Lies, Damn Lies, and Statistics
And how to rescue science
Lies
We all know what lies are. For example, if I toss a coin, and it falls heads, but I tell you,
“the coin landed tails”,
then that is a lie.
Statistics
Many of us also know what statistics are. If I told you,
“There is a 50% chance of the coin landing heads”,
then that is a statistic.
Damn Lies
I couldn’t find a credible definition of “Damn Lies”, so I thought of inventing one myself.
We usually use the word damn “to emphasize or express anger or frustration with someone or something” [2]. More severely, in Christianity, it means “be condemned by God to suffer eternal punishment in hell”. I thought of borrowing from both these usages and defining a Damn Lie as,
“A lie that is so egregious that it should instantly generate anger or frustration in the audience and summarily condemn the liar to suffer eternal punishment in hell.”
Characteristics of Damn Lies
So, what would be an example of a “lie so egregious”?
To me, the worst sort of lie is a lie that we believe to be true. The more people who believe the lie to be true, the more dangerous it is.
For example, consider my previous statement,
“There is a 50% chance of the coin landing heads.”
On the one hand, if you read this is an elementary mathematics textbook that says, “assuming the coin is fair”, then this statistic is true.
On the other hand, what if this statement refers to a real-world coin toss? Would it be true?
Consider why the previous textbook example is true. It is true by tautology; because a “fair coin” is one, by definition, that lands 50% heads and 50% tails. However, in real life, there are no fair coins.
You counter,
“But what if we design a symmetrical coin? Then wouldn’t it be fair?”
The only way we can say an “absolutely symmetrical” coin is “fair” is by conducting a scientific experiment. We toss the symmetrical coin many (say 1,000) times, and if the number of heads is close to 50, we conclude the coin is fair.
But what if we get 511 heads and 489 tails? Is the coin fair or not? This is where statistics comes to the rescue.
Damn Lies and Statistics
Now, suppose the coin was fair. If one were to toss this fair coin 1000 times and repeat this experiment (of 1000 tosses) many times, in all but 5% of the experiments, the number of heads would be between 469 and 531.
Since 511 is in this range, a statistician might conclude that “There is no evidence that the coin is not fair”.
The (somewhat arbitrary) 5% is the probability that we conclude the coin is “not fair” when it is fair, and we could want this probability to be small. “5%” is often a statistician’s preferred value for “small”.
So if the coin were fair, our test would have a “small” probability of getting it wrong (i.e. concluding that it is not fair).
Deduction and Induction
“All dogs have four legs. Snowy is a dog. Therefore, snowy has four legs.”
is an example of deductive reasoning. We have some premises (“All dogs have four legs” and “Snowy is a dog”) and a conclusion (“Snowy has four legs”) that logically follows from the premises. Hence, if the premises are true, the conclusion must also be true.
Inductive reasoning is the reverse. We assume that if the conclusion is true, the premises must also be true. For example,
“All dogs have four legs. Garfield has four legs. Therefore, Garfield is a dog”.
Note, inductive reasoning is flawed; for example, Garfield could be a cat.
“If the coin were fair, then all but 5% of experiments would have between 469 and 531 heads” is a deductive argument. “If the experiment result is between 469 and 531 heads, then the coin is fair” is an inductive argument.
Our statistical conclusion above is based on such an inductive argument and is hence flawed. For example, a non-fair coin could yield the same results, just as a cat has four legs.
What Science Really Says
Our examples about tossing coins and dogs that might be cats are trivial. Sadly, more serious scenarios fall into the same bucket.
Consider, for example, a Vaccine. The WHO says that
“If the vaccine works, then it will provide more than 50% of takers with full immunity”
A sound, deductive argument. However, to test the vaccine, we vaccinate a large number of test subjects. If more than 50% have full immunity, we approve the vaccine. However, this is based on an inductive argument:
“If more than 50% of takers have full immunity, the vaccine works”.
All conclusions resulting from the scientific method are similar. They are not sound conclusions based on deductive reasoning, which we can “know” to be true. They are flawed conclusions based on inductive reasoning, which we can only “believe” to be possibly true.
How Science became a Damn Lie
I’m no anti-vaxer. I’ve taken dozens of vaccines in my life and am impatiently waiting for my COVID19 vaccination. So don’t I “believe” in science?
I don’t believe in science because I don’t “believe” in anything. “Belief” is code for “I don’t know, but I pretend to know”; a dangerous excuse of a word. I do know a few things, and with the many things I don’t know, I’m happy to say “I don’t know” instead of saying “I believe”.
Then, don’t I “know” any scientific theory to be valid?
Science doesn’t say,
“The Vaccine works”.
Science only says,
“If the vaccine works, then we would expect these results”
and
“We got those results”.
If we conclude,
“Therefore, the vaccine probably works”
…is belief based on induction, not knowledge.
Conversely, if, as in the case of many vaccines, the full immunity rate was less than 50%, then we can deduce (not induce) that “The vaccine does NOT work”. Hence, the only truths in science are invalid theories. All claims about “valid theories” are beliefs, not truths.
Hence, again no. I don’t “know” any scientific theory to be valid. I do “know” some to be invalid — but none to be valid.
Sadly, many people, including eminent scientists got say that many scientific theories are true. Richard Dawkins famously talks about the “fact of evolution”. Science doesn’t say that the Theory of Evolution is true; it merely explains the evidence — for now. As Einstein replaced Newton, a new, different theory might explain the evidence better. Hence, Richard Dawkings cannot possibly “know” that that the Theory of Evolution is true. At best, be “believes” it to be true.
This is why, sadly, much of science has acquired “Damn Lie” status, and this is what is really killing science.
How to Rescue Science
Some readers might conclude that I’m an anti-vaxer-intelligent-design conspiracy theorist. But hopefully, the more discerning will see my real point.
Obviously, Anti-Vaxing and intelligent Design are damn lies; since they are lies that many humans believe to be true. Sadly, many advocates of vaccines and evolution have countered one set of damn lies with another set of damn lies. When a fundamentalist theist says,
“God created the world in 6 days, and that is a fact”,
Dawkins says,
“No, the world is a result of evolution, and that is a fact”.
To counter religion, Dawkins turned science into religion; and, thus, nailed it to a cross.
My real point is that just as you can’t fight fire with fire or anger with anger, you can’t fight lies with lies. You have to fight it with something else.
If one is confronted with a lie, the easiest response is the truth. But if we don’t “know” the truth, we shouldn’t pretend. We risk degrading something that might be true into a damn lie (just as Dawkins has degraded evolution).
If we don’t know the truth, the best thing to do is to say, “We just don’t know” — as all true scientists have been saying throughout the ages.
…
References and Notes
[1] https://en.wikipedia.org/wiki/Lies,_damned_lies,_and_statistics
Mark Twain popularized the saying in Chapters from My Autobiography, published in the North American Review in 1907. “Figures often beguile me,” he wrote, “particularly when I have the arranging of them myself; in which case the remark attributed to Disraeli would often apply with justice and force: ‘There are three kinds of lies: lies, damned lies, and statistics.’”
[2] “Damn” — https://languages.oup.com/google-dictionary-en/
- Used to emphasize or express anger or frustration with someone or something
- (In Christian belief) be condemned by God to suffer eternal punishment in hell. E.g. “be forever damned with Lucifer.”
[3] https://stattrek.com/online-calculator/binomial.aspx
