Claude Brasseur September 12, 2016 Belgium

By Claude Brasseur, mathematician
brasseurvossen@skynet.be

The premises

1. According to (among others) studies by Doctor Rachel LEPROULT (ULB-Belgium), chronic lack of sleep shortens life.
2. The vicinity of wind turbines appears to be causing chronic sleep deprivation to an increasing number of residents.
3. All other things being equal, if the mortality records kept by retirement homes located in the vicinity of wind turbines evidence a faster “turnover” since the turbines were installed, this would constitute objective proof that the life of their residents is being shortened by their presence.

Discussion

Let’s imagine, ideally, the case of a residence for the golden age where occupancy has not changed for, say, 20 years. They have a waiting list and, as soon as a bed becomes vacant due to mortality, a new resident occupies it. And the mortality rate is, say, 10% per annum.

Let’s suppose that, in a given year, a wind farm is installed nearby. In the following years, mortality increases to, say, 11% in the first year, 12.5% in the second, 15 % over the third, etc.

If, as I stated in the premises, all other things have stayed equal (no change in the diet, in the environment, in the temperature, etc. and no significant change in the age pyramid of residents and their state of health), this would be an indication that the operation of a wind farm might be shortening the life of people living nearby.

Of course, all other things never remain absolutely equal, and this will cause distortions. For example, the turnover rate might change for other reasons, with or without wind turbines. But this variance can be taken into consideration, for we have records going back many years before the installation of the turbines, so we can establish a range of variations in the baseline, which can then be compared with the range of variations after the inauguration of the wind farm.

In any event, statistical significance can only be achieved if we have a large database. In other words, if we are studying the mortality rates of a large population of elders, variances upwards and downwards will likely tend to even out. This is precisely what makes this idea interesting: in my small country, Belgium, there probably is a sizeable number of retirement homes that have wind turbines within, say, 20 km. Statistical significance can be achieved relatively easily.

The other advantage is that the mortality data is already available in the records of each institution, and probably as well in those of the public administration. It would just be a matter of gathering, organizing and processing it. The difficult part would be the work of the statisticians, as they try to eliminate the distortions created by, for instance, unusual vacancies in some of the residences, and by any other factor – e.g. the distance to the nearest wind turbine, its power capacity, etc.

This idea came to me as I heard someone in a “residence” say: “since the wind turbines, they (the elderly) are dying like flies”. Being a mathematician, I have built a model which could be used to process the data once it is made available to statisticians. — Note: I will use the “Anova” method.

THE STATISTICAL MODEL
Procedure I suggest to follow for processing the collected data

First, let’s suppose that the hosting capacity of a given home for the elderly has remained unchanged over the study period. Let’s also suppose that no bed stays empty any significant lapse of time, because there is a waiting list for that home.

Then, let’s imagine a study period starting January 1, 1995 and finishing December 31, 2014, and that wind turbines started to operate in the vicinity of the residence on January 1, 2005 — i.e. we would study mortality data for 10 years without wind turbines (baseline), and 10 years with their influence.

Let’s now suppose we have collected the following data:

• yearly turnover before the wind turbines:
  93, 105, 115, 82, 75, 110, 75, 98, 101, 120
• yearly turnover after the wind turbines:
  104, 98, 125, 132, 117, 89, 131, 115, 122, 117

The average turnover before the wind turbines is: M1 = 97.4 each year
The average turnover after the wind turbines is: M2 = 115.0 each year

Standard deviation concerning M1 is S1 with
S1² = {(93² + 105² +....) / 10} − M1² = 231.04
S1 = 15.2
Standard deviation concerning M2 is S2 with
S2² = {(104² + 98² +....) / 10} − M2² = 178.8
S2 = 13.4

Now let’s suppose that the two above data samples would in fact produce an unchanged average (M1 = M2), meaning that wind turbines would have no effect on the health of the senior citizens sampled. In statistics this is called the null hypothesis. We would then calculate:

Standard deviation of the difference of averages M1 and M2:
S of M1 − M2 = {(15.2 / 10) + (13.4 / 10)}1/2 = 1.7

To be able to compare data distributions between them, we’d calculate the reduced centred variable:
Z = (97.4 − 115) / 1.7 = − 10

With a unilateral meaning test of 0.05 (risk of error of 5%), the null hypothesis would then be refused, because Z = − 10 is smaller than − 1.645 (the figure supplied by the normal law of statistics).

This would mean that wind turbines are harmful to the health of the sampled senior citizens. We could then want to find out by what process they shorten their lives. We already know from abundant scientific literature authored by independent researchers that low frequency noise, infrasound and amplitude modulation are the likely culprits (this is why governments refuse to look in that direction).

I suggest the use of the ANOVA method for the statistical analysis of the data, and insist that the larger the number of senior citizens encompassed in the study, the more reliable will be the results.