Check the Crosswalk, Not the Candy

Plenty of us grew up watching our parents inspect the Halloween haul at the kitchen table. They held apples up to the light and threw out anything unwrapped, because a stranger somewhere might have slipped in a razor blade or poison. I also remember for the whole of October the local news ran the same segment about the dangers of getting free candy.

Sociologist Joel Best at the University of Delaware spent years tracing those stories. With his coauthor Horiuchi, he reviewed decades of news reports about "Halloween sadism." Most of the incidents involved no injury, and many turned out to be hoaxes. Best found no substantiated case of a child killed or seriously injured by a contaminated treat or chomping down on a hidden razorblades in a candied apple.

Candy did take the blame for two deaths, but neither involved a stranger. In 1970, Kevin Totson died after eating heroin later found at a relative's home. Four years later, Timothy O'Bryan's own father poisoned him.

Last week I wrote about the denominator, the population underneath every health number. For candy tampering, the denominator is every child who trick-or-treats in a given year. The numerator is essentially zero verified harm.

Cars, however, are the real danger on Halloween. Researchers Staples, Yip, and Redelmeier used federal fatal-crash records from 1975 to 2016 to compare pedestrian deaths on Halloween evenings, 5 p.m. to midnight, with the same hours one week before and one week after. Because the control evenings fall on the same weekday and in the same fall darkness, the holiday is the main difference.

Across 42 Halloweens, 608 pedestrians died, compared with 851 on the 84 comparison nights. Per evening, Halloween averaged about 14.5 deaths and the comparison nights about 10.1.

The researchers expressed the gap as an odds ratio of 1.43, meaning the relative risk of a pedestrian fatality was 43% higher on Halloween. A relative measure shows how much bigger one number is than another. Absolute risk is the number of people affected, here roughly 4 extra deaths per Halloween nationwide (14.5 minus 10.1).

Either alone can mislead. Forty-three percent sounds like a jump scare, while four deaths across the country is small. For any single family the odds are slim, but each of those deaths was a child or adult who didn't come home. Over the study, the absolute rate fell by about half, from 4.9 to 2.5 per 100 million Americans, comparing 1975 to 1985 with 2006 to 2016.

The risk is highest for little kids. For ages 4 to 8, the odds ratio was 10, the largest of any age group. Deaths peaked around 6 p.m., when the youngest trick-or-treaters are out and commuters are driving home in fading light.

The American Academy of Pediatrics recommends walking with young kids on their neighborhood rounds. Add reflective tape to dark costumes, give kids a flashlight, and swap masks for non-toxic face makeup so they can see traffic clearly. Cross at corners, where drivers expect pedestrians. If you're driving, especially if you're a teen, slow down in neighborhoods and watch for small kids darting out from driveways. I mean, really, this should all be common sense.

So when you sort through the candy this year, ask the same questions as last week: what's being counted, out of how many, over what period of time, and compared to whom. By those numbers, the street is a bigger threat than the candy bowl.

Sources: Staples JA, Yip C, Redelmeier DA. Pedestrian Fatalities Associated With Halloween in the United States. JAMA Pediatr. 2019;173(1):101-103; Joel Best, Halloween sadism; Best J, Horiuchi GT. The Razor Blade in the Apple: The Social Construction of Urban Legends. Social Problems. 1985;32(5):488-499; American Academy of Pediatrics, Halloween Fun & Safety Tips; Out of How Many? The Denominator Is Half the Story.