Sunday, July 11, 2021

It May be a Mass Delusion that People Are Living Longer

Pretty much everyone accepts that people live longer these days. Lots more folks are in their 90s, 100 no longer seems so exceptional; people dying in their 70s now seems surprising.

But statistics say otherwise (average life expectancy may be rising *slightly*, but mostly due to reduced infant mortality), and scientists absolutely do not share this empirical observation.

This is normal. The public frequently draws unsupportable conclusions, and scientists frequently debunk. But in this case, I've never met a scientist who'd even concede that this is a widespread observation. It's "news to them".

Thanks for playing today's edition of "Who's Hallucinating?"

I think I figured it out.

Life expectancy for a 60 year old has gone from 78 (in 1975) to 83 (today). That’s a significant rise, but can’t account for the massive demographic changes we’ve emperically observed. Hence, my question: is this gap due to statistical aberration or mass delusion?

But those figures are means. We might notice more people in their 90s and 100s even with the mean increasing only modestly. It takes a lot of change to really move a mean. And while we’re observing powerful changes, they’re not happening in bulk.

Today, many people have a centennarian in their lives, while in 1985 very few of us did. That’s a significant observational change, but it couldn’t move the mean very much (unless you built your statistical model around the very narrow issue).

Same for nonagenarians. There seem to be “many more” of them around, but we don’t need vast quantities to foster an impression of change. Say the average person knows four nonagenarians in 2020, versus one in 1990. That could create a powerful impression without significantly budging the mean.

Note I’m just calling out numbers here (e.g. “four nonagenarians”) very roughly and informally.

I’ll call this “statistical aberration” rather than “mass delusion”. It’s actually an example of my “Green M&M Fallacy”. Scaling creates subjective shifts disproportional with the degree of scaling.

No comments:

Blog Archive