The data is what you want it to be

I was just browsing kottke.org when I came across a short little post about a neat page on Wikipedia aptly called “List of common misconceptions“. The post contained an excerpt of that Wikipedia page on life expectancy, a misconception that I myself had (somewhat embarrassingly) up till only recently:

It is true that life expectancy in the Middle Ages and earlier was low; however, one should not infer that people usually died around the age of 30. In fact, the low life expectancy is an average very strongly influenced by high infant mortality, and the life expectancy of people who lived to adulthood was much higher. A 21-year-old man in medieval England, for example, could by one estimate expect to live to the age of 64.

I’d always wondered what it felt like to live in the middle ages, where people died on average forty, fifty years earlier than they do now: What would I do differently? Did people get “old age” issues younger? Were there great grandparents? 

When I realised that infant mortality was the one greatest factor in affecting average life expectancy, it opened my eyes up to the possibilities of data story-telling.  You can pretty much tell any story you want, with any sort of data, depending on what you leave in or out.

The fact that the average life expectancy in one place and/or time is two-fold that of another place and/or time doesn’t really mean anything without context.

That life expectancy in Monaco is 87 years while life expectancy in Sierra Leone is 47 is a fact.

That Sierra Leone’s average life expectancy is affected greatly by infant mortality is a story. That Sierra Leone’s average life expectancy is affected greatly by deaths of both mother and child during childbirth is a story. That Sierra Leone’s average life expectancy is affected by poor access to healthcare services in rural areas is a story. That Sierra Leone’s average life expectancy is affected by civil war is a story.

That Monaco’s life expectancy is helped by having the lowest infant mortality rate in the world is a story. That Monaco’s life expectancy is helped by compulsory state-funded health services is a story. That Monaco’s life expectancy is helped by their Mediterranean diet is a story.

In my work as an analyst, I work with data quite a bit. Many times someone would come up to me and ask, “so, Donn, what does the data say?” And I can’t help but answer that question with another: “What do you want it to say?”