The Ladder of Inference

The other day, a message in one of the ASQ forums pointed me to a short video (under four minutes) about a psychological concept called The Ladder of Inference. Apparently this idea has been current for years, but somehow I missed it. But it was coined by the business theorist Chris Argyris, and it neatly summarizes how our perceptions and assumptions interact to generate beliefs and actions.

The Ladder of Inference has six or seven rungs. (The video below lists only six, but other sources online give a seventh.) They proceed, from the bottom up, as follows:

• observable data
• selected data
• interpretations or meanings (This level is silently dropped in the video.)
• assumptions
• conclusions
• beliefs
• actions

Expressed in more narrative form, the idea is that our minds operate like this:

• I start by perceiving stuff out there in the world.
• There's lots of stuff going on and I can't possibly track it all, so I select a small subset to focus on and forget the rest. 
• I use this small subset of the data I've perceived to generate interpretations. These are just stories I tell myself about what the data mean.
• I use these interpretations to build operating assumptions for approaching the world.
• Then I use the assumptions I have built up to evaluate new data in order to form conclusions about what is happening around me.
• These conclusions harden into beliefs.
• And finally my beliefs are the background underlying the actions that I take in the world.

But the fun part is that this whole structure is self-sustaining, because I use my assumptions and beliefs as a filter to simplify the selection of new data. Once I have built up a certain set of beliefs (never mind how they got there or how bizarre they might be), I use those beliefs in the ongoing process of perceiving new stuff. Therefore when I select which data I'm going to remember, I tend naturally to choose whatever agrees with the beliefs I already hold. If you ask me how I can be so sure of those beliefs, I tell you that I see them confirmed every single day. And in a literal sense, that's absolutely true! Of course you might think my beliefs are pretty weird. But because I use those beliefs to cherry-pick which data I bother to remember—and because the whole process takes place subconsciously—I remain convinced all the while that my crazy beliefs are perfectly empirical.

It gets better. Remember that my beliefs are the basis for my actions. If my beliefs are weird, my actions might be a little weird too. Those actions will cause others to react to me in a certain way, and from their reactions I derive new perceptions with which to start the cycle all over again.

How can we get out of this trap? By recognizing how the process works, and then taking steps to make it less automatic.

First, I should question my assumptions and conclusions. I should watch to see what kinds of data I regularly select and remember; and also what kinds of data do I regularly reject or forget?

Second, when I notice all my data tends in a common direction, I should seek out contrary data on purpose. If there is none, well and good. But if I can find it, then it's time to relax my assumptions and broaden my beliefs. 

As a method, this sounds simple to describe. It's no more than disciplined skepticism, or what I have described elsewhere as "the scientific method applied to business." (See for example this post or this one.) But that doesn't mean it is always easy to do.