Cognitive bias sorting

Jurij Fedorov
Jan 20 · 15 min read

Let’s get to it. This is a mini-guide on bias sorting ideas for psychologists and laymen.

What is a bias?

The Handbook of Evolutionary Psychology has a great intro on biases overall and a very loose definition of them as they are not well defined overall:

By cognitive bias, we mean cases in which human cognition reliably produces representations that are systematically distorted compared to some aspect of objective reality.

But humans are not evolved to fail in a natural environment. So why do we actually have biases in our head?:

To the evolutionary psychologist, however, the evaluative task is not whether the cognitive feature is accurate or logical, but rather how well it solves a particular problem, and how solving this problem contributed to fitness ancestrally. Viewed in this way, if a cognitive bias positively impacted fitness, it is not a design flaw — it is a design feature.

How can we sort biases?

I will present some categories people have tried to sort biases into. There are obviously many other ways to do it, but they are largely similar to each other in some ways. With biases and emotions there are no clear structures, models or categories because they are brain modules with just the basic function in common. Biases are also unfortunately defined in a wrong way: by how they fail. Imagine if your arms were defined by how they fail to lift a piano. Biases furthermore are also named after how they fail in a modern setting. Naming a bias, IKEA effect, would be like calling your arms, piano failures. Or like calling the emotion of anger, car accident effect. If a person loses IKEA effect, arms or anger then that person will lack a tool to solve certain problems in life. So removing a bias actually causes one to have more faulty thinking overall.

Some biases are more potent than others which makes them kinda top layer biases. Some of them are very similar to each other and some seem to belong to several vastly different categories at one time. So if you want to create categories for biases or emotions you have to look somewhat outside the science of psychology and just go ahead and do your best. Categories are created to simplify matters and explain complicated concepts.

All the bias cards I mention are completely free so go ahead and download them. I already put most of them into Tabletop Simulator so you can play around with them in 3D. If you can’t find something just message me and I will send you the material.


Uncategorized

There are some of the bias cards and posters online that just have biases unsorted or sorted by names.

Heuristicards

https://medium.com/@agkdesign/heuristicards-a-flash-deck-of-cognitive-bias-4dac372ac802

These bias cards are beautiful and simple. Instead of categorizing them into groups he just sorted them by name and used varied color ranges and numbers. It’s the most simple way to do it as it requires no arbitrary categories. But adding or removing a bias will cause all numbers and colors to change. So how would you add the remaining 200 biases to such a list?

Yourbias.is

https://yourbias.is/shop

These short 20 to 50 biases lists are great for most people because a lot of biases are very minor and particular to specific situations. Just be aware that not all biases are actually scientifically shown to exist. And some are just misperceptions about how humans actually judge situations.

The Yourbias.is site also has a list of logical fallacies. As biases are irrational looking thinking it implies that logical fallacies are the ways we are biased.


Wikipedia bias list sorting

Pretty much all bias lists with over 60 biases use the Wikipedia bias list as a starting point. It’s the best bias list we have, but unfortunately, it’s not complete. When I made my own bias list I found 30 extra biases and had to remove some biases on the list that appeared twice.

The current Wikipedia bias list article uses 3 categories for biases:

  1. Decision-making, belief, and behavioral biases
  2. Social biases
  3. Memory errors and biases

This seems to be the most simple way you can possibly sort biases.

Decision-making biases make you able to take effective decisions in life. Social biases are needed to fit into an ingroup setting. Memory biases make us able to remember the most essential problems, solutions, and resources in life.

Cards that autogenerate using a Wikipedia list:

That’s pretty darn impressive! Some people… man…

Wiki cards. Made for easy print:

https://github.com/RichardJActon/CognitiveBiasCards

This one looks so simple yet has so many details in it. It’s a bit old and lacks some of the newly added biases.


Bias Codex

universal point of view

https://commons.wikimedia.org/wiki/File:Cognitive_bias_codex_en.svg

Bias codex is made by sorting the Wikipedia bias list into more detailed categories so that laymen can easily access biases. Since a bias can belong to multiple categories it does make this sorting method confusing at times. It’s weird when you have 3 very similar biases next to each other, but then a 4th one just as similar to them, but in a totally different group.

There are 4 main categories and 20 sub-categories. All really great and to the point. The creator clearly knows his way around psychology which is a breath of fresh air compared to some more arbitrate sorting methods.

https://medium.com/thinking-is-hard/4-conundrums-of-intelligence-2ab78d90740f

The 4 categories are: attention, logic, proactivity, memory. They seem like very natural categories for biases. Of course, a social bias category may also have been useful.

You can find the interactive list here. It has 8 duplicate biases and is also missing crucial biases. So while the categories are super cool the bias codex itself uses an older and shorter Wikipedia bias list.

The maker of the list also created 13 different categories for an internet site/iPhone app:

https://pocket-biases.glideapp.io/

There are also these flashcards:

Here are 3 fancy posters with biases and descriptions:

Cognitive bias codex — can be found in all sizes and in different languages

220 cognitive biases compiled, organized and described in 5 columns

220 cognitive biases compiled, organized and described in 1 column

My own cards

I noticed that there are no complete and easy to get to bias list online which inspired me to make my own cards using the bias codex categories. I’m still updating the text in the cards. I have an Excel file with all the text that is free to use. I also have the nanDeck files so you are welcome to change anything in the cards. You can make your own cards in 15 minutes in nanDeck and save them all with 2 clicks. Obviously I’m no designer…

PDF file with cards

Zip folder with card images


Ux in Lux cards

creative projects point of view

https://stephaniewalter.design/blog/52-ux-cards-to-discover-cognitive-biases/

Categories:

1. Decision-making & behavior

These biases affect people’s decision-making abilities, behaviour and the decisions they make based on the different information they get.

2. Thinking & problem solving

These biases can change the way people think or solve problems and lead them to come up with wrong conclusions.

3. Memories & recalling

These biases can influence choices by either enhancing or impairing the recall of a memory or altering the content of a reported memory.

4. Interview & user testing

These biases can directly influence designer, during interviews or user testing, and may change the outcome of our research. They influence the behaviour of people we interview or people who will test your products and services.

5. Team work, social & meetings

These biases can change the way groups of people work collectively and interact with each other, whether in a meeting room or in their daily lives in general.

As you can see the 3 Wikipedia bias list categories seem to appear here too: problem-solving, social and memory biases. Problem-solving and ego biases have been split up into: behavior, thinking, and user testing.

You can see all their biases online here.

The “Interview & user testing” category is clearly arbitrary and not based on evolutionary psychology. There are maybe 70 of these kind of biases that are described from an artificial point of view. We could have named “Observer-expectancy effect” something less lab focused, but these biases are researched via small experiments so that’s how they appear to scientists. I suspect that once bias lists make their way into the business world some biases will be renamed and described in a more scientific/evolutionary psychology way.


Sitback bias cards

business point of view

https://blog.sitback.com.au/bias-cards-free-download

Here the biases are sorted into project-focused categories:

  1. Planning
  2. Observing
  3. Designing

Then there are 2 other groups on top of that.

  1. Participant
  2. Researcher

It seems like these categories are for business settings. These sorting methods are fine for very specific practical uses, but they don’t really explain what these biases exist for.

You need to pick the right list for your needs whether you are a researcher, manager or school teacher.


Cognitive Bias Cards For Learning Practitioners

teacher point of view

https://www.dashe.com/cognitive-bias-in-learning-cards-free-download

These cards define common cognitive biases that influence the four goals of learning:

  1. Behavior change
  2. Performance improvement
  3. Learning retention/recall
  4. Critical thinking/problem
  5. Solving

These cards seem practical for school teachers.

A similar set of cards exists but is not free, so I can’t really tell you much about these categories.

Mental Notes — Psychology on the Web Card Deck

These non-free cards are sorted into these categories:

1. Comprehension

2. Attention

3. Persuasion

4. Memory

A bias can belong to multiple categories. It seems like the best way to go about this as a bias may for example be a memory bias but also an ego bias.


TitleMax cards

household economy point of view

https://visual.ly/community/Infographics/other/50-common-cognitive-biases
  1. Memory
  2. Social
  3. Learning
  4. Belief
  5. Money
  6. Politics

A card can be in all 6 categories at once. This is pretty useful to find the biases that may influence your decisions in a certain area of modern life. Probably the most simple categorization of all as even kids could place biases into these categories.


A Visual Study Guide to Cognitive Biases

PDF file from 2010. They have 4 categories:

  1. Social biases
  2. Memory biases
  3. Decision-making biases
  4. Probability/belief biases
https://www.academia.edu/36478961/A_Visual_Study_Guide_to_Cognitive_Biases

acaps biases

humanitarian projects point of view

https://www.acaps.org/sites/acaps/files/resources/files/acaps_technical_brief_cognitive_biases_march_2016.pdf
  1. Selection Biases
  2. Social Biases
  3. Process Biases

Selection biases are caused by choosing non-random data for analysis. Some information is unconsciously chosen or disregarded, misleading us into a wrong conclusion.

Social biases are a result of our interactions with other people. The way we are processing and analysing information depends on our relations with the persons who provided us with information or hypotheses

Process biases impact our ability to process information based on evidence. They prevent us from accurately understanding reality even when all the needed data and evidence are in our hand.

These 3 categories have weird names, but seem logical. Social biases is an obvious category. Selection biases are what we focus on in our environment. Process biases is how we understand and solve problems. As humanitarian work is proactive/current they don’t use a memory bias category.

Download the poster here.

Complicated categorizations only for lab use

These 3 next categorizations below are not really useful for laymen. But they may still be interesting to look into and get inspired by. After that, I will go into the actual psychology behind biases.

Bias categories for biomedical scientists

only from biomedical science point of view

https://jech.bmj.com/content/jech/58/8/635.full.pdf

Group of bias:

  1. Selection bias
  2. Information bias
  3. Execution of an intervention

Then there is a really big subgroup of biases. Too many to list here. And there is even another subgroup pointing out the experimental design being affected.

This is too complicated a list for anyone to easily use. Even the bias definitions in the article are hard to understand. But scientists working with big data math may want to look into this. It’s good that they have their own very unique bias list.

A Task-Based Taxonomy of Cognitive Biases for Information Visualization

lab experiements point of view

https://www.researchgate.net/publication/327907209_A_Task-Based_Taxonomy_of_Cognitive_Biases_for_Information_Visualization

This bias list is made from the Wikipedia bias list. Here again we have 2 side-by-side categories. It’s just 2 different ways to sort biases combined into one.

As these groupings completely ignore evolutionary psychology it’s hard to use this sorting for anything besides experiments focused on visual biases. The thinking behind these groups is: biases exist, we don’t try to explain what they are or what they do, but let’s group them in a way that’s useful for our specific field.

And in lab settings or arbitrary settings biases may just look like logical mistakes:

If problems presented in the laboratory are not those for which the human mind is designed, we should not be surprised that people’s responses appear to be systematically irrational.

I also made cards out of this list:

Cards Images

Cards PDF file

The site below is a much more simple presentation of the visual bias concept. And it presents all biases in a way that it becomes clear why they need to be thought about when designing a user interface. A site like Amazon.com uses all the visual bias concepts and it becomes obvious once you study this list below:

https://www.mobilespoon.net/2019/04/collection-cognitive-biases-how-to-use.html

Studying Biases in Visualization Research: Framework and Methods

lab experiements point of view

https://homepage.univie.ac.at/michael.sedlmair/papers/calero-valdez2017framework.pdf

This bias list is yet again a super nerdy experiment-focused list. But here we have an actual categorization of biases that is not just made for lab use but also rooted in bias studies and human behavior.

I tried to sort biases into these categories, but it’s too complicated to apply such terms without any guide. It’s nice that they have overlapping categories as that is exactly what is needed for some biases.

They also have an interactive list with bias codex categories in the paper. And they have a guide for how to sort biases into clear number categories. Very nice! Number categories are very useful because many biases have 3–4 different names. And then they have names in other languages too. For such a list to be useful biases do need either a number system or just very specific categories that cannot be confused.

A Neural Network Framework for Cognitive Bias

universal point of view

This model is a bit vague overall. It’s a framework that makes sense but could be better explained.

  1. Association
  2. Compatibility
  3. Retainment
  4. Focus

Here they try to explain how the brain picks up and works with information generally.

Evolutionary psychology bias sorting

universal point of view

Evolutionary psychology explains biases in how they work and why they exist instead of just explaining how they look like in a lab setting. Imagine a slow and clumsy robot trying to solve the problem of how hard it would be to climb a certain hill. It may first calculate how tall a hill is, then calculate how steep it is, and then calculate how hard it would be to climb in every single location on it. Then after that, it may conclude that its wheels are way too small to climb the hill so all calculations were wasteful. In this example, the robot does a lot of detailed calculations that are a waste of brainpower. Knowing the exact measurements of the hill is just overkill and the resources and time spent on calculating this could be used for resource gathering or reproduction. The robot just needs to know how hard it is to climb the hill, nothing else. The answer “Looks freaking steep from here. Too dangerous to climb.” is more productive than an answer with 1000 different measurements all fighting for attention. The example below shows why humans are so clever at being biased. We don’t need to perceive the reality of the situation. We need to understand the world in how we can use it optimally. So we perceive a hill steeper if we are on top of it as that is what is useful for us to understand. And our mood, energy level, and physical condition also change this perception:

In one series of studies, people consistently overestimated the steepness of hills — both real and computersimulated (Proffitt, Bhalla, Gossweiler, & Midgett, 1995). Failing to properly descend a steep hill is far more costly than failing to properly ascend it. An error management perspective therefore predicts that this bias toward overestimating slopes will be greater when people view from the top than from the bottom, which was exactly what is found (Proffitt et al., 1995).

https://onlinelibrary.wiley.com/doi/pdf/10.1002/9781119125563.evpsych241

There are different kinds of biases as you see in the image above. They left out heuristics that are just defined as calculation biases (logic shortcuts) that make us take a wrong direction in specific calculations. Evolutionary psychology redefines the concept of biases into artifacts and error management biases.

Artifacts are research strategies that may be great for solving natural problems but fail in unnatural settings. Wason selection task illustrates this as college students don’t seem able to solve the logical math puzzle but can solve the same problem very easily if it’s presented as a social problem with humans.

An error management bias would be when you, for example, have more to lose than to gain from going into a dark forest, so your brain will overestimate the danger of the small forest. Or the steepness of a hill. It’s an incorrect conclusion that overall is helpful for survival and reproduction long-term.

In the image below you can see how error management biases can be sorted into 3 broad categories:

  1. Judgment of threat
  2. Evaluation of interpersonal relationships
  3. Evaluation of the self
https://onlinelibrary.wiley.com/doi/pdf/10.1002/9781119125563.evpsych241

Here the idea of resource cost is introduced. Lack of false positives or false negatives biases may lead to deadly situations or make us miss out on great resources. So it’s better to be safe than sorry. Better be mistaken 100 times about there being a lion in the bush than being eaten one time.

I furthermore suggest that such a model can be useful to sort biases into potency. False negatives and false positives can be illustrated with numbers and our natural environment can be compared to our modern environment in how biases fail more in a specific unnatural setting. This is why it’s useful to build bias categories for single settings like lab, business or project. Because the biases may have different potencies based on the arbitrary setting they are in. The various potencies could go from 1 to 10 and decide the size and location of the bias on a 2D graph.

So in error management biases, we have these 3 categories: Danger biases, social biases, ego biases. Then we have artifacts biases that don’t just over- or underreact, but rather react in a wrong area altogether.

Conclusion

Despite widespread claims to the contrary, the human mind is not worse than rational . . . but may often be better than rational. — Cosmides and Tooby

Only one bias model was scientific and properly explained, evolutionary psychology itself. All the other bias sorting methods I have looked into are theoretically vague and unclear. They are more practical than name sorting, but they also all are so unspecific that I cannot really add biases to the lists without feeling like I’m misplacing some biases.

Bias codex is currently the most logical detailed bias sorting tool as it’s based on actual psychology theories and not just the practical application of biases themselves. This makes it useful in all settings: lab, school, business. But it does have the problem of a single bias fitting into multiple categories at once and a dual-sorting method could fix that.

So, how does one sort the 300+ biases out there? It depends, is it for a specific or general use? General sorting must use evolutionary psychology in some way. Error management theory sorts biases in what problems they are created to solve. So error management biases guide us in a specific direction in life. While artifacts are misperceptions like seeing a face on the surface of Mars because we are born to spot human faces easily in our environment. Then in the second layer, one could add a memory category. So, for example, ego biases will furthermore have an ego memory bias category. The duplicate and semi-duplicate biases should be in sub-groups. And the potency of biases should be taken into account too so that strong biases could be visually made to appear bigger if one used a 2D graph to show them or some list chart.

Anyhow, I just like colorful things and lists. Contact me if you have any questions. I’m bored and want something to do.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store