In 1993, the British Medical Journal advised people to stay home on Friday the 13th. A study suggested that the risk of hospital admission after a traffic accident might be 52% higher that day. The evidence came from just six pairs of dates, and later Swiss analyses pointed in opposite directions. Is Friday the 13th really more dangerous?

The advice appeared in a paper published by the British Medical Journal on 18 December 1993:

“Staying at home is recommended.”

The authors reported that on Friday the 13th, the risk of hospitalization after a traffic accident might rise by 52%. [1]

It was the Christmas issue. The BMJ has a tradition of publishing genuine research on unexpected, sometimes amusing subjects at the end of the year. Hollywood had already made the date famous with the 1980 horror film Friday the 13th. Now the date had a medical statistic to add to its reputation.

But that 52% came from surprisingly little data.

Where did the 52% come from?

Between 1989 and 1992, British researchers compared Fridays that fell on the 6th and 13th of the same month. They counted vehicles on sections of the M25 highway around London, shoppers at nine supermarkets, and emergency hospital admissions after transport accidents. The original data are still available. [2]

What was measuredFriday the 6thFriday the 13th
Hospital admissions after transport accidents4565
Vehicles on M25 sections1,283,8531,265,495
Customers at nine supermarkets223,673225,765

On Friday the 13th, traffic on the roads surveyed was 1.43% lighter. Hospital admissions rose from 45 to 65, an increase of 44.4%. The supermarkets had slightly more shoppers.

Why did headlines repeat 52% rather than 44.4%? The higher figure was the authors’ estimate of relative risk. The paper does not explain in detail how they arrived at it from the raw counts.

Traffic was counted on selected sections of the M25 in Surrey, while hospital admissions covered the much wider South West Thames region. The two totals came from different areas. Dividing admissions by the observed traffic would not tell us the risk on an individual trip. We explored a similar gap between overall figures and individual risks in our analysis of athlete mortality statistics.

The authors acknowledged how few admissions they had counted. What stuck was the advice to stay home.

Just six pairs of Fridays

In November 2026, The Null Institute, a research project I lead, reanalyzed the original BMJ data. We published the calculations and code separately. [3]

There were just six pairs of Fridays. In five of them, more people were hospitalized on the 13th than on the 6th. On reanalysis, one standard statistical test found a significant difference; another did not.

The crucial caveat: with just six pairs of dates, we cannot reliably distinguish an effect of the 13th from random variation, changes in travel volume or weather.

Friday the 6th and Friday the 13th are a week apart. That is enough time for dry roads to turn icy, or for a snowstorm or downpour to hit. Across six date pairs, a single unusual day can swing the average.

Switzerland: more crashes, then fewer

In 2022, the Touring Club Switzerland (TCS) examined crashes recorded from 2019 to 2021. On Friday the 13th, the average was 184.2 crashes, compared with 163.93 on other Fridays, roughly 12% more. [4]

In 2026, TCS examined 2020–2024 instead and found 2.6% fewer crashes on Friday the 13th than on an average Friday. [5]

PeriodCrashes on Friday the 13thAverage crashes on a Friday
2019–2021184.2163.93
2020–2024160.62164.88

Average number of recorded crashes per day. The first dataset included five Fridays the 13th; the second included eight.

One difference was 2019, which included two unusual Fridays the 13th. On 13 September there were 212 crashes amid heavy traffic. On 13 December, authorities recorded 253 crashes as heavy snow fell in some Swiss cantons.

The second analysis excluded 2019 and included more recent years. This time, the Friday-the-13th average was lower than the average for other Fridays.

There was another way to read the 2026 report. Compare Friday the 13th with every day of the week, not just Fridays, and the figures become 160.62 crashes versus 142.60: 12.6% more. Friday itself was the day of the week with the most crashes.

The same report could support two headlines: “12.6% more crashes on Friday the 13th” and “2.6% fewer crashes on Friday the 13th.” Both are arithmetically correct. What changes is the comparison group.

Finland and Japan: conflicting results

In 2002, Finnish researcher Simo Näyhä examined fatal road accidents from 1971 to 1997. Over those 27 years, 82 men and 41 women died on Fridays the 13th. After statistical adjustments, the risk for men was close to that on other Fridays; for women, it was 63% higher. Näyhä suggested that anxiety about the date might be responsible. [6]

Two years later, Igor Radun and Heikki Summala studied road accidents involving injuries from 1989 to 2002, including 21 Fridays the 13th. They found no rise in crashes involving women. They had measured a different outcome, but the proposed “female effect” did not show up again. [7]

In Japan, where 4 and 9 are more widely regarded as unlucky than 13, a researcher examined roughly 1.88 million police records of crashes from 2019 to 2024. The resulting 2026 preprint likewise found no increase in crashes on Friday the 13th after statistical adjustments. [8]

Could superstition itself change the statistics?

In the British study, traffic on the roads surveyed was slightly lighter on the 13th. Perhaps some drivers simply chose to stay home.

If some people made fewer trips, there might have been fewer crashes. Others might have been more nervous behind the wheel. The overall figures would depend on how many people drove and how they behaved.

The date itself does not change the weather or the condition of the road. But it may change drivers’ plans, and with them the number of trips.

So should we stay home?

Studies from different countries do not show a consistent rise in crashes attributable to Friday the 13th. The British 52% figure was never as solid an argument as the headlines suggested.

In Switzerland, one snowy Friday the 13th made it into the first report but not the second. Change the dates, and the conclusion changed with them.

Before heading out, check the weather forecast, not the calendar.

References

  1. Scanlon T. J. et al. (1993). Is Friday the 13th bad for your health? BMJ, 307, 1584–1586. Full original paper. https://pmc.ncbi.nlm.nih.gov/articles/PMC1697765/
  2. OpenIntro. Friday the 13th. Open dataset from the British study. https://www.openintro.org/data/index.php?data=friday
  3. Nikitin I. (2026). Friday the 13th: We Rechecked the Study Behind the Accident Scare. The Null Institute. Reanalysis, methods, code and limitations. https://nullinstitute.org/friday-the-13th-accident-risk-reanalysis/
  4. Touring Club Schweiz (2022). Analysis of Swiss road crashes in 2019–2021. https://www.tcs.ch/assets/docs/presse/2021/communiqu-de-presse-vendredi-13-2022.pdf
  5. Touring Club Schweiz (2026). Analysis of Swiss road crashes in 2020–2024. https://www.tcs.ch/assets/docs/presse/2026/120326-medienmitteilung-freitag-13.pdf
  6. Näyhä S. (2002). Traffic Deaths and Superstition on Friday the 13th. American Journal of Psychiatry, 159(12), 2110–2111. https://doi.org/10.1176/appi.ajp.159.12.2110
  7. Radun I., Summala H. (2004). Females do not have more injury road accidents on Friday the 13th. BMC Public Health, 4, 54. https://pmc.ncbi.nlm.nih.gov/articles/PMC544193/
  8. Shirai M. (2026). Friday the 13th and Traffic Accidents in Japan: A Natural Cultural Control Study of 1.9 Million Police Records. SSRN preprint. https://doi.org/10.2139/ssrn.7170660