“Errors using inadequate data are much less than those using no data at all”
There is a war against Western civilization. So there is a war against objective science, seeking to replace it with Marxist agenda driven science.
Objective knowledge is considered a “white male construct” devised to oppress people of color. There is a war against statistics, the language of proof in science. Monuments to the great statisticians who created statistics are being torn down. Statistics is considered inherently racist.
In 1989, a stained glass window commemorating British statistician and geneticist Sir Ronald Fisher was placed in Gonville and Caius College, Cambridge. After the George Floyd riots in 2020, the window was removed.
Sir Ronald Fisher was simply the greatest statistician to ever live. He is also widely considered the greatest biologist since Darwin. Whenever we enjoy the fruits of science, we benefit from his work. He is credited with saving millions from starvation due to his research on agriculture, millions in India alone.
Of course leftists prefer Trofim Lysenko, whose ideologically pure Marxist genetics caused the starvation of millions and the execution of many legitimate geneticists. Fisher was obviously the greatest person to ever set foot on the campus of Gonville and Caius College, and he has been dishonored in favor of a criminal thug. Such is the sick, crumbling, suicidal West of 2020.
This is an actual quote from a “woke” data scientist, “Objectivity is extremely overrated…Just because you haven’t measured something doesn’t mean that it’s not there. Often, you can see it with your eyes, and that’s good enough.”
No, it is not good enough. Your eyes are fooled, they are the worst judge. Science is about objectivity, science is about data. That is the purpose and value of statistics. It is the only way to determine if my data support my hypothesis. It is the only was to determine if we have proof of something. That is the first thing I would tell my classes.
“Measure what is measurable, and make measurable what is not so”, said Galileo, to capture the spirit of the new emerging science. Of course Galileo got in trouble with the cancel culturerists of the time, who today are persecuting his heirs.
I taught statistics on the college level for 30 years. I am no longer well enough to do much of anything. But more than a decade ago, I volunteered to write the blogs for a company dealing with statistical analysis. I thought it would be a way to get my foot in the door, and lead to future professional advancement. Unfortunately the company went belly up. And I was left with having written a lot about statistics that now resides on my computer, never seen by another human.
One of the things I was working on was a defense of reductionism. Reductionism is part and parcel of Western science, where the attempt is to understand complicated systems by isolating and studying objectively defined variables that are part of the system. The classic example is an experiment holding every variable but one constant, and examining the effects of that one variable on a dependent measure.
Western reductionism was the reason science advanced, the reason the West advanced. It is counter “to holism” and getting answers by intuition or magic.
It is counter to the ideas that underlie Eastern thought, and the passed down “traditional wisdom of the elders”.
Notions that would have been considered superstition and nonsense when I was born, as belonging to the dark ages, as belonging before the enlightenment, before science, are now labeled “indigenous people's wisdom” and replacing science. Just another symptom of the total collapse of society. Yet for some reason, the acolytes of indigenous knowledge do not abandon their cellphones and rely on ESP.
Statistician George Box, who passed away in 2013 at the advanced age of 93, pioneered many statistical techniques bearing his name, yet is as well known for a short simple quote as anything. He wrote “all models are wrong, but some are useful”. Box did not mean that fashion models were stupid but can be fun, although perhaps this is true.
Rather Box meant models simplify an otherwise too complicated to understand reality, and are thus inherently imperfect. But models make predictions that are far more accurate than not having a model at all. A scientific theory makes objectively testable predictions. That is the scientific method and what drove progress in the West. Box’s quote pertains to the heart of statistics, anytime we predict something based on a model we will have error. Yet even with an imperfect model, the predictions are more accurate than those based merely on intuition.
My demonstration of the value of reductionism was to be showing a surprisingly high correlation between a true "unknown" model and a simplified model, in the context of investing in drugs and hoping to get FDA approval. There are 600 potential drugs, we can invest in only 30, but which 30? The true model, which we can not know, consists of an additive relationship between ten variables . For each of the ten variables for each of the 600 “drugs”, SPSS was used to assign a pseudo random score. The true score is simply the sum of all these ten scores for each drug. If we knew the true score we would always be right. The observed model uses only 3 variables picked at random from the ten. If the true model is seen as the criterion and the obtained model is the test, then their correlation represents a validity coefficient.
Let us define as success a drug getting FDA approval. Imagine that 20% of these “drugs” if invested in would get FDA approval. We can see the the percentage of correct decisions made using the observed model, as compared to the base rate of 20% (making decisions by guessing.)
The correlations between the true and obtained models while using only three variables was almost .7. This would mean the following, if 30 drugs were selected for further investments out of the 600 by using the model with 3 variables, we would expect a success rate of 70% in our choices, compared to the base rate of success which is again 20%. In other words, using the model, the expected results on our investments would be 21 hits and 9 misses, and the expected results of not using any model would be 6 hits and 24 misses! That is quite a difference. If each hit meant earning a million dollars and each miss losing a million dollars, it’s the difference between earning 12 million dollars and losing 18 million dollars! And notice again how "poor" our model is, we managed to measure only 30% of the relevant variables!
Charles Babbage (1791 –1871) was an English mathematician and mechanical engineer, credited with inventing the idea of a programmable computer. Though he died before the birth of modern statistics he wrote “Errors using inadequate data are much less than those using no data at all.” He was right.
Read next: While The Left Attacks Christianity, Russia Defends Western Values
Email Us | Reader Friendly | Comments
Dr. Alexander Nussbaum has had articles in a number of magazines including articles on intelligent design and on the history of statistics and is a contributor to a personality textbook
