Statistical Crystal Ball

Dean Foster, Lyle Ungar, and Choong Tze Chua at the University of Pennsylvania have created a mortality calculator. It’s pretty cool–you enter all kinds of information about your health, habits, family history, etc., and it predicts how long you’ll live. Not to brag, but my predicted life span is 94 years, with upper quartile 103.99.

Dean Foster’s website also links to Northwestern Mutual’s Longevity Game . I like the Penn version better–Northwestern Mutual has multiple choice questions, and sometimes the choices don’t seem all-inclusive. For example, their alcohol question has the following 3 choices:

(a) Don’t drink or never drink more than 2 drinks a day
(b) 3-4 drinks more than 2 times a week
(c) 5 or more drinks at one time, more than once a month

I sometimes have more than two drinks in a day, but I don’t think I have 3-4 drinks more than 2 times a week. If I choose (a), my expected lifespan is 92 years, pretty close to the Foster et al. calculator (which asked for number of alcoholic drinks per day). If I choose (b), it drops down to 88 years (no error bars reported).

I wonder if my numbers are inflated, though–my family history is free of almost everything, but that could just be because my parents are still fairly young.

This entry was posted in Uncategorized by Sam. Bookmark the permalink.

2 thoughts on “Statistical Crystal Ball

  1. The questionnaire seems to overlook people who have themselves had cancer; it only questions whether close relatives have had it. Isn't this a serious flaw in the calculation?

Comments are closed.