Lately the themes have been a bit sensitive and heavy, so without calling it a course correction, I'm shifting back toward thinking for a bit.
The question: do you need objective data to prove your claim is right?
The answer is probably "not necessarily."
Using objective data to decide or verify a hypothesis is normal now, but that approach still has a short history.
Medicine, for example — with how often EBM comes up — is one field where decisions rest on the most credible data available.
The span for stacking data, like clinical trials, is long, and decisions are careful.
Even so, as the famous story goes, in the 19th century Nightingale used huge amounts of data in wartime field hospitals to show what was causing patients' conditions — and to the people around her (officials and so on), it apparently sounded dubious.
And if we go further back, in ancient Greece the idea of "data" barely existed. I can't really picture Aristotle saying in debate, "Gentlemen, look at this data that supports my claim." (Then again, maybe he could have. Statistical methods, though — probably not.)
But the absence of a data-based approach doesn't mean their claims lose validity.
Of course you could later pull data in to strengthen them after the fact.
For the sharp readers: I'm not about to yell "Down with data-driven!" or "Fuck you data!!"
If anything, whatever you call "data-driven," I welcome best practices across organizations, policies, and opinions — and those best practices often rest on plenty of solid data and proof (history), so in a broad sense that's data-driven too.
Still: data is not always required for an accurate claim.
I like arguing inside my own head about hypotheses I form — inventing counterarguments myself — and I think there are cases where explaining everything with word-logic alone is better.
The core premise of data-based decisions, as I see it: "Stack (collect) data, extract the best method from it, and produce an effect that beats the old method."
Setting aside ways of monetizing data itself, the goal of that flow is obviously "produce an effect that beats the old method."
So in the extreme, until you hit that goal, everything before it is sunk cost.
Until "produce an effect that beats the old method" is achieved, the strongest method dreamed up by a data-analysis specialist and the strongest method dreamed up by some idiot kid (me, for example) are the same.
...I wrote that far and it was starting to sound like an anti-data-driven rant, so let me shift the axis.
As I said at the start, I like data-driven as "best practice," and I use it in many situations.
But logic built only from words, if it covers contradictions and gaps, is extremely strong too — and as a method it doesn't lose.
Hmm. Having written this far, it didn't become a very sharp insight.
Either way, I want to keep insisting on hardening logic to the limit in my own words.
Tweet© 2020 ZUUHE