All things environmental

Musings about the environment and all it touches, from education to city planning

Posts Tagged ‘monarchs; logical fallacies; correlation; time series; Agarwal

Logical fallacies and the environment: correlation and causation

leave a comment »

Last year I published a series about logical fallacies that abound in pronouncements about environmental issues (here is an installment, as an example).  But I recently came across one of the best examples that illustrate how correlation and causation are distinct.

The example comes from the recent book by Anurag Agrawal, a world authority on monarchs.  Monarch numbers are clearly in decline, but what is causing it?  There are multiple causes, as always.  One possible cause that has often been pointed at is the use of GM corn and soybeans.  These plants are resistent to herbicides such as Roundup (glyphosate), and therefore much more of the herbicide has been sprayed as before.  If the drifting herbicide is also killing milkweed, it would deprive monarchs of their food source (monarchs eat only milkweed – nothing else).  Could that be the main reason that monarchs are becoming rarer?

On page 236 of the book, Agrawal proposes a graph (reproduced below) that seems to indicate that this is the case.

The caption reads: a) correlation between the percentage of genetically modified, herbicide-tolerant corn and soybeans planted each year in the United States and the monarch overwintering population size in Mexico, and b) the number of cell phone subscriptions in the United States and the same estimate of monarch populations.  Each point represents a year (1993-2014).  Note that the points are not in chronological order.

The top graph seems to strongly indict the GM plants, doesn’t it?  But wait, the bottom graph shows that maybe radiation from cell phones is to blame.  This is not as farfetched as it may originally seem: we still don’t know exactly how monarchs orient themselves for their yearly migration to mexico, and it is conceivable that the EMF from the now cell phones and towers is affecting their ability to navigate using the magnetic field.

But if the monarch decline is plotted against the S&P January stock market values, we get as strong a correlation – what is going on?  The reality is that all three correlations are very strong because the percent of GMO plantations, the number of cell phones, and the stock market have all increased over the time period under consideration, while the monarch numbers have consistently declined.  Agrawal states clearly that the correlations, in themselves, are not an indication of what is causing the decline.  (He also indicates that this doesn’t get the GM crops off the hook, either – merely that causation and correlation are different.  Saying that the fact that this is a mere correlation shows that GM crops are innocent would be equally spurious, a case of either-or fallacy).

The book is full of little gems like this, and I thoroughly enjoyed it – I recommend it to anyone interested in ecology.  If time permits, I will try to post a true review.  But meanwhile, for completeness sake, in case anyone doubts that monarch numbers are declining, here’s the data (from page 214):

We do have a bit of a crisis on our hands, when it comes to the monarch.  But thankfully, this is not the main theme of the book; co-evolution, with all its mysteries and wonders, is what the book is about, and it is written in a very accessible style.  Meanwhile, though, here’s the key lesson: one can really easily find correlations by plotting against one another the Y-axis of time series that have strong time trends.  And quite likely, just as easily can you use it to convince folks that you’ve found the environmental culprit!  But don’t.

 

 

Agrawal, Anurag 2017.  Monarchs and milkweeds: a migrating butterfly, a poisonous plant, and their remarkable story of coevolution.  Princeton U Press.

Advertisements

Written by enviropaul

August 11, 2017 at 12:02 pm