16

Here we've got a claim that infant mortality in select US cities on the west coast has increased by 35% since the nuclear meltdown. This claim is based on two data points:

4 weeks ending March 19, 2011 - 37 deaths (avg. 9.25 per week)

10 weeks ending May 28, 2011 - 125 deaths (avg.12.50 per week)

Elsewhere, people are worried that an increase in infant mortality in British Columbia is the fault of Fukushima.

“There have been 21 sudden infant deaths in B.C. so far this year, while there were 16 sudden infant deaths for all of 2010″

Increases in infant mortality rates is now almost across the board in North America since the disaster and the STILL continuing nuclear crisis in Japan. But in Canada it is because of “unsafe sleep practices”? Are we really going to fall this nonsense?

Are these increases significant? Is there any evidence that they are the result of Fukushima?

11
  • 4
    And I thought radiation caused cancer... ;-)
    – Sklivvz
    Jul 14, 2011 at 20:58
  • 2
    @Sklivvz just because it causes cancer does not mean I can not cause other issues as well. That said I think it is too early for this to be considered anything more than a statistical burp. With out statistics for similar time periods for 10+ years. And continued elevated numbers of SIDS above the norm. This data is not even enough to show that the numbers are outside of the margin of error for datacollection.
    – Chad
    Jul 14, 2011 at 21:15
  • 2
    (This is somewhat of a rhetorical question. I already know that the US "study" is cherry-picking. I don't know about the BC claims, though.)
    – endolith
    Jul 15, 2011 at 0:48
  • 7
    @endolith - It only cites the 4 weeks prior then chooses 10 weeks after. that seems fishy to mean. Like I said it reeks of a statistical burp. Basically a statistically low period followed by a spike. And there is no talk of what the general average is for the time period/area/overall. What are the general trends. If you have more babies born it makes more sense that more babies die too. Too many unknowns to do anything with this.
    – Chad
    Jul 15, 2011 at 17:17
  • 3
    @endolith No, I mean it’s been debunked. I’ve read it somewhere, if not here then elsewhere. Basically the weeks around what’s been cited completely reverse the trend. So Chad guessed right. Jul 16, 2011 at 13:52

2 Answers 2

38

Both of the statistics quoted lacks several vital pieces of information. These include:

  • how much the death rate varies naturally from year to year;
  • Why the "select cities" mentioned in the article were selected, and what the rate becomes if you select different ones;

It also lacks any explanation of how the alleged radiation could have caused these deaths. The usual statement has to be made about all such allegations:"correlation is not causation".

The Canadian article is just a rant with no information. It doesn't even state whether the infant death rate went up before or after the Fukushima tragedy (which took place halfway through the period that article claims to have figures for).

Against this is the truth that the amounts of radiation which reached the US (and Canada) were absolutely trivial. Here is an article giving some information: initial readings are "about a billion times beneath levels that would be health threatening." People who don't understand radiation often assume that "any radiation causes disease", forgetting that we are all subjected to much more radiation than reached North America from Japan just from natural sources.

5
  • 2
    +1 for clearly articulating that why statistics devoid of context are most likely bogus.
    – matt_black
    Oct 4, 2011 at 22:19
  • Well, they proposed that the radiation was concentrated in small "hot particles": "These [hot particles] can cause cancer, but you can't measure them with a Geiger counter. Clearly people in Fukushima prefecture have breathed in a large amount of these particles. Clearly the upper West Coast of the US has people being affected. That area got hit pretty heavy in April."
    – endolith
    Nov 19, 2012 at 15:02
  • The 'hot particles' might have an effect over the long term. They aren't really relevant to the infant death statistics over the period asked about. Nov 19, 2012 at 15:27
  • 1
    all those "hot particles" were long rained out over the open ocean before reaching US shores, diluted to insignificance in the massive volume of water that is the Pacific Ocean, and by now mostly reduced by fission processes to non-radioactive isotopes.
    – jwenting
    Mar 27, 2013 at 11:20
  • Did they establish what the radiation environment was like in the sampled cities before and after the disaster? What about ruling out other potential environmental changes? Without that they can't even show a correlation, let alone demonstrate a causal link.
    – GordonM
    Aug 15, 2018 at 9:43
17

No, this is fraudulent statistics

Here is a real life example of Lies, Damned Lies, and Statistics: they fudged the numbers to make it look bad.

The long answer is very long indeed. Here is a blog post on the subject. A complete analysis accompanies that post and can be found here.

In summary though: they cherry-picked the numbers, selecting towns and time periods that make it look like there is a spike in infant mortality after Fukushima. But when looking at a longer time perspective and over more cities, this spike disappears among the statistical noise. Instead it appears there was a negative spike in those cities, with a drop in infant mortality just before Fukushima, and then going back to average level after.

enter image description here

The data that Mangano and Sherman chose to show

enter image description here

Same cities, but including a longer time period before the accident

enter image description here

Expanded time period, and going for national average instead of just the selected cities

3
  • 1
    National average may be irrelevant, since the question was about the West Coast.
    – GEdgar
    Aug 13, 2018 at 20:56
  • 2
    @GEdgar considering there is as much Fukushima radiation on the west coast as in the rest of the US, i.e. barely detectable, at 1-2 Bq per ton (1000 kg) of sea water - compare to bananas or humans at 50 000 - 100 000 times more at 100 Bq / kg... the numbers are relevant.
    – user32299
    Aug 13, 2018 at 21:00
  • I do love me some charts.
    – Harabeck
    Aug 16, 2018 at 20:17

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .