December 24, 2019

Grandma got run over by a reindeer

The Accident Compensation Corporation, a wonderful NZ institution, always takes advantage of the season to try to get coverage of Christmas accident risks. There are two popular angles, and this year our newspaper sites have one each.

The NZ Herald writes about the total cost of Christmas claims.  These, according to the story, are steadily increasing: $448k in 2014, $497k last year. The increase, of nearly 11%, needs to be compared to the population increase of about 8.4% (June 2014 to June 2018) and to inflation of about 5%.  Christmas accident claims have increased very slightly per capita in nominal dollars, and decreased very slightly per capita in constant dollars.  And, as the story admits, Christmas Day is not a particularly high-risk day — it’s compared there to New Year’s Day, but it actually has lower ACC costs than a typical day.

On Stuff, the story is about holiday-specific accidents. Here we’re on stronger ground.   The Christmas tree, pavlova, and turkey injuries are reliably attributable to Christmas; we’re talking causation, not just correlation.  If I were feeling pickier than is appropriate for the season, I might complain about “Turkey was a slightly safer option”  comparing turkey and ham-related injuries: we don’t know that turkey was safer, since we’d need the number of people at risk from turkey and ham to draw that conclusion.

The song referenced in the title makes clear the difference between injuries occurring at Christmas and those caused by the season — “she had hoof-prints on her forehead / And incriminating Claus marks on her back”.  Even though there are fewer injuries on Christmas Day, we can still tell that some of them are Christmas injuries. It makes sense for ACC to try to reduce these, especially given the reduced competition for news coverage at this time of year.

 

avatar

Thomas Lumley (@tslumley) is Professor of Biostatistics at the University of Auckland. His research interests include semiparametric models, survey sampling, statistical computing, foundations of statistics, and whatever methodological problems his medical collaborators come up with. He also blogs at Biased and Inefficient See all posts by Thomas Lumley »