Lying with statistics

In My Humble Opinion
1

I recently received the Penn Arts and Sciences Magazine, fall/winter 2009. It has an article titled, “Smoking’s impact on longevity” which gives data divided by sex and smoking status for life expectancy at age 50. But the real purpose of the article, as is made clear in the opening sentence of the text, is to expound the theory that the reason the US does so poorly in global health stats among developed countries is that so many Americans smoke. So among smokers the US is 14 (M) and 17 (F) among 19 developed countries, while among non-smokers the US ranks 9 and 7.

While the data are interesting and show that non-smokers live 2 to 3 years longer than smokers (I would have expected it to be more than that), they certainly do nothing to show that the US health care system is adequate.

Consider the following facts:

  • A disproportionate precentage of people who die between 50 and 65 are smokers

  • At age 65, nearly all Americans get medicare and medicare is what the younger ones lack

and let me conjecture that unisured people are liklier to be younger anyway, then it is obvious that what those data show is that it is people who lack insurance who have the shortest life expectancy compared to their counterparts, say in Canada.

The point would be made most clearly had they started their stats at age 65. Why choose 50 anyway? Most smokers start before 20, a fact the tobacco companies are all too aware of. It would be interesting to see how that would go. A lot of the difference is in mortality before 1 anyway and this is entirely concealed.

The article is unsigned, incidentally, so I must assume it is the dean herself who authorized it. She should go found a tea party.

2

This is General Questions.

Do you have a question?

3

It seems a bit questionable. I checked the life expectancy and cigarette consumption lists and noticed the countries like Greece and Japan and Switzerland consume more cigarettes and live longer than Americans.

http://en.wikipedia.org/wiki/List_of_countries_by_life_expectancy

http://en.wikipedia.org/wiki/List_of_countries_by_cigarette_consumption_per_capita

You might find a better correlation with obesity. These charts suggest that Americans are much more likely to be obese than other countries.

http://en.wikipedia.org/wiki/Obesity#Epidemiology

You don’t need much to move the numbers. Add a couple of years and we are right up there with New Zealand and Norway.

4

I apologize. I actually had two TSDs loaded (inadvertently) and meant this for MPSIMS, although GD or IMHO might be more appropriate. Could a mod move it please?

5

Let’s try IMHO first.

samclem Moderator, General Questions

6

I hope my post isn’t off-topic. It has to do with smoking statistics but cases where statistics that look like they might be lies are actually true! I can’t find links for some of the more surprising statistics I think I’ve seen, but here’s a few that showed up with Google:

Smokers are less likely than non-smokers to die when hospitalized for heart failure.
The oldest people tend to be smokers
In a sample of women, non-smokers were more likely than smokers to die within 20 years

(A fuller view, instead of a single statistic, would not yield the pro-smoking message!)