (joint cooperation), each individual in each group would earn €3; unilateral cooperation (A by participant, B by out-group) would result in €0 to each in-group member and €4 to each out-group member; unilateral non-cooperation (B by participant, A by out-group) would result in €4 to each in-group member and €0 to each out-group member; and joint non-cooperation (B by participant, B by out-group) resulted in €1 to each individual in each group. Thus, a cooperative choice would benefit the out-group at the expense of the in-group when the out-group would choose non-cooperatively, and benefit in-group and out-group equally when the out-group would (also) make a cooperative choice. A non-cooperative choice would protect the in-group against non-cooperation by the out-group, and benefit the in-group most in case the out-group would choose cooperatively. Accordingly, to serve one's in-group best, participants should make a non-cooperative rather than cooperative choice [25
Participants made three confidential decisions without feedback, with one decision being randomly chosen for pay-out to oneself, one's fellow in-group members and the out-group. On average, the instructions and decision-making took approximately 10 min. Confirming earlier findings [25
] across the three decisions, males given oxytocin were less cooperative (M
= 2.39) towards the out-group than males given placebo (M
= 2.88; directional t
(70) = 1.76, p
< 0.042; range 0 = non-cooperative to 3 = fully cooperative).
Following decision-making, participants read that in many competitions, players can choose who is in their team. Then they were shown 12 pictures, each on a new screen and randomized per participant. Participants indicated for each picture whether they would select the target for their team (0 = no; 1 = yes), as well as how dangerous and how useful the target was (both 1 = not at all; 5 = very much). We used six different pictures of faces morphed into low and high threat by Oosterhof & Todorov [29
; ). Although people rely on a multitude of cues when perceiving and interpreting faces [30
], Oosterhof & Todorov [29
] identified trustworthiness and dominance as the two orthogonal dimensions that are sufficient to describe face evaluation. Although face-trustworthiness is more sensitive to features signalling whether the person should be avoided or approached, dominance evaluation is more sensitive to features signalling physical strength/weakness. Threatening faces—the focus of the current experiment—should be both untrustworthy (signalling that the person may have harmful intentions) and dominant (signalling that the person is capable of causing harm). Although these computer faces are somewhat artificial, the advantage is that other features of the face (e.g. symmetry) can be kept constant, thus creating optimal conditions for a clean hypothesis test [29
]. Indeed, participants saw high-threat targets as more dangerous (M
= 2.75, s.d. = 0.67) than low-threat targets (M
= 2.01, s.d. = 0.62; F1,70
= 92.88, p
< 0.001). The lack of effects involving treatment (all F
< 1) confirms earlier research showing that oxytocin does not modulate the human ability to perceive emotional expressions [32
]. The total selection task lasted approximately 7 min. Upon completion of the experiment, participants were paid and debriefed.
Figure 1. Examples of (a) low- and (b) high-threat facial morphs used as targets (adapted from ).