Smoking saves money

Philip Morgan

Insight about risk is like catnip to me. Stories about the reality of research are too. Put the two together and you've got catnip and… cocaine! Cocainip! Sign. me. up!

Peter Sandman writes about risk communication and how we humans respond to risk. This, from a recent blog post of his, is fascinating:

In 1981 I began volunteer communication work with an American Cancer Society state chapter. Then as now, a big ACS activity was corporate smoking cessation programs. To help sell these programs to companies, we commissioned a study of how employee smoking affected companies’ bottom line. We expected to show a big cost due to tobacco-related medical expenses. Instead, the study found that employees who smoked saved their companies both pension money and healthcare money by dying more rapidly after retirement. Smoking cessation simply wasn’t in a company’s economic self-interest.

We suppressed the study results and kept telling companies they would save money by sponsoring our smoking cessation programs. I lost my argument for dropping the false claim. Health is a higher value than truth, I was told.

Source: http://www.psandman.com/col/Corona64.htm


Point of view is both content and context. It is both what you say to your market, and why you see things that way.

A point of view that sees things in a certain way because it uses data to understand the world has the tailwind of our culture's bias towards data. Tailwinds are nice!

But the point of view that uses data to understand the world also has to grapple with issues like Peter Sandman discusses. What if the data is inconvenient? Cuts counter to the dominant narrative? (Or the narrative we'd like to establish as dominant?) Or adds complexity to an otherwise simple narrative?

I think Sandman's core argument in the aforelinked is that people are able to understand properly-communicated data, and making this data available -- even if the data complicates a simple narrative -- is essential to being trustworthy.

I think I agree.

-P