Hi A fantastic site for looking at this kind of data over time and for many countries is www.gapminder.org.
One way that you can get some idea about the issues raised by Mike is to look not only at statistics like life expectancy at birth, but also at child mortality rates, which is a major factor in longevity (thank heavens for Joseph Lister, vaccines, ...!). As to the increased longevity of higher SES women and not lower, the decrease in smoking rates has been much greater for higher SES than lower SES, thus producing an even more marked relationship between SES and mortality than existed earlier. The NY Times had a piece on that a few years ago. Take care Jim James M. Clark Professor & Chair of Psychology [email protected] Room 4L41A 204-786-9757 204-774-4134 Fax Dept of Psychology, U of Winnipeg 515 Portage Ave, Winnipeg, MB R3B 0R4 CANADA >>> "Mike Palij" <[email protected]> 24-Mar-13 6:59 PM >>> The NY Times Sunday magazine has an interesting article on life expectancy and the factors that affect it. See: http://www.nytimes.com/2013/03/24/magazine/who-lives-longest.html?hpw&_r=0 It starts off with saying that a Swedish baby born in 1800 had a life expectancy of 32 years -- Sweden was the first country to keep extensive records of births and deaths and allowed one to calculate such a value. Now, to tell the truth, I don't know if "life expectancy" as reported here is a simple descriptive statistic (i.e., mean or median time to death) or is calculated in some other way but the article goes on to point out how misleading this value is *if one considers it as a point estimate*. Because so many infants and children died early in life, the implication is that the distribution of the variable "age at time of death" is a seriously positively skewed distribution (whether it a skewed normal or some other distribution is an empirical question that is not addressed). What is missing is a measure of variability for "age at time of death", something like the simple range but that is likely to be uninformative (e.g., what is one to make of the range 0-70?). The interquartile range might be more informative because it would represent that age range for the middle 50% which would be more informative (one could also argue that if the above life expectancy of 32 years is the arithmetic mean, then perhaps some robust measure of central tendency should be used). So, for pedagogical purposes, the article provide a basis for comparing an argument based on a point estimate (like the life expectancy of 32 years) versus an interval estimate like the interquartile range or some other measure of variability. The rest of the article is also of interest because it goes into why a measure like life expectancy, beyond the point vs. interval distinction, is problematic because it might cause one to focus to much on longevity and how to increase instead of focusing on other factors that influence it, for example, the article's example of white women with college degrees having their life expectancy increase by 3.5 years during the period of 1990-2008 while women without a high school diploma lost 5 years. How does one fix that? -Mike Palij New York University [email protected] --- You are currently subscribed to tips as: [email protected]. To unsubscribe click here: http://fsulist.frostburg.edu/u?id=13251.645f86b5cec4da0a56ffea7a891720c9&n=T&l=tips&o=24505 or send a blank email to leave-24505-13251.645f86b5cec4da0a56ffea7a89172...@fsulist.frostburg.edu --- You are currently subscribed to tips as: [email protected]. To unsubscribe click here: http://fsulist.frostburg.edu/u?id=13090.68da6e6e5325aa33287ff385b70df5d5&n=T&l=tips&o=24506 or send a blank email to leave-24506-13090.68da6e6e5325aa33287ff385b70df...@fsulist.frostburg.edu
<<attachment: Jim_Clark.vcf>>
