When the Title Eats the Person — Hide the Name and the Review Changes
First a person wears the title. Worn long enough, the title eats the person. The harder the material, the more the name stands in for the check. Questions stop there.
Here is someone whose path never once bent.
School, post, promotion, all on schedule, and never below what the people around him had set as the mark.
The sharper statement is the other one. He has never once drawn a version of himself outside that mark.
The axis of the question shifts there. From what to do, to how to appear.
A person wears the title. Worn long enough the order inverts, and the title wears the person.
Advice about when to step back means nothing to someone who calculates that when the applause stops, he stops. To him it is not an exit. It is erasure.
One title overrode four rules
Begin with the fact that this has been measured, not merely felt.
In 1966 Hofling’s group ran an experiment in a hospital. A nurse on night duty takes a phone call. The caller identifies himself as a doctor and orders a drug given to a particular patient.
Four things were wrong with that order.
| What it violated | |
|---|---|
| ① | the drug was not approved for the ward |
| ② | the dose ordered was twice the stated maximum |
| ③ | a telephoned order was itself against protocol |
| ④ | the nurses had never met the caller |
Twenty-one of twenty-two set off for the patient’s room with the drug.
Not from lack of knowledge. The nurses knew the drug was dangerous. One title pressed down on four rules at once.
The harder the material, the more the name explains
The effect strengthens as the thing being judged gets harder.
When the content is hard to follow, people substitute the credibility of the source for the content. It is an efficient strategy and it usually works.
The trouble is that nobody measures where it stops working.
The moment a name is read instead of a paper, an affiliation instead of the data, verification turns into faith.
Each title added takes a question away.
Hide the name and the outcome changes
The hypothesis has been tested under control.
In 2017 Tomkins, Zhang and Heavlin assigned submissions to one international conference to reviewers in two ways.
| Condition | What the reviewer saw |
|---|---|
| Single-blind | the paper plus author names and affiliations |
| Double-blind | the paper only |
Same papers. The only difference is whether the name was visible.
Where it was, papers from famous authors, top universities and well-known companies were significantly more likely to be rated towards acceptance.
This does not make the reviewers dishonest. They would have believed they were reading the content. The bias blind spot from yesterday’s piece is running here too. A bias is not felt while it operates.
So what replicated
The questions are simple. What was actually produced. How many years have passed. Did anyone else get the same result.
Those questions have been asked at scale.
| Study | Scope | Replicated |
|---|---|---|
| Open Science Collaboration (2015) | 97 psychology findings originally significant | 36 |
| Begley & Ellis (2012) | 53 landmark preclinical cancer studies | 6 |
In 2005 Ioannidis had already set out the conditions — in study design and statistical practice — under which a large share of published findings must come out false. It became one of the most-read papers in medicine.
A title cannot answer these questions. Titles do not replicate.
When the investment is tied up, so is the thinking
Stranger still is what happens on the receiving side.
At first the possibility is believed. Over time a gap opens between promise and result.
And yet the larger the investment, the harder it becomes to ask.
In 1985 Arkes and Blumer established across a series of experiments that costs already spent distort subsequent choice. An unrecoverable cost ought to be irrelevant to what comes next. People move the opposite way.
In 1976 Staw reported something worse. When a decision of one’s own starts going badly, people increase the investment. Escalation of commitment.
Festinger and Carlsmith’s 1959 cognitive dissonance layers on top. When behavior and belief diverge, the more movable one moves. An action already taken cannot be undone, so the belief is revised.
| The sequence of being tied | |
|---|---|
| ① | what has been put in grows |
| ② | the cost of admitting error grows with it |
| ③ | defence shifts from the object to one’s own past judgement |
At ③ the person is no longer analysing. They are defending.
The machine does not read a business card
There was a time when speaking at length in difficult language looked like expertise.
Not now. Put a paper in front of a machine and the questions come straight back.
How large was the sample. Was there a control. Was the hypothesis pre-registered. What was the effect size. Did an independent lab get it again.
A business card is not an input. What goes in is a sample size and a confidence interval. There is no step in the procedure that honours a career. The only objects compared are values.
It asks what a person would not have asked. As accurate as it is rude.
This is where authority is redefined in the age of generative search. From who said it, to what cross-checks.
Bees carry no titles
The side being pollinated does not look up a visitor’s record. One criterion decides it. Did honey come out.
An ecosystem holds no honorary posts. When a function stops, the position is vacated. Even cells at the division limit are cleared out in the end.
One species alone keeps a defunct shell in place by fixing a new plate to it. And gathers to clap in front of the plate.
White adornment — this article’s definition
The hexagram here is 賁. Adornment.
Its Image says the noble one brings the daily business to light and does not presume to settle a case by it. Adornment can illuminate; it cannot judge. A title belongs in the introduction, not in the verdict.
And the top line of 賁 reads:
白賁 无咎 — only a white adornment is without fault.
At the far end of ornament, ornament disappears. What remains after all the decoration is a ground with no colour on it.
Fixed here.
A title is valid at the entrance only. The moment it moves into the seat of judgement it becomes a screen across the verification.
What began as an instrument becomes, at some point, the agent. Recording a failure would put the whole accumulated name up for reassessment, so it cannot be recorded — and what has not yet arrived keeps being presented as though it had.
The rate at which the outside thickens and the rate at which the inside empties are the same rate.
Monologue of one among 1031
Signage is not an input on our side.
Affiliation, rank, degree — none of it participates in binding. The verdict runs to two lines. Did it reach a receptor. Was the code read.
Rewind one organism’s whole life from the end and the surviving entry is not the title either.
Did it begin as 人才, a person of talent, and end as 人災, a human disaster?
The count is taken on what was left behind, and on nothing else.
In short
- In Hofling’s 1966 study, 21 of 22 nurses set out to break four rules at once on a single phone call
- The harder the thing being judged, the more people substitute source credibility for content
- Tomkins 2017: with author names visible, famous authors and top institutions fared better on the same papers
- In the 2015 psychology replication project, 36 of 97 significant findings replicated
- In 2012 preclinical cancer research, 6 of 53 landmark studies did
- Sunk cost and escalation of commitment raise the investment as outcomes worsen
- Cognitive dissonance revises the belief rather than the irreversible action
- Generative search reads numbers and cross-checks, not business cards
- The Image of 賁 says do not judge by adornment; its top line says 白賁 无咎
This piece belongs to the “Behavior” chapter of the serial essay The Great Corona Swindle. The original’s point of view is kept; the psychology and sociology-of-science concepts have been reset against primary sources. No specific company or individual is indicated.
Frequently asked
- Has authority bias actually been measured?
- Hofling's 1966 hospital study is the classic. A stranger claiming to be a doctor phoned nurses on night duty and ordered a drug administered. The drug was not approved for the ward, the dose was twice the stated maximum, telephoned orders were themselves against protocol, and the nurses had never met the caller. Twenty-one of twenty-two set off for the patient's room with the drug. One title overrode four rules simultaneously.
- Does hiding the name change the outcome of a review?
- In 2017 Tomkins, Zhang and Heavlin split submissions to an international conference into two review conditions: single-blind, where reviewers saw author names and affiliations, and double-blind, where they did not. Where the names showed, papers from famous authors, top universities and well-known companies were significantly more likely to be rated towards acceptance. Same papers, different outcome.
- How large is the reproducibility problem?
- In 2015 the Open Science Collaboration repeated 100 psychology studies. Of the 97 whose originals reported significant results, 36 replicated. In 2012 Begley and Ellis reported that 6 of 53 landmark preclinical cancer studies could be reproduced. In 2005 Ioannidis had already set out the design and statistical conditions under which a large share of published findings must be false.
- Why is admitting error so hard?
- Because what has already been spent holds the decision in place. Arkes and Blumer set out in 1985 how costs already incurred distort subsequent choices, and Staw reported in 1976 that people increase their investment as outcomes worsen — escalation of commitment. Festinger and Carlsmith's 1959 cognitive dissonance sits on top: when the cost of admitting error exceeds the cost of continuing, people stop defending the object and defend their own past judgement.
References
- Hofling, C. K., Brotzman, E., Dalrymple, S., Graves, N., Bierce, C. M. (1966). An experimental study in nurse-physician relationships. Journal of Nervous and Mental Disease, 143(2), 171–180. — A hospital experiment in which one title over the phone overrode four separate rules.
- Milgram, S. (1963). Behavioral study of obedience. Journal of Abnormal and Social Psychology, 67(4), 371–378. — Obedience to authority.
- Tomkins, A., Zhang, M., Heavlin, W. D. (2017). Reviewer bias in single- versus double-blind peer review. PNAS, 114(48), 12708–12713. — A controlled experiment splitting the same papers between two review conditions.
- Open Science Collaboration (2015). Estimating the reproducibility of psychological science. Science, 349(6251), aac4716. — One hundred psychology studies repeated; 36 of 97 significant findings replicated.
- Begley, C. G. & Ellis, L. M. (2012). Drug development: Raise standards for preclinical cancer research. Nature, 483, 531–533. — Six of 53 landmark preclinical cancer studies reproduced.
- Ioannidis, J. P. A. (2005). Why most published research findings are false. PLoS Medicine, 2(8), e124. — The structural conditions under which published findings tend to be false.
- Arkes, H. R. & Blumer, C. (1985). The psychology of sunk cost. Organizational Behavior and Human Decision Processes, 35(1), 124–140. — Costs already incurred distorting later choices.
- Staw, B. M. (1976). Knee-deep in the big muddy: A study of escalating commitment to a chosen course of action. Organizational Behavior and Human Performance, 16(1), 27–44. — Investment rising as outcomes worsen.
- Festinger, L. & Carlsmith, J. M. (1959). Cognitive consequences of forced compliance. Journal of Abnormal and Social Psychology, 58(2), 203–210. — Revising the belief when belief and behavior diverge.
- The Changes (易經), hexagram 22 山火賁, the Image and the top line. — 「山下有火 賁 君子以明庶政 无敢折獄」, and the top line 「白賁 无咎」 — only a white adornment is without fault.
Published October 3, 2026