The Field Guide › Paper
A controlled behavioural experiment testing whether a Captain's risk-taking spreads to First Officers, and through what mechanism. Seventeen student pilots on the verge of qualifying decided, across 50 landing scenarios spanning four pre-rated risk levels, whether to continue an approach or go around — first alone, then paired with a professional Airbus A380 Captain (introduced via a scripted, uniformed handshake designed to establish a strong status gap) who was scripted to always choose to land except in the most extreme scenarios. The crew condition separated two moments: a pre-decision made before the Captain's choice was revealed, and a final decision made after. Two distinct effects emerged. First, First Officers raised their risk-taking in moderately risky scenarios the moment they were merely paired with the Captain, before his decision was known — a social-facilitation effect the authors read as wanting to look competent in front of a superior rather than compliance under direct pressure. Second, after learning the Captain had chosen to land in highly risky scenarios, First Officers raised their risk-taking further still, and the size of that reactive shift correlated with how authoritarian participants rated the Captain — but, notably, not with how trustworthy they rated him (rs = 0.109, p = .677). The study directly measured something close to cognitive trust — a post-induction 1–5 rating of the Captain's perceived trustworthiness, alongside authority, skill and kindness — and found it did no explanatory work at all: deference tracked perceived authority, not generalised trust in the Captain's competence or good faith. It also took First Officers measurably longer to decide to oppose the Captain than to go along with him, consistent with Bienefeld and Grote's finding that speaking up is the more effortful path. Even after being told the Captain himself had called a go-around in the most extreme scenarios, most participants insisted on continuing to land at least once anyway, a pattern the authors attribute to a Dunning-Kruger-style overestimation of their own piloting skill in pilots with essentially no real flight experience. The design deliberately maximises the status gap — a near-qualified trainee against one of the highest-status captain roles in commercial aviation — to isolate the effect, so the findings describe an upper bound on hierarchy's influence rather than a typical crew pairing, and the small, all-male sample (n = 17) limits how far the specific correlations should be trusted.