Lies, Damn Lies, and Life Expectancy

“[that life expectancy in Canada is higher than in the USA] is to be expected, Peter, because we have 10 times as many people as you do. That translates to 10 times as many accidents, crimes, down the line.” Bill O’Reilly

via Good Math, Bad Math

4 thoughts on “Lies, Damn Lies, and Life Expectancy

  1. Don’t get me started on all this health care bull ****. Yesterday I got my insurance manual. It’s 120 pages long. Oh? I’m sick? Just let me read through this epic to see how much it’s going to cost me.

    I miss Canada. And yes, health care is objectively better in Canada. And, in the month that I have been back in the US, I have spent more time dealing with health care (signing up for plans, talking about plans, figuring out where I can see a doctor) than I have in my whole life in Canada. (And, I haven’t even been sick here.)

    Please excuse my rant.

  2. There is no doubt that Bill O’Reilly is dumber than a block of concrete, and the statement is clearly wrong.

    … well except for one small doubt of mine (not that it would validate Bill O’Reilly’s incredibly stupid argument, but that it raises a curious question in theory of social networks). In a society consisting of a single individual, there is no opportunity for crime by one person against another. In a society of two individuals, there is opportunity for crime by one individual against another. Therefore you would expect the crime rate to be higher in a population of two individuals than a society of one individual. The question is whether social network effects would cause a higher rate of crime as the society grows (or as the average degree of social interaction grows). The actual crime rate could be modeled as a function of many factors in a social network such as the distribution of physical vulnerability of individuals, the inequity in resources, or the global shortage of resources. The interesting question is how the expected crime rate might scale in a society as the number of individuals grows. As the population grows, it might put pressure on resources (which is probably why rats attack each other in the presence of food shortages – see Prisons also seem to have higher crime rates as the population increases, but that is probably because of close physical proximity.

    It pains me to think that there might be a grain of truth to an argument by Bill O’Reilly, so I would prefer it if there is no scaling effect on the crime rate. On the other hand, I don’t think we need science to prove that Bill O’Reilly is an idiot.

  3. Kevin, homicides in the US are about 0.7% of all deaths in the US every year (~18000 out of over 2.4 million), ranking 15th in the causes of death. For comparison’s sake, the number one cause of death, heart disease, causes over 600,000 deaths per year in the US alone.

    Every year, a whopping 0.0062 percent of the population dies by homicide. Compare that with 0.67 percent of all babies born each year that die at birth (vs. about 0.5 percent in Canada – a very significant difference). A substantially larger portion of the difference in life expectancy is explained by infant mortality than is by homicide – if every one of those homicide victims was aged 10, it would make only 6 months worth of difference in the life expectancy of Americans. If those 0.17% of babies that die in the US but live in Canada had lived to be only 10 years old, that would’ve made up for those six months and then some.

    So put your mind at rest – O’reilly’s a douche.

  4. Not talking about O’Reilly (I saw enough of him in TV last year, when I visited US).

    But modeling crime in social networks is an interesting topic. Okay, sure Kevin is right,
    increased population density has an effect in increasing crime rate. But how about,
    regulating factors to it? I am a postdoc in Singapore right now, where we have
    very high population density and low crime rate.

    So let’s see, some hand waving here, individual i has some probability P_{ij} to
    commit crime to a other individual j (over a social network, I assume we are not talking about complete graph). But this probability is regulated by a some other
    external factors, like fear of getting cought, punishment etc.

    You guys know any models dealing with this kind of situations?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s