Authority bias

Milgram’s obedience experiment 1961-1963 (published in 1963)#

Biases at play: obedience to authority, displacement of responsibility.

What happened. At Yale University, Stanley Milgram recruited volunteers for an alleged study on “memory and learning.” An experimenter in a gray lab coat asked the participant, in the role of “teacher,” to administer increasingly strong electric shocks to a “learner” (an accomplice) for each wrong answer. There was no real shock, but the learner screamed, protested, then fell silent. When the participant hesitated, the experimenter responded with standardized prompts (“Please continue,” “You have no other choice, you must go on.”). In the study published in 1963, 26 out of 40 participants went all the way. In later variants (compiled in 1974 in Obedience to Authority), obedience depended heavily on the situation: it dropped sharply when the experimenter gave instructions by phone, when the experiment took place in non-prestigious facilities, and it became rare when two other “teachers” refused to continue.

Tenerife: the flight engineer speaks up, the captain decides March 27, 1977#

Bias at play: hierarchical authority in a cockpit.

What happened. On March 27, 1977, planes were diverted to Los Rodeos Airport in Tenerife (Canary Islands), which was congested and engulfed in fog. A KLM 747 began takeoff while a Pan Am 747 was still on the same runway. The cockpit recorder shows that the KLM flight engineer had doubts: he asked whether the other plane had really cleared the runway. The captain, chief flight instructor and one of the airline’s most senior and respected pilots, insisted yes and continued. The collision killed 583 people. Investigators identified several causes, including ambiguous radio communications, fog, and pressure from duty-time limits; the authority gap in the cockpit was one of them. It led to crew resource management (crew resource management) training, which teaches crews to object and captains to listen.

Elf’s “sniffer planes” (France) 1976-1979 ; revealed in 1983#

Bias at play: authority bias (with wishful thinking).

What happened. Two inventors claimed they could detect oil deposits from an airplane thousands of meters underground; the state-owned oil group Elf funded them for three years for hundreds of millions of francs. The process was a hoax. The process was endorsed by group executives and top-level public figures, and, according to published accounts, the case was taken all the way to the Presidency of the Republic. That endorsement discouraged counter-expertise for several years: challenging the demonstration meant contradicting those who had approved it.

The full account, with sources and limitations, is in the Wishful thinking page.

The contaminated blood case (France) 1983-1999#

Bias at play: authority bias (context, framing).

What happened. In the early 1980s, hemophiliacs and transfusion recipients were infected with HIV through blood products. As early as 1984-1985, product heating (virus inactivation) and donation screening were known and available. On May 29, 1985, at an internal meeting of the National Blood Transfusion Center (CNTS), its director stated that all batches were contaminated; distribution of unheated products nonetheless continued, and was not truly stopped until fall 1985, according to the Senate inquiry commission, which estimated around 1,200 infected hemophiliacs (with infected transfusion recipients, according to the commission, being four times higher). Systematic screening took effect on August 1, 1985. In 1992, the Paris criminal court convicted the CNTS director on October 23 for deception; in 1999, the Court of Justice of the Republic acquitted two former government members and found a former junior health minister guilty of breaching an obligation of prudence or safety, with exemption from punishment.

Where the bias appears. The Senate report describes a “sacralized” transfusion system: a prestigious institution, surrounded by illustrious names, which the medical profession allegedly trusted without really knowing it. According to the commission, the system was “never challenged nor even contradicted” and doubt “had hardly any place in it”; it writes that “it is difficult to doubt an institution one trusts and to which one delegates responsibility,” and its foreword challenges the “blind assertions of famous or powerful physicians.” This is exactly the mechanism of authority bias: an alert is judged by the prestige of who speaks, not by arguments. What is established is these parliamentary findings; the label “authority bias” is our reading grid, and it explains only part of the case: industrial and financial stakes, rivalry between screening tests, system monopoly, and fragmented responsibilities weighed heavily, and the CNTS director was convicted for deception—that is, deliberate behavior. Level of evidence: established. A parliamentary report based on hearings itself describes deference to the institution and its figures.

Élysée wiretaps: “on orders and with eyes closed” (France) 1983-2008#

Bias at play: authority bias (orders from the top of the state), among other causes.

What happened. From 1983 to 1986, a “coordination, information, and anti-terrorism action mission,” nicknamed the “Élysée cell” and made up of gendarmes, DST officials, and general intelligence police, carried out phone surveillance of public figures (journalists, lawyers, artists) and intercepted their contacts, on instructions from the deputy chief and then chief of staff to the President of the Republic. No law then provided for these administrative interceptions, later regulated by the law of July 10, 1991. After press revelations (1992) and seizure of archives (1997), the Paris criminal court ruled on November 9, 2005; according to reports, it identified François Mitterrand, deceased in 1996 and never prosecuted, as the inspirer and decision-maker for essential elements, and convicted former aides for invasion of privacy. Defendants did not appeal these convictions; the court of appeal (March 13, 2007) ruled on civil interests, and the Court of Cassation dismissed appeals on September 30, 2008.

Where the bias appears. Defendants invoked as justification command by legitimate authority—that is, orders from the top. The court of appeal, upheld by the Court of Cassation, rejected the argument: even assuming the order was given, it came from an authority without the power to give it, and notification of an illegal order from the President “does not exempt, any more than an order from a hierarchical superior, a civil servant […] from his obligation of loyalty to constitutional principles.” The ruling reports one defendant saying he signed wiretap requests “on orders and with eyes closed,” and notes that the control procedure was “often purely formal.” This is the authority-bias pattern: the source of the order substitutes for verification. But judges do not speak in psychological terms: they retain intent and stress that “political leaders knowingly took the risk of letting this system persist,” which points to political choices and career interest as much as deference. Level of evidence: accepted, not proven. The invoked order and execution without review are established; reading them as authority bias, rather than as deliberate choice, is an interpretation.

PIP breast implants: CE marking stands in for oversight (France) 2001-2018#

Bias at play: authority bias (certification, context).

What happened. Poly Implant Prothèse (PIP), founded in 1991 by Jean-Claude Mas, placed its silicone-gel implants on the market on April 18, 2001. A gel not compliant with what was declared during certification was used for years. According to the DGS/AFSSAPS report, authorities did not inspect the company between 2001 and 2010: regulatory oversight was provided by the notified body chosen by the company, whose annual reports were not sent to the agency. Device-vigilance rupture reports increased starting in 2008; an inspection took place March 16–18, 2010 and the agency suspended implants on March 29. The report concludes that implanted women, healthcare professionals, and authorities were victims of organized deception by company leadership. PIP executives were convicted on December 10, 2013 in Marseille; the founder was sentenced to four years in prison, a sentence upheld on appeal (May 2, 2016) and made final on September 11, 2018.

Where the bias appears. Prescribers and clinics implanted CE-marked devices, issued by a private body; authorities, for their part, relied on this delegated oversight and on device vigilance. It is established that certification stood in for oversight; it is interpreted, not demonstrated, that trust in this stamp delayed doubt. The dominant explanation is deliberate and organized fraud, which, according to the criminal ruling, also deceived the certifying body; the report even notes that “it is not obvious that an unannounced inspection would have uncovered this fraud.” Level of evidence: accepted, not proven. Delegation of oversight to a body chosen by the manufacturer is documented; authority bias is only the reading grid.