Lifespan Calculators Don't Add Up

If you've ever taken an online life expectancy test, don't be too satisfied -- or too worried -- about the result. The Wall Street Journal's Carl Bialik tried several and got a fourteen year spread in his estimated time remaining. One actuary told him "I don't think any of them are very good."

My own results were closer -- only a three year spread among three tests -- but each test emphasizes different factors, and no doubt weight them somewhat differently.

I looked at LivingTo100.com, a calculator run by a professor at the University of Pennsylvania's Wharton School, and MSN Money's tool.

One of the primary differences is how they evaluate your stress level. Stress is a killer because of the physiological problems associated with it, such as coronary artery disease and cardiovascular disease, and also because it can provoke other dangerous problems such as substance abuse, overeating, and physical inactivity. The three calculators determine your stress level very differently. LivingTo100 asks "How do you usually cope with your stress<" and gives you four choices such as "Very Well. It helps me to get motivated" or "Not very good. Stress eats away at me and I can not seem to shake it off." You then check from a list of items that you think are currently stressful to you. The MSN calculator doesn't trust your opinion; it asks whether a friend would describe you as "easy-going and relaxed" or "aggressive, intense, and quick to anger." It seems odd that those are the only two choices. The Wharton quiz is the most objective, asking whether you've experienced any of ten specific stressful events such as the death of a family member or divorce / separation.

Another difference was in auto use. The Wharton test is especially concerned with how much you drive, asking six questions to measure risk whether you or someone else is driving. The MSN test has four car-related questions, but LivingTo100 doesn't ask any.

While all three considered diet, LivingTo100 focused more on it than the other two with questions about processed and red meat consumption.

LivingTo100.com asks your marital / relationship status, which the other two ignored. Another unique question at that site was the number of new friendships you've made in the last twelve months (and they don't mean Facebook friends!). MSN is the only one to ask about marijuana use, while the Wharton test is unique in asking for family income.

Overall, the fastest was the MSN test while LivingTo100 was the most detailed. The Wharton test was in between but has the advantage of being a single page, rather than ten or so different pages you have to click through.

Unfortunately, the sites don't reveal the weighting they give factors, or why they emphasize some and not others, so there's no reasonable way to evaluate the lifespan number they provide. A more open scoring system would be a plus for any site in this space.

The WSJ's Bialik observed that the tests he looked at don't consider correlations and interactions. Each answer adds or subtracts years from a baseline lifespan, but the formulas don't consider whether answer 3 has a more significant impact than average because of the answer to question 7.

So these are interesting experiments that will give you a feel for the things you can do to extend your life (don't smoke, drive safely, eat right, exercise ...). But take the accuracy they pretend to have with a grain of salt.