Inferring the general from the particular

I typically don’t veer too far from web analytics on this blog, but a couple of news stories have caught my eye this week. One has the headline Ex-Google employee says Google+ has ruined the company (Source: CNN) and  Goldman exec quits, calling firm ‘toxic’ (Source: CNN).

What is remarkable about these stories is that they are the opinions of two single individuals, yet somehow they were able to make national headlines. I suspect a lot has to do with whether we agree with those opinions or not. If we are confirmed in our beliefs then it must be true, even though all we have is one data point. Actually, it’s not even “data” – it’s an opinion.

I am reminded of a memorable passage in Daniel Kahneman’s book Thinking, Fast and Slow:

“Subjects’ unwillingness to deduce the particular from the general was matched only by their willingness to infer the general from the particular.

Kahneman further states that there is a “deep gap between our thinking of statistics and our thinking about individual cases”. Isn’t this what is going on with these news stories? People are of course entitled to their opinions, but are we too quick to assume that they are true and general?

I actually think that we as outside observers have basically no relevant information at all to form a decent opinion. We don’t even know what we don’t know about any of these companies. It’s like being able to only see a single visit with a 100% bounce rate in your web analytics and then inferring that all visits must have a bad bounce rate.

Thankfully we have more data!