PolySciFi Blog

In response to this post, commenter San, Juan made some good comments on why people are bad marginal utility maximizers which I feel should be given a wider read...¹

I read an interesting theory presented by Terry Burnham and Jay Phelan in their book “Mean Genes”, which relates to why humans (in general) are bad utility maximizers (perhaps maximiser would be a more appropriate spelling in this case). They claim the reason behind our ineptness is two-fold. First, as a species, we derive pleasure from taking risks, and second, we are bad at calculating the odds for the risks we take.

The reasons behind our love for risk are biological. Risky behavior stimulates the dopamine reward systems in our body. Some people are born with systems that that muffle the buzz they get from taking risks. These individuals with unusual dopamine receptors – and hence reduced stimulation of this pleasure pathway -- will go to extreme lengths in pursuit of the dopamine high. They are the risk freaks; impulsive and extravagant: bungee jumpers, explorers, and the highest of rollers in Vegas (studies have shown these individuals also tend to prefer spicier foods.)

The gene which influences this type of behavior is referred to, by the press, as the “novelty-seeking” gene. Recent evidence has also shown a strong link between the prevalence of this single gene in a population and how far that group has migrated. This makes sense from an evolutionary perspective as well. Humans first arose in East Africa, and spread out to cover the rest of the world. If we imagined two types of humans, those who cowered in their caves, and those who explored new areas…. we can come to the logical conclusion that, although many of the risk takers may have died, those who gambled and won populated the entire globe. So we are risk takers because we are the great-great….grandchildren of humans who placed risky bets.

As far as the second point about us being horrible mathematicians, Burnham and Phelan provide an explanation that is not quite as clear. They make a case for certain flaws in statistical reasoning, such as predicating causes of death. For example, Americans typically believe that natural disasters are a bigger threat to life then diabetes, yet in 1997, 62,636 Americans died from diabetes compared to the 227 from tornados, floods, and lightning. The authors claim this skewed fear is based on things more likely to harm are prehistoric ancestors. We most wildly overestimate the risk of death from accidents, homicide, venomous snakes, and pregnancy compared to our underestimate of death from disease or vaccination.

This idea of living in a modern world which is very different from our ancestor’s is extrapolated to explain why we can’t calculate expected utilities in other situations.
One possibility is that our ancestors evolved in a world with very few people, many groups were made up of less then 100 people, and not long ago on a genetic time scale there were only 18 thousand humans in the entire world. So people care less about odds and focus on the prizes. It’s still tough for us to wrap our brains around the enormous size of our modern population, and when we see a lottery winner on TV, we think…I can be next.

One final point made in the book was that in certain situations, we risk too little. The social arena is one example. For ancestral humans, social failures were presumably much more costly than they are for us. In the small populations that our ancestors lived in, they were doomed to hear about their social mistakes for years as the same group joked around the campfire. Perhaps worse than jokes, were the consequences of offending the wrong people within a village…banishment would often be a death sentence. Nowadays we should be much bolder at the local bar since chances are we’ll never run into our rejecter again. (Personally, my experiences with this one have been more along the lines of Murphy’s Law then correct statistical odds.)

I think San point on why people gamble is spot on - a) dopamine high, b) we tend to focus on the prize and c) we really don't understand how many people there are in the world and thus muck up our statistics.

I also think his assessment as to why we poorly assess probabilities of death is partially correct, but I think it's changing.² Certainly, there has been some calibration lag as to what we should fear, but we're catching up. As we've started living longer, our fears have come more in line with the modern causes of death. Cancer, diabetes, AIDS, Alzheimer's, heart disease, and lots of other diseases were not feared by out ancestors yet are very much feared today, and a lot of societal energy is put into combatting these diseases.

Specifically, I think that we tend to especially fear the following things: things that are new, things we do not understand or feel that we have no control over (or are unpredictable), and things with particularly disturbing results (the reverse of San, Juan's gambling rationale).

Accidents, homicides, and natural disasters fall into the second category. Ebola falls into all three. SARS into the first and BSE into all three. How much we fear a particular kind of death (here's my favorite way to die) is a function of the sum of the contributions from each of these three components. If something is very new, we fear it, even if it's no worse than the flu (ala SARS). Ebola, other than the the low probability of it occuring in the US outside of Reston, VA, (but how many of you are willing to take a boat ride on the Ebola river?), is nonetheless quite feared because it's not well understood, is still relatively new, and is just a horrible way to die.

As further support for my thesis, consider the public reaction to AIDS. In the 80s, AIDS was quite the scary disease as it was new, we didn't understand the disease much less its transmission, and the results were quite scary (like Kapposi's Sarcoma). Since then, the disease is obviously not as new, we have a pretty good understanding of the disease and its transmission (don't share needles or have gay butt sex - other kinds of gay sex are statistically ok), and thanks to improved treatments, the results are not so bad (see Magic Johnson). I feel safe in saying that the public's fear of AIDS has dropped significantly.

So keep up the interesting comments. Hearing others' thoughts in response to my musings is one of the primary reasons I post, plus they're fun to read.

Footnotes
1. Note: I added a hyperlink to San, Juan's comments to link to the Mean Gene website and inserted blank lines between paragraphs for readability. Other than these minor changes, San, Juan's comments are as originally posted.

2. I think we overestimate, not underestimate the probability of death or injury from vaccination (see the Autism-vaccination scare or the Nigerian fear of polio vaccines), but that's secondary and not crucial to the larger point, and hence relegated to a footnote.
Comments(1) |