A number. Just a number — four strokes of a pen, one syllable, a digit sitting between three and five on every calendar ever printed. And yet researchers at the University of California, San Diego sat down with twenty-five years of mortality data and came back with a claim that stopped people cold: Chinese and Japanese Americans were dying of heart attacks at measurably higher rates on the fourth day of every month. Not in war. Not in famine. On an ordinary Tuesday or Thursday, simply because the date contained a four.
The researchers named it the Baskerville effect, after the Arthur Conan Doyle novel in which a man is literally frightened to death. The implication was stark. If the numbers held, then superstition — pure, inherited cultural anxiety about a single digit — was killing people.
Then other scientists looked at the data. And the story got complicated fast.
---
To understand why this study landed with such force, you need to understand what the number four means in Chinese, Japanese, and Korean cultures. The Mandarin and Cantonese word for four — 四, pronounced *sì* or *sei* — is nearly homophonic with the word for death — 死. In Japanese, the same character carries the same shadow. The result is a form of numerical avoidance called tetraphobia, practiced widely enough that it shows up in architecture. Hotels in China, Japan, and Korea routinely skip the fourth floor entirely, the same way American skyscrapers used to skip thirteen. The California Yellow Pages offered a quiet confirmation: Chinese and Japanese restaurant phone numbers contained fours in their last four digits only 7.71 percent of the time, against an expected 10 percent by pure chance. American restaurants showed no such pattern. The avoidance was real, measurable, and culturally specific.
David Phillips and his colleagues at UC San Diego had spent years studying what they called the "psychological postponement hypothesis" — the idea that people could delay death around meaningful dates, surviving long enough to reach a birthday or a holiday before letting go. Phillips had previously published work suggesting Jewish men died at lower rates just before Passover, and Chinese women died at lower rates just before the Harvest Moon festival. The Baskerville effect was, in some ways, the dark mirror of that research: not postponement, but acceleration. Not holding on for a meaningful day, but being pushed over the edge by one.
---
The study analyzed daily death rates from heart attacks among Chinese and Japanese Americans between 1973 and 1998. The headline finding: a 7 percent higher rate of cardiac deaths on the fourth of the month compared to the monthly average. A more focused analysis of roughly 20,000 computerized death certificates from Asian-Americans in San Diego sharpened that number to a 13 percent uptick on the fourth. White Americans showed no such spike. The researchers argued the disparity could only be explained by one variable — the cultural weight of the date itself.
The paper was published, the name stuck, and the Baskerville effect entered the lexicon of psychosomatic medicine. It was cited in discussions of nocebo phenomena, the evil twin of placebo — the documented capacity of negative expectation to produce real physiological harm. The mechanism proposed was straightforward enough: psychological stress triggers the autonomic nervous system, elevates cortisol, strains the cardiovascular system, and in someone already vulnerable, tips the balance toward cardiac arrest. Fear, the argument went, was not metaphorically lethal. It was literally lethal.
For a while, the study sat largely unchallenged, its conclusions elegant and its cultural detail compelling. The phone book data felt like a smoking gun. A community so conditioned by tetraphobia that it scrubbed fours from its business listings — surely that same community would feel the fourth of the month as a physiological burden.
---
Gary Smith was not convinced. In 2002, the economist and statistician published a pointed critique. He went back to coronary death data from 1969 to 1988 and from 1999 to 2001 — periods that bracketed and partially overlapped with the Phillips dataset — and found no statistically relevant peaks on the fourth day of the month. Nothing. The spike that Phillips had reported simply did not appear in the adjacent data.
Smith's deeper accusation was methodological. He alleged that Phillips and colleagues had selectively used heart disease subcategories — cherry-picking the categories that happened to show elevated rates on the fourth, rather than applying the same all-inclusive approach Phillips had used in his earlier studies on Jewish and Chinese mortality patterns. When Smith used all heart disease categories, the effect evaporated. The implication was uncomfortable: the Baskerville effect might not be a discovery at all. It might be an artifact of how the data was cut.
A year later, in 2003, Nirmal Panesar and colleagues examined mortality data from Hong Kong's Chinese population between 1995 and 2000. If the Baskerville effect were real, Hong Kong was where you'd expect to find it most powerfully — a predominantly Chinese population, deeply embedded in the cultural traditions that give the number four its lethal reputation, living in a city where tetraphobia is not a relic but a present-tense reality. Panesar's team looked at deaths on the fourth, the fourteenth, and the twenty-fourth of each month. They found no statistically significant difference. Not a trace.
That absence was damning in a specific way. Jonathan Jarry, writing in early 2018, put the problem plainly: if the cultural stress mechanism were real, you would expect a larger effect in Asia than in the United States, where Chinese and Japanese Americans are a minority population, somewhat removed from the most intense expressions of tetraphobia. The Phillips study found the opposite — a detectable effect in San Diego, nothing in Hong Kong. Cardiologist Christopher Labos, quoted in Jarry's analysis, added the statistical fine print: while deaths on the fourth might be nominally higher than the monthly average, they are not statistically higher than deaths on any other individual day of the month.
---
What investigators confirmed is narrow but real: the cultural phenomenon of tetraphobia is documented, measurable, and genuinely embedded in the communities Phillips studied. The phone book finding is not in dispute. The avoidance of four in hotel floors, hospital room numbers, and business listings is observable across multiple countries. The nocebo mechanism — the capacity for negative expectation to produce physiological harm — is supported by independent research in other contexts.
What remains contested is whether the original mortality data actually shows what Phillips claimed. Smith's reanalysis found no effect in adjacent time periods. Panesar found no effect in the population most theoretically susceptible. Labos's statistical framing suggests the original finding may not survive rigorous scrutiny. No independent team has successfully replicated the core result using comparable methodology.
The community of skeptical researchers came to view the Baskerville effect as a cautionary tale about the seduction of a clean narrative. The story was too good — a literary name, a cultural hook, a mechanism that felt intuitive. When a finding is that elegant, it invites scrutiny. And the scrutiny, here, did not go well for the original claim.
---
The Baskerville effect remains in the literature, still cited, still debated. The Wikipedia entry for it exists in a state of permanent tension — the original claim presented, the critiques documented, no resolution offered. Phillips and his colleagues never publicly conceded the methodological objections. Smith and Panesar's counter-evidence has not been refuted.
What lingers is the question the study raised, even if it couldn't answer it. Nocebo effects are real. Cultural belief shapes physiology in documented ways. The mechanism Phillips proposed is not impossible — it is simply unproven in this specific case. The number four did not kill anyone on the fourth of the month in any way that statistics can confirm.
But somewhere in San Diego, a hotel still doesn't have a fourth floor. And nobody who works there thinks that's strange.