Anti Confirmation Bias: to be unbiased you need to bias your bias
/By Duncan Anderson. To see all blogs click here.
Reading time: 4 mins
Summary
Confirmation bias = the tendency to search for, interpret, favor, and recall information in a way that confirms or supports one's prior beliefs or values.
Anti confirmation bias = IMO what is needed to counter ‘confirmation bias’ to attempt to be unbiased
Confirmation bias = overweighting information on why your idea is correct and underweighting information that goes against your idea
Anti confirmation bias = looking for how your idea can be wrong (AKA ego distortions)
Anti confirmation bias = looking for what you have missed (AKA blind spots)
Jingle: to be unbiased you need to bias your bias.
We come preloaded (baked into our source code) with a number of cognitive biases.
IMO you don’t get to choose whether you have cognitive biases, you just get to choose how you deal with them.
‘Anti confirmation bias’ is my attempt to counter ‘confirmation bias’.
Here is a nice list of some well known cognitive biases.
The closest possible connection to reality is the basis upon which a good life, business, etc is built.
For your own ideas… IMO default is confirmation bias, AKA you see your ideas as better than they actually are.
“We see the world the way we want it to be, not the way it actually is.”
IMO you want to ‘lean against’ ‘confirmation bias’ with ‘anti confirmation bias’.
Effectively if you are thinking about your idea or doing user testing etc I think you want to try and consciously spend more of your time thinking about why your idea is wrong and what you could be missing than trying to see why you are right.
A rough rule of thumb I have:
⅓ of the time on why your idea is right
⅓ of the time why your idea is wrong
⅓ of the time on what your idea is missing
If you can do this then hopefully you can have a relatively strong understanding of reality.
Some ways to try and route out confirmation bias AKA how to do anti confirmation bias
If I’m doing user testing or just speaking to someone about the idea and I only have 5 questions at the end I’ll ask something like the following:
Preamble: we are obviously looking to build this product to help people. We’ll find out one way or another if we are helping… but the sooner the better. So, if ok, please let us know what you think. ‘Our baby might be ugly’, we need to know ASAP so we can try and improve the product / idea!
What did you like?
What didn’t you like?
What would you change?
Is there anything you think we have missed?
Would you recommend this to someone else, if so who please? If not, why please? (This is an NPS question as is an orthogonal way of asking if they actually like the product. )
Comment:
I find that this sequence of questions can allow someone to be more frank with you, as you are warming them up, making the space to say something good, then actively asking for constructive feedback which makes the space for someone to say things.
IE people are often much more comfortable saying something constructive if they have first said something positive!
There are many many more questions you can ask, but basically I try to spend ⅓ of time on why the idea might be a good one, ⅓ on where we could be wrong (ego distortion) and ⅓ of the time one what we are missing (blind spot).
If I’m trying to develop an idea myself I try to lean against ‘confirmation bias’ through thinking about things like this:
"Common sense is the collection of prejudices acquired by age eighteen." - Einstein
“The difficulty lies, not in the new ideas, but in escaping from the old ones.” – John Maynard Keynes
“Any year that you don't destroy one of your best-loved ideas is probably a wasted year.” - Charlie Munger.
If you only take away one thing
You have nothing to fear from the truth (how reality actually is), but that doesn’t mean the truth won’t hurt!
I find the sooner you sooner you have an accurate reflection of how good your idea is the less the hurt… and the more the joy!