If you've heard of Stanley Milgram's obedience experiments, you've probably also heard the obvious objection: the "shocks" weren't real, the "victim" was an actor, and maybe — on some level — the subjects in the lab coat's room suspected as much. Chase Hughes's book raises this objection directly, then walks through a study built specifically to answer it.1
In July 1961, Stanley Milgram, then at Yale, began the experiments that would end up in nearly every introductory psychology textbook written since. Subjects were told to deliver increasingly severe electric shocks to a person in another room every time that person answered a question incorrectly. An authority figure — lab coat, clipboard, calm insistence — instructed the subject to keep escalating. The person supposedly receiving the shocks was, in fact, an actor experiencing nothing. What shocked observers wasn't that some subjects obeyed. It was how many did, and how far they went — continuing to deliver shocks even while protesting, even while visibly distressed, even while the "victim" wept in the next room.2
Charles Sheridan and Richard King's response to the "maybe they knew it was fake" objection was direct: remove the possibility entirely. They reran the paradigm with a real, non-human subject receiving genuinely uncomfortable — if not seriously harmful — electric shocks: a puppy. The book's summary of the result: half of the male participants and all of the female participants continued delivering shocks to the end of the protocol. Many showed visible distress. Some openly wept while continuing to comply.3
The puppy replication matters structurally, not just as a grimmer variant of the same finding. It closes the specific escape hatch critics had available for the original study — the possibility that subjects were, on some level, performing compliance for an experimenter they suspected wasn't asking them to do anything real. A puppy receiving a real, if mild, electric shock removes that ambiguity. If obedience to authority can produce this level of compliance even when the subject has no plausible reason to doubt the harm is genuine, the phenomenon isn't resting on some hidden social contract between subject and experimenter. It's something closer to what it looks like on the surface: authority overriding personal moral resistance, even against direct evidence of suffering, even in the population — women in this replication — that showed the highest compliance rate of any group tested.
The book pairs the puppy replication with a different kind of obedience study — this one about disobedience by proxy. In 1955, researchers staged an experiment at several intersections in downtown Austin, Texas: a model, dressed either in a sharp business suit and polished shoes or in wrinkled, soiled street clothes, would deliberately cross against a "do not walk" signal. The measured variable wasn't whether the model crossed — it was whether bystanders followed. When the model wore the suit, far more people at the intersection followed the illegal crossing than when the same model, doing the exact same thing, wore rough clothing. The book's own gloss: "Simply perceiving someone to have higher social status drastically altered the behavior of subjects at the intersections. Class counts."4
Where the puppy study is about direct obedience to an instructing authority, the Austin crosswalk study is about something adjacent but distinct: unconscious permission-granting based on perceived status, with no instruction given at all. Nobody told the bystanders to jaywalk. The well-dressed model didn't ask anyone to follow. The behavior spread purely through status-based inference — if a person who reads as high-status breaks a rule without consequence, the rule itself starts to feel more optional to everyone watching.
Both studies cited here predate 2015 by decades — the Sheridan and King puppy replication traces to the early 1970s literature on Milgram-adjacent obedience research, and the Austin crosswalk study to 1955. Neither is cited in this source with a modern replication study, a meta-analytic reassessment, or any indication that either finding has been checked against the post-2011 replication-crisis standards that reshaped how much of mid-twentieth-century social psychology gets weighted today. Per this vault's general source red-flag checklist, that combination — pre-2015, no cited modern replication — earns both studies a [PLAUSIBLE — needs corroboration] tag rather than [VERIFIED], regardless of how frequently either appears in popular psychology writing.
There's a second, smaller fidelity flag worth naming alongside the replication concern: the book cites the puppy study with the parenthetical "(Milgram, 2009)" — a citation that doesn't parse cleanly, since Milgram died in 1984 and the study itself is Sheridan and King's, not Milgram's own. The most likely explanation is that Hughes is citing a 2009 reprint, anthology, or textbook compilation that included Milgram-adjacent material without disambiguating original authorship in the in-text citation — a minor but real instance of the source's own inconsistent citation practice, already documented in this book's source stub, showing up in exactly the kind of place where precision matters most.
You don't need a lab coat or a crosswalk experiment to notice this mechanism operating around you — it shows up anywhere a group is deciding, in real time, whether to follow one person's lead. Watch a crowded sidewalk when the signal says wait. Notice who crosses first, and notice what they're wearing, how they're carrying themselves. Watch whether the second and third people to cross look like they're making an independent decision, or whether they're reading the first person's presentation as tacit permission.
Notice the same pattern in meetings, in group settings, anywhere someone breaks from an expected norm — speaks over time, disagrees with the room, does something mildly against the grain. Track how quickly other people follow, and correlate it, honestly, against how much perceived authority or status the first person carried into the room. The mechanism isn't limited to crosswalks. It's a general permission-cascade, and status is the fuel it runs on.
Evidence: Milgram's original 1961 obedience paradigm is among the most cited findings in social psychology, though even the original has faced serious methodological critique over the decades (ethical concerns about deception and distress, questions about experimenter-demand effects, and debates over how faithfully later "replications" actually reproduce the original protocol). The Sheridan and King puppy study and the Austin crosswalk study are both real, published findings, cited by name in the source, but neither is checked here against contemporary replication standards. [PLAUSIBLE — needs corroboration]
Tensions: The puppy study is presented as settling the "maybe they suspected it was fake" objection definitively — but it introduces a different confound the source doesn't address: a real animal in visible distress activates different (and arguably stronger) empathic and moral machinery than an adult human actor a subject has just met, which could plausibly increase the pressure to stop rather than serving as a clean like-for-like replication of the original human-subject paradigm.
Authority Costume documents how a tactical operator deliberately constructs the visual signals of authority or status — clothing, bearing, props — specifically to trigger compliance without needing to justify the underlying instruction. The Austin crosswalk study is a direct, real-world instance of exactly this mechanism operating passively: the well-dressed model didn't perform an authority costume technique deliberately, but the effect on bystanders was structurally identical to what an operator using the technique on purpose would be trying to produce.
The insight the pairing produces: this page explains why the costume mechanism works at the population level — because status perception genuinely does license rule-breaking-by-association in observers, not just in a lab, but on a real street corner with real jaywalking consequences on the line. The authority-costume page shows an operator exploiting that same effect deliberately. Read together, they suggest the costume technique isn't manufacturing a new vulnerability — it's aiming a pre-existing, well-documented social-inference mechanism, one that operates whether or not anyone in the room intended to trigger it.
The Autopilot Mode and Three Hijacking Methods documents a framework for deliberately disrupting habituated behavior to install directed compliance — built on the premise that people run on automatic, role-based scripts until something breaks the pattern. Milgram's subjects and Sheridan and King's puppy-study subjects weren't in an autopilot state in the technical sense this other page describes — they were fully alert, distressed, and consciously aware they were causing harm, and continued anyway. The insight the pairing produces: obedience-driven compliance and autopilot-disruption compliance are two genuinely distinct routes to the same behavioral outcome. One works by exploiting inattention; the other works by installing a stronger, socially-sanctioned instruction that overrides a fully-conscious, fully-distressed moral objection. The puppy study is arguably the more disturbing of the two precisely because it shows compliance surviving full conscious awareness of harm — autopilot bypass isn't even a necessary ingredient.
Sharpest implication: If the puppy replication holds up, the uncomfortable finding isn't really about obedience in the abstract — it's that the population showing the highest compliance rate in that specific replication was women, the opposite of whatever intuition might predict about who's "more likely" to override an instruction to cause harm. Any account of who obeys and why that leans on assumed gender-based empathy differences runs directly into this finding, unresolved, sitting in a popular-practitioner book that cites it without much comment on the implication at all.
Generative Questions: