Faith-Based Science

How statistics is turning science into a religion

Religion and science are often at odds, as people search for the truth and strive for certainty. Is it better to live our lives scientifically, and follow medical prescriptions, or to live according to faith and follow religious prescriptions?

The goal of science is to apply logic and reason to understand ourselves and our world. Science is a product of the rational mind, and requires keen observation and critical thought. A scientific mind must be able to think for itself, create hypotheses that can be tested objectively with measurements and data collection, and reject hypotheses that are not supported by the evidence. It is an inexact mental process, but over time it can lead to deeper understanding.

Religions are based on faith, not on logic or reason. Instead of relying on human rationality as science does, religion relies on writings considered “Holy Scriptures” that are regarded as divine truths which are accepted axiomatically. This means religion is about faith, not reason.

This gives religion an advantage over science. To a religious person, truth is written in the Holy Scriptures for all to see. The search for knowledge is about interpreting these writings. The faithful can feel confident in their knowledge about the world when it is given to them by their religion.

Scientists, however, are not so confident, and shouldn’t be. The search for knowledge using scientific methods is never complete. There are always variables that have been ignored, or haven’t even been considered, that can affect the experiment and affect results. “Certainty” is a goal that can never be achieved, since the human mind is limited in its ability to reason, sense reality, and process ideas logically. Scientific research must therefore be open to criticism, and is a work in progress.

As a result of the inability to achieve absolute certainty in the sciences, mathematicians have developed statistical models to determine the probability that a scientific conclusion is correct. The goal of statistics is to give some confidence to scientific conclusions, with the understanding that certainty is impossible.

The way this works is by somehow turning all scientific observations into numerical data, which is why measurements are essential for science. These numbers can then be crunched in a certain mathematical way, giving such numerical outcomes as mean, average, standard deviation, and other statistical parameters, which are supposed to suggest the likelihood that the scientific results have some degree of truth. It’s basically making educated guesses using statistics.

This means that if something can’t be measured, then it cannot be scientifically studied. This is an advantage of religion, which does not need any evidence, apart from the written word in religious texts.

Some statistical approaches we were all taught in school, like how to get the average of a set of numbers. However, statistical methods have evolved into very complex equations and formulas, and require computer programs to run. There was a time when you could read a scientific article that shows its data and statistical conclusions, and you could verify the statistics. But you can’t do that anymore with modern research.

These days, scientists search for truth by using artificial intelligence and highly complicated statistical models. In medical research, for example, mountains of data from huge databanks are used to find meaningful connections between genetic or other factors and disease. This often includes meta-analyses of multiple studies, each done differently, making it difficult, if not impossible, to come to any conclusions without the aid of complex statistical models. Modern statistical models take inhuman amounts of complex, confusing, contradictory, and even false information and applies an almost magical mathematical method of making sense out of this chaos, yielding a probability determination of whether A is connected to B.

For the average layperson trying to read this research, there is absolutely no way you can challenge study results. Unfortunately, for the average scientist, it’s the same problem. Statistical methods have become black boxes that take in chaos and put out probabilities. You can’t crunch the numbers yourself to see if the statistical conclusions are valid. At one time, you could do that by using Chi square and other simple statistical methods. But those methods are in the dark ages compared to the computer-driven, AI-driven statistical models used today.

Statistical models have become like oracles. You feed them data, and the math does the rest, spitting out a number that tells you that chances of something being true or false. And since the statistical model may conclude mathematical associations that are subtle, strange, and not obvious, you need to accept its assessment without the ability to argue with numbers. The numbers speak for themselves, and you need to listen.

This means that scientists must have faith in these statistical models and the numbers they generate. They can’t run the numbers for themselves, even if they wanted to. We are expected to take the conclusions of these models as true — on faith.

And this is why science is becoming a religion. It is relying on faith in statistical methods that cannot be easily verified or refuted, even by scientists.

Keep in mind that statistical significance is not necessarily practical significance in reality. Statistical significance is a term that means that the statistical black box found an association that is probably real from a math standpoint, but that association may be insignificant from a practical standpoint. For example, there may be a statistically significant increase of 1% for a disease due to a particular factor, which means that you can feel confident that the study looking into this association did indeed find a 1% increase. But who cares about a 1% increase? It’s insignificant from a practical standpoint, despite being significant from a statistical standpoint. Conversely, there could be a study that found a 50% greater risk of a disease, which is highly significant, but the statistical analysis may say that there is not enough data, making the results insignificant.

This means that statistics decides what passes as truth. This is why anecdotal evidence, or just common sense, are not considered valid in scientific circles. Statistics relies on lots of data points, the more the better. You need to “prove” a theory with lots of data, not just accept a theory on personal experience or insight. And if a theory can’t be proven, it can’t be considered true, even if it is true. On the other hand, if the statistics say something is true, it must be considered true, even if it seems wrong.

This means that scientists subordinate their sense of truth and reality to the black box of statistics. They need faith in their statistical models to determine if what they see in their research is real or not. These models are opaque to the average scientist, not to mention a layperson, so faith is essential. And this makes science a religion.

Uncertainty is the problem. Scientists strive for certainty, and say they rely on reason and logic and data, but they need faith in their black box statistical models to reach any conclusions with any degree of probability. They can’t think for themselves, or broaden their theories beyond what they can measure. They are stuck in their mathematically-defined models, boxed in by statistics.

In the end, the scientist is just as faithful in his methods and models as the religious person is in his Holy Scriptures. But the difference is that scientists think they are rational and committed to using facts, proof, and reason to discover reality, while the religious-minded manage their uncertainty by having faith in their doctrine. In the final analysis, there is really no difference, and I say that with 95% confidence.

The solution is to free our minds from the limits of math and its stats. We need to see past the blinders of statistics. Science cannot see anything that the black box does not condone.

While mathematics and statistics have clearly enabled technological progress, it hides from our minds those aspects of life and reality that are not measurable and testable. We are both aided and limited by statistics. There is no doubt that statistics and probability are valuable to decision-making. Any gambler will tell you that. But odds are that statistics have killed many a great idea, simply due to lack of enough data to reach a statistically valid conclusion.

It is just as dogmatic to rely on statistics to define our world as it is to rely on Holy Scripture. Each requires faith, although religion deals with certainty and absolutes, while science deals with probabilities and ever-changing theories. So whether you believe in science and its predictions, or in religion and its proclamations, there’s no big difference.

Keep in mind the famous words of computer scientist George Fuechsel, who in the 1950s coined the term, “Garbage in, garbage out”. Statistics are only as good as the data fed into the black box. Alternatively, the data used to feed into the box can be selected to make the box have a certain output. This way the statistical narrative can be managed, which is a very unscientific thing to do. As Mark Twain was quoted as saying, “There are lies, damn lies, and statistics.”

Twain was also credited with saying, “Figures don’t lie, but liars figure”. The mathematical nature of statistical models gives the impression of objectivity, but don’t believe the numbers if it defies common sense, or if there is an agenda behind the facade of figures.

Statistics also rely on data, and data relies on studies, which rely on funding. Therefore, statistics rely on funding, so follow the money before you decide on faithfully believing the numbers.

Be careful about confirmation bias. When the statistics are on your side, you are more likely to believe them. Conversely, you will more likely reject numbers that defy your bias. This is a human thing to do, and scientists are not immune from being biased, especially since grant money is probably a biasing factor. Biased scientists feed biased data into the black box, and get out biased results, which may be statistically significant!

We need to use common sense, and our intuition and innate knowledge, instead of relying on statistics. There are things we know, deep down, that are instinctive, and feel true, and we judge the truthfulness of what we are told by this innate standard of truth. Our common sense is a survival tool. We need to use it when assessing scientific information, regardless of what the statistics say. They are likely brought to you by some self-interested party.

Finally, get comfortable with uncertainty. It’s the human condition.