On April 2, 2019 the President of the United States delivered a speech in which he said that the noise from windmills causes cancer. Watch it yourself.
The question is so bizarre, I doubt it has been scientifically studied, so I gave a quick stab at it. I acknowledge the methodology was crude, but this is what I did.
Data Sources And Variables
I gathered data on national per capita cancer rates, and national per capita megawatts generated from wind power. For the scientifically minded, the hypothesized independent variable (the cause) is national per capital megawatts generated from wind power. The dependent variable (the effect) is national per capita cancer rates.
The Theory, Hypothesis and Null Hypothesis
The theory being tested is that wind power causes cancer. The tested hypothesis, based on this theory, is that as national per capital megawatts generated from wind power increases, national per capita cancer rates shall also increase. As usual in statistics, the Null Hypothesis is that there is no statistically significant relationship between these two variables.
I decided to use only industrialized countries, and I was limited to countries who were included in both lists. That produced the following chart.
Data Analysis and Regression Analysis
The next thing to do was a simple scatterplot of that data.
The scatterplot reveals no obvious relationship. But this is an effort at statistics and science. I commanded the spreadsheet program to perform the traditional least squares standard linear regression analysis line fit. That produced this chart.
Two things stand out about this chart. First, and perhaps foremost, to the extent there is a relationship between the independent variable of wind power and the dependent variable of cancer rates, that relationship is negative. Put another way, to the extent there is a relationship, as national per capita wind power increases, national per capita cancer rates decrease. This is the exact opposite of the relationship hypothesized based on the President’s theory.
The Question of Statistical Significance
Of course, that relationship certainly does not look very strong. It isn’t. Here are the basic tests of statistical significance from those numbers.
Pearson’s R= -0.166
R Squared= 0.028
N (sample size)= 17
The R Squared of 0.028 is actually a calculation of a percentage. It means that to the extent the statistical relationship measured actually exists, then a rather meaningless 2.8% of the decrease in cancer rates is explained by the increase in wind power megawatts per capita.
Which leads to the question of whether the measured statistical relationship exists. That is reflected in the variable “p” above. The .53 means that there is a 53% chance that the apparent negative relationship between wind power and cancer rates is completely explained by . . . chance.
Standard rules of science require much, much, much higher standards of proof. Generally, p must be < 0.05 reflecting less than a 5% chance of being explained by chance. This is not even close. So what are the conclusions?
The hypothesis based on Trump’s theory that wind power causes cancer is not supported. There is stronger support for the opposite claim.
The Counter Hypothesis
The reverse hypothesis that wind power is associated with less cancer is poorly supported, with nothing close to scientific significance.
The Null Hypothesis
The best support is for the Null Hypothesis. There is no relationship between wind power and cancer.
Weaknesses Of The Study And Suggestions For Further Research
My sample size was obviously limited. A larger one, involving more nations and more years might produce statistically stronger results.
In addition, the variables used were crude. Cancer, and in particular national cancer rates, involve enormously complex multiple variable interactions. This study did not control for any other potential variables. However, one wonders whether the President’s anonymously sourced “they say” controlled for any variables at all. Not knowing who the “they” are, I cannot say.