The Fallacy of Observational Statistics

Here's how the argument usually goes: I was at site X where there were a lot of photographers, and most of them had Brand Y. 

One way this plays is like this: "I was at the Grand Canyon recently and all I saw were DSLRs. Almost nobody had a mirrorless camera." The usual conclusion is then "mirrorless cameras must not be selling well." 

Unfortunately, the second statement doesn't derive from the first observation. One data point does not make a measurable trend. Second, almost no one who makes such a statement actually enumerates it. They don't say they saw 100 different people photographing, and that 20 were using Canon DSLRs, 15 using Nikon DSLRs, 35 were using cell phones, 25 were using compact cameras, 4 were using mirrorless cameras, and the other person was using "other." Nor is the observation given a measurement time ("in the hour I counted…") or was it repeated. Just a random observation.

A couple of years back I spent two full days at the Snow Monkey pool in Japan (not natural at all, more like a hot tub maintained for snow monkeys, which was a bit disappointing). Because I was with a group of DSLR users and we stayed all day, and because buses would occasional come and drop off a gaggle of tourists who'd stick around for a short period, photograph, and leave, someone "counting" cameras there would have had two remarkably wrong statistics had they only counted for a short period.

First, in between buses, the only ones there were those of us using DSLRs. Statistical observation: "everyone is using DSLRs." Since we were a small group, every time a bus came, we were badly outnumbered and most of those people were shooting with cell phones. Statistical observation: "most people shoot the snow monkeys with cell phone cameras." Now, I happened to have a mirrorless camera with me at the time, and was using it, plus I saw a couple of others here and there in the busers, so yet another statistical observation might be: "I'm starting to see mirrorless cameras being used." 

So which of those things is correct?

But there's another problem lurking in the background of observational statistics: cumulative sales. DSLRs have sold well over 60 million units in the past decade. How many mirrorless cameras have been sold total? Probably not even 1/10th that number. Even if the mirrorless camera sales were replacements for DSLRs you wouldn't see "mostly mirrorless" cameras in the wild yet. It takes awhile for installed bases to change, and we haven't reached a tipping point yet, though in some places in Asia the numbers are starting to skew.

So don't judge whether mirrorless cameras are successful based upon random observations in the field. From a sales perspective, mirrorless cameras are quite successful in Japan, successful in most of Asia, heading towards successful in Europe, and only here in America do we see success still lag.

text and images © 2012 Thom Hogan -- All Rights Reserved   //    Follow us on Google+: Thom Hogan or on Twitter: @bythom, hashtags #bythom, #sansmirror