Imagine two creatures a few hundred thousand years ago. One is a human that is casually observing the other, a 750-pound Eurasian cave lion. The beast (the lion, that is) lunges at the human (who is not a beast at all – the cranium of humans at that time was about the same as ours, and the cranium of Neanderthals was bigger). The lion shreds our would-be ancestor like so much pulled pork. 

What happened? Why did this human with as much grey matter as you or I, or more, fail to see this coming?

For the first few million years of our evolution, humans drew conclusions in the same way that other animals did – from their senses. Their database was what they personally saw, heard, smelled, felt and tasted. This was a poor method for gathering data from which to draw conclusions about nuanced matters, because the amount of data gathered was small. It was limited by the lifespan and experiences of the single person gathering it.

Conclusions drawn from small data pools are inherently unreliable because they may be skewed by a few outlying points, whereas in larger data pools the outliers get swamped by the in-liers which, by definition, are more numerous.

Toss a coin. It’s not very unusual to get heads three times – 100% of the time – in three tosses. The odds are one in eight. But it’s exponentially harder to get heads 100% of the time in 10 tosses. The odds of that are about one in a thousand.

Now apply that principle to human affairs. A person hundreds of thousands of years ago who’d never seen a lion attack a human, even though he’d seen lions in proximity to humans several times, might reasonably conclude that lions don’t do that. People witness a specific event, and on that basis they generalize.

Extrapolating from specific anecdotal evidence to the general is natural and is even now a legitimate part of the scientific method. But the reliability of this method is limited.

Computer programmers say “garbage in, garbage out” but that’s not quite the right expression for this statistics principle. I’m not talking about input data that are garbage. To the contrary, the input data can be highly accurate. What I’m talking about is not its quality but its quantity. A better expression to capture this principle is “limited data in, limited reliability out.”

As we developed more sophisticated communication methods through complex language and societies, data gathering improved. A person’s database on which to draw conclusions expanded beyond his own experiences and began to include to some extent the experiences of people he talked with.

This was a good thing for the advancement of humankind. Humans surely became careful around lions even though they’d never seen one attack a human, because they were told by others who had seen such attacks that they do in fact occur.

The advent of inexpensive mass communication improved data collection and dissemination even more. The printing press was followed by newspapers and then electronic communication and today’s internet where virtually all public information is available with a few clicks.

You might think, as many did early in the internet, that the availability of all this data would improve decision-making. In theory it could, and in practice it often does.

But it can also have exactly the opposite effect.

People are still wired to absorb a story rather than process data. People go to the movies – and to the internet – to see a story, not to collect data. Stories are fun, data are boring. People especially like stories that play to their primitive instincts such as violence, sex and conflicts with other tribes.

The presenters of internet content – the media – are of course aware of this. That’s why they present clickbait stories instead of data. They’re in the infotainment business.

People draw their conclusions from this clickbait. That’s a problem because although the clickbait is typically accurate (though sometimes not), it’s limited in scope. Stated another way, the problem is not that clickbait is disinformation; the problem is that it’s outlying data. This internet tool that enables us to gather vast quantities of information also enables us to filter it to suit our preferences and biases.

Back to our noble but naïve and dead savage. If he’d been getting his news from Facebook or Fox News or CNN, he’d have been bombarded daily with every gory detail of every lion attack on the whole flat earth. Look what they do with shark attacks. The internet algorithms worsen his plight by collecting his interest in lion attacks in order to feed him still more lion attacks to generate more clicks which generate more advertising revenue.

From those stories of lion attacks, the reptilian part of his brain would conclude (as would mine, since modern brains are arguably no more dexterous than prehistoric ones) that lions are on the rampage and probably under his pine bough bed right now. He’d immediately go into a lion lockdown in his cave where there’s a high probability he would starve to death. But in the meantime, he’d comfort himself with the knowledge that he avoided the low probability of being eaten to death.

The prehistoric human I introduced at the outset got eaten because he had insufficient data to realize that cavorting with lions was dangerous, while his counterpart on the internet starved to death because he had too much data about that same danger without context showing that the danger is not large enough to justify a lockdown. Both made bad decisions that left them equally dead.

Today’s political junkies are similar to the second human, the one who exaggerates the danger of lions because he’s been fed a steady diet of lion attacks. Political partisans on both sides generalize from accurate but outlying clickbait to draw the conclusion that the other side is evil and stupid. I do it myself sometimes. Its takes real effort not to. It was hard-wired into our brains as they evolved in a time when that was the best data collection and analytical method available.  

So, what should we do?

My advice to readers is to get off Facebook. Get off Twitter. Get off Fox News and CNN. Get your news from a variety of sources, and not just sources that filter and spin it the way you like. RealClearPolitics.com is pretty good at presenting points accompanied by counterpoints (though they often fall into the trap of each being extreme).

My advice to the media is to rescue your industry from the clickbaiters. I realize that clickbait is a good business, but there’s a profession calling out to you. It’s called journalism. Try it. Our species is at stake.