Flying Is Safer Than Driving

If the study of decision theory, probability, and statistics is to help us in daily life, then we should be able to point to decisions we have made based on rational choice. That does sometimes happen, but often we grossly mis-estimate the relative risks of various activities. We all know people who are nervous about flying, but have no problem driving to the airport. Arguing with such people that they are more at risk during the drive is unlikely to get them to change their minds even though on a death/passenger mile basis, flying is about 34 times safer than driving.

What about terrorism? Since even with 9/11 we have not experienced as much carnage as some other countries, let’s consider Israel, where terrorist attacks are much more common than in the USA. The latest figures I have are for 2002 in which 238 people were killed. Look at the total population of Israel and the USA and check out the total number of deaths in 2002 due to automobile accidents in the USA, and you will compute that the probability of being killed in the USA by an automobile accident is three times higher than being killed by a terrorist attack in Israel during the same year. This is not a joke or trick of mathematics. Almost any reader would, without thinking, believe that the danger of being killed by a terrorist attack in a hotspot like Israel would be much greater than being killed simply going about business in an American automobile.

Do you think more people were killed in New York in 2001 by automobiles than by terrorist attacks?

Is the quality of life getting better or worse? If you look at the newspapers, you can be overwhelmed by the daily listing of killings, rapes, and corporate misdeeds without even leaving our shores to consider the wars. Are the news media a good source of data about the quality of life? I say no because for some perverse reason, we all seem to enjoy reading about bad things, but get bored by most good things. If that were not true, the profit-driven media would feature good things. Why did the killing of a pregnant woman in California gets so much nationwide coverage when many people are killed every day? During that same period that killing was being featured, how many good things happened? Why did my friend who is still alive ten years after getting a new heart not become famous?

Most of us base our internal assessment of risks on our daily experiences. Most of us drive every day and rarely have an accident. Most of us fly relatively infrequently and read about aircraft crashes when they happen. Based on this skewed sampling, we can easily come to the conclusion that flying is more dangerous than driving. Not many of us are willing to comb through the national accidents statistics that are published annually by the federal government to determine what activities are safer. (BTW, do you know what is the most dangerous room in your house? How dangerous is it? We think motorcycles are more hazardous than automobiles. Are they? If so, how much? Are bigger, heavier cars safer than smaller, lighter ones? If so, how much?)

So we have two issues: (1) assigning incorrect values to the relative risks of activities due to biased sampling, and (2) failure to act on accurate risk analysis in a rational manner.

The second point comes under the general heading of psychology and not really part of this series. The first point, learning to collect unbiased data and assign correct relative values to risks is only peripherally related. Most exercises in decision theory start by assuming you have good data on which to make the relative risk assignments, and a payoff matrix. The best decision theory analysis in the world is worthless if it is based on faulty data. The best decision theory analysis in the world is worthless if it is followed by irrational behavior.

None of this discussion will likely influence how you drive (gas prices will do that!), but what happens when our government makes decisions? Do the various agencies get the sums right when assessing the relative value of the risk of pursuing programs? If they get the sums right, do they behave rationally based on the analysis?

In response to the interest my original tutorial generated, I completely re - wrote and expanded it. Check out the tutorial availability through Lockergnome. The new version is over 100 pages long with chapters that alternate between discussion of the theoretical aspects and puzzles just for the fun of it. Puzzle lovers will be glad to know that I included an answers section that includes discussions as to why the answer is correct and how it was obtained. Most of the material has appeared in these columns, but some is new. Most of the discussions are expanded compared to what they were in the original column format. I have also recently posted a greatly expanded and improved tutorial on tutoring seniors in computer literacy.

In response to the interest my original tutorial generated, I have completely rewritten and expanded it. Check out the tutorial availability through Lockergnome. The new version is over 100 pages long with chapters that alternate between discussion of the theoretical aspects and puzzles just for the fun of it. Puzzle lovers will be glad to know that I included an answers section that includes discussions as to why the answer is correct and how it was obtained. Most of the material has appeared in these columns, but some is new. Most of the discussions are expanded compared to what they were in the original column format.

[tags]decision theory,statistics,probability,good news,terrorism statistics,newsworthy events,bad news[/tags]