Don't trust numbers from the innumerate

From conspiracy theorizing about food plant fires by el gato malo.  El Gato Malo is asking the right questions.  The mainstream media is full of stories of food processing plants going up in flames at a startling rate but they do not actually specify what that rate might be rate.  They list twenty-eight (or whatever the current number might be) plants having been burned but are obscure about the timeframe in which they are counting and their definition of a fire (a fire or a plant burned down?)   

This is an old mainstream media failing (or trick, depending on your level of jaundice).  I first came across it sometime back in the mid- or perhaps in the late 1990's.  

There was a slew of headlines lamenting the frequency with which black churches were being burned across the county and heralding the return of the blatant and systemic racism of which journalist seem to be convinced exists.  The first indication that there might be less to the story than was being presented was the absence of any comparative base line evidence.  If X number of churches are being burned now, how many are usually burned in this time frame?

Then it became apparent that black churches were being burned in Canada as well which raised another red flag that perhaps the mainstream media narrative might not be especially empirical.

The screeching and gnashing of teeth went on for some months before someone bothered to answer the two obvious questions.  How many churches usually are burned in this time period?  Is there a differential between white churches being burned and black churches being burned?

I don't remember the numbers from that episode but they are probably not proportionately far from the current numbers.  There are some 380,000 churches in the US and there are some 1,500 churches burned each year.  

To know whether there is some significant change in churches being burned, we need to know both how many are being burned versus how many are usually burned in the same time period.  Without those two pieces of information, this issue is merely mainstream media induced panic porn.

Really, we need a third piece of information - the variance, year-to-year.  If, over a decade, 1,500 churches are burned each year, we also need to know the variation per year.  Does it usually range from 1,400-1,600 or is it 1,000-2,000 variance per year.  

If there are 1,000 churches burned in year X and 1,500 in year X+1, I can panic over the 50% increase or I can remain calm knowing that the variance is from 1,000-2,000 in any given year and that 1,500 merely reflects the reversion to mean.  

Newspapers which do not provide numbers and context to those numbers are either innumerate or deliberately conjuring fear.  Or both.  

Are there an unusual number of food plants being burned this year?  I don't know.  Per the USDA, there are some 38,000 food processing plants in the US.  There have been headlines claiming as many as 15-30 food processing plant fires in 2022.  Though, significantly, most headlines and corresponding articles are pretty nebulous about the starting date for the supposed recent trend in fires with some of them reporting food processing plant fires which occurred as long ago as 2020.   

However, the baseline appears to be at least 2,000 fires at food processing plants each year (or more depending on definition.)  15-30 fires reported against an average annual rate of 2,000 is not indicative of anything.  At that rate, we should expect 35 or so food processing plant fires per week.  A report of 30 food processing plant fires for the first four months of the year has no particular relevance.

Seems like whenever the mainstream media wants you to be suspicious or concerned about something, you always have to do your own fact-checking because they are almost invariably wrong.