If you have spent some time on rationalist forums, you might have come across images that try to visualize cognitive biases that humans are prone to:

This specific one has been created by John Manoogian III and Buster Benson, who compiled the list of biases from Wikipedia.

It is a great way to get a sense of the sheer number of biases that exist, but it doesn’t tell you much about how much of the popular mindshare each bias has. All the biases having the same size implies that they are all equally important, but that is obviously not the case. Arguably, for someone who has just started to learn about cognitive biases, confirmation bias should be more important than, say, the Peltzmann effect.

To measure and visualize the popularity of each bias, I…

• ran a Google search with the format "<insert cognitive bias here>" cognitive bias using a SERP API,
• got the number of search results for each term,
• created a wordcloud using the wordcloud Python package,
• used logarithms of the search count for better scaling,
• used the same colors as the Cognitive Bias Codex for consistency,
• used a shape mask of a brain to make it look cool.

Here is the result:

The bigger the font, the more Google search results there are for that bias, the assumption being Google search results are a good measure of popularity.

Why should you care about the popularity of biases? The more popular or common a bias is, the more likely you are to be affected by it. So it makes sense to study them in decreasing order of popularity, to maximize the benefit to your own thinking. However, this is all statistics—you could still be impacted more by a bias that is smaller in the wordcloud. For example, there was a time when I was very prone to the sunk cost fallacy, even though it doesn’t show up so large in the wordcloud.

Below is a version of the image without the shape mask:

Below are the top 10 biases ranked by Google search result count:

Cognitive bias	Search result count
Prejudice	8,560,000
Anchoring	1,100,000
Stereotyping	1,080,000
Confirmation bias	992,000
Conservatism	610,000
Essentialism	436,000
Loss aversion	426,000
Attentional bias	374,000
Curse of knowledge	373,000
Social desirability bias	319,000

Click here to see the search result counts for each 188 biases included above.

I have also computed the average search result count for each category of biases, by dividing the total search result count for each category by the number of biases in that category:

Category	Average count
We discard specifics to form generalities	1,494,378
We notice when something has changed	237,141
We fill in characteristics from stereotypes, generalities, and prior histories	160,170
We are drawn to details that confirm our own existing beliefs	93,350
We think we know what other people are thinking	81,555
To act, we must be confident we can make an impact and feel what we do is important	72,435
We notice things already primed in memory or repeated often	70,835
To get things done, we tend to complete things we’ve invested time and energy in	65,822
To avoid mistakes, we aim to preserve autonomy and group status, and avoid irreversible decisions	65,750
We edit and reinforce some memories after the fact	59,503
We favor simple-looking options and complete information over complex, ambiguous options	52,491
We tend to find stories and patterns even when looking at sparse data	46,375
To stay focused, we favor the immediate, relatable thing in front of us	37,940
Bizarre, funny, visually striking, or anthropomorphic things stick out more than non-bizarre/unfunny things	37,081
We imagine things and people we’re familiar with or fond of as better	34,379
We simplify probabilities and numbers to make them easier to think about	33,881
We notice flaws in others more easily than we notice flaws in ourselves	31,390
We project our current mindset and assumptions onto the past and future	29,418
We reduce events and lists to their key elements	27,638
We store memories differently based on how they were experienced	20,440

Notice that the top few biases such as prejudice and anchoring highly skew the ranking.

Similarly, I have computed the average search result count for each top category of biases:

Top category	Average count
What Should We Remember?	316,297
Too Much Information	101,842
Need To Act Fast	64,568
Not Enough Meaning	64,134

You can see the code I used to create the figure here.

I will not try to reason as to why some biases are more popular than others, and instead leave that for another post.