Fallacies and biases
A name is not an argument. Each entry has a definition, three illustrations of the same example, a way to answer it, and a practice question that also sits in the test.
Cognitive Biases List with Examples
Fallacies
Ad hominem
An ad hominem answers a claim by attacking the person who made it, instead of the reasons. A fact about the speaker is relevant only when it actually bears on the evidence, such as a shown conflict of interest or a record of false reports on this subject.
Straw man
A straw man replaces someone’s actual view with a worse or simpler one, then attacks the substitute. The real view is left standing. Quoting loosely, dropping a qualification, or leaping to an extreme all do this.
False dilemma
A false dilemma pretends there are only two options when others are available, or treats the options as if you must take one extreme. Sometimes the pair is real. The fallacy is forcing a pair that has not been shown to be complete.
Slippery slope
A slippery slope says a modest step will lead, without a shown mechanism, to an extreme result. Slopes can be real when each step is likely. The fallacy is the bare claim that one concession dissolves the whole line.
Appeal to authority
Citing an expert is reasonable when the person is expert on this question, the field has real standards, and serious disagreement is not hidden. The appeal fails when the authority is in the wrong field, is unnamed, or is treated as settling a question experts still dispute.
Appeal to popularity
An appeal to popularity treats the fact that many people buy, believe, or repeat something as proof that it is true or best. Popularity can be a clue about taste, price, or what is easy to find. It is not, by itself, a test of quality or truth.
Appeal to nature
An appeal to nature says something is good because it is natural, or bad because it is artificial. “Natural” is also vague: lightning, infection, and medicine are all parts of the world. The origin of a thing does not settle whether it helps or harms.
Appeal to emotion
Feeling can explain why a question matters. It becomes a fallacy when pity, fear, anger, or pride is offered instead of a reason that the claim is true. A sad case can be real and still not answer the question that was asked.
Appeal to ignorance
An appeal to ignorance says a claim is true because it has not been proved false, or false because it has not been proved true. Absence of proof is sometimes a reason to withhold belief. It is not automatically a reason to believe the opposite.
Circular reasoning
Circular reasoning assumes what it is trying to prove. The conclusion reappears as a premise, sometimes in different words. A reason has to be independent of the claim it is meant to support.
Post hoc
Post hoc ergo propter hoc treats “after this” as “because of this.” Order in time is necessary for a cause of that kind, but it is not sufficient. Other changes may explain the result.
Hasty generalisation
A hasty generalisation draws a wide conclusion from a sample that is too small, too narrow, or too casually gathered. One vivid case can be true and still be a bad basis for “everyone” or “always.”
Red herring
A red herring introduces a point that may be true, or even interesting, but does not bear on the question being decided. The discussion slides, and the original claim is never answered.
Tu quoque
Tu quoque answers a criticism by saying the critic does the same thing. Hypocrisy may hurt someone’s standing. It does not show that the criticism is false. A bad example can still state a true rule.
Equivocation
Equivocation slides between two meanings of a word or phrase in the middle of an argument. The conclusion looks supported only if you pretend the word never moved.
Composition
The fallacy of composition says that what is true of the parts must be true of the whole. Sometimes that holds, as with weight. Often it does not: a team of strong players can still play badly together.
Division
The fallacy of division says that what is true of the whole must be true of each part. A heavy orchestra is not made of heavy notes. A rich club need not have rich players.
No true Scotsman
No true Scotsman saves a generalisation from a counterexample by quietly changing what counts as a real case. “No Scot would do that” becomes “no true Scot” once someone does. The claim is made immune to evidence.
Moving the goalposts
Moving the goalposts changes the standard of proof after it has been met, so the claim can never be accepted. Asking for better evidence at the start is fair. Changing the demand only when the evidence arrives is not.
Shifting the burden of proof
The person who asserts a claim carries the burden of giving a reason. Shifting the burden tells the sceptic to disprove it instead. Scepticism is not itself a positive theory that needs the same kind of proof.
Appeal to tradition
An appeal to tradition says a practice is right, or a claim is true, because it is old or customary. Age can explain why a practice survived. Survival is not a proof that it is fair, safe, or accurate.
Genetic fallacy
The genetic fallacy judges a claim by its origin: who first said it, or the shady way it entered the conversation. Origin can be a reason to look harder. It is not, by itself, a reason the claim is false or true.
False analogy
An analogy argues that because two things are alike in some way, they are alike in the way that matters to the conclusion. The analogy fails when the similarity is real but irrelevant, or when a relevant difference is ignored.
Middle ground
The middle-ground fallacy says the truth is halfway between two positions simply because they conflict. Compromise is sometimes wise about what to do. It is not a method for finding a fact. One side can be much closer than the other.
Special pleading
Special pleading applies a rule to others and exempts your own case without a relevant difference. An exception can be fair if the difference is stated and it matters. “Because it is mine” is not a difference that matters.
Loaded question
A loaded question hides a disputed assumption, so that either a yes or a no accepts it. “Have you stopped wasting the budget?” assumes the waste. The fair move is to answer the assumption first.
Anecdotal fallacy
An anecdote is a story about a case. It becomes a fallacy when the story is used to dismiss a broader pattern, or to stand in for one. A story can show that something is possible. It rarely shows how common it is.
Cognitive biases
Confirmation bias
Confirmation bias is the tendency to search for, remember, and accept what supports a view you already like, and to skim past what would hurt it. It is not a claim that every belief is stubborn. It is a pattern in how attention is spent.
Availability bias
Availability bias judges how common something is by how easily examples come to mind. Vivid, recent, or personal cases feel frequent. A dull record of ordinary cases is harder to recall and gets underweighted.
Anchoring
Anchoring is the pull of the first number or fact you meet. Later figures are judged against it, even when the first figure was arbitrary or a deliberate high price. A relevant comparison breaks the pull.
Sunk cost
Sunk-cost thinking lets a past loss decide a future choice. Money or time already spent is gone whether you continue or stop. The live question is which option is better from now on.
Survivorship bias
Survivorship bias looks at the cases you can still see and forgets the ones that disappeared. The successes are interviewed. The attempts that failed are quiet. A rate needs both.
Framing
Framing changes a decision by changing the wording of the same facts. “90% survival” and “10% mortality” describe one risk. People often treat them as different risks because one wording highlights the gain and the other the loss.
Halo effect
The halo effect lets one striking trait colour judgements that the trait does not speak to. Warmth, beauty, confidence, or a famous employer spills onto unrelated questions such as accuracy or honesty.
Gambler’s fallacy
The gambler’s fallacy treats independent events as if they owed you a correction. After a run of one result, the other result feels “due.” If the events are independent, the run does not change the next chance.
Overconfidence
Overconfidence is certainty that outruns the evidence, common when someone has just met a subject and cannot yet see its difficulties. Specialists often hedge because they can see the ways a claim fails. High confidence is not high competence.
Selection bias
Selection bias distorts a result because the people or cases you reached are not a fair stand-in for the group you want to talk about. The flaw is in who got into the sample, not only in how you do the arithmetic afterwards.
Hindsight bias
Hindsight bias makes a past event feel as if it was obviously going to happen, once you know that it did. The feeling of “I knew it” arrives after the result and rewrites how uncertain the situation was.
Fundamental attribution error
The fundamental attribution error explains other people’s behaviour by their character, and underweights the situation. Your own similar behaviour is easier to explain by the circumstances. The bias is the uneven use of those two explanations.
Planning fallacy
The planning fallacy is underestimating the time, cost, or difficulty of a task, especially one you are inside. You picture the smooth path and leave out delays that, on similar tasks, usually appear. Hope is not a base rate.
Base-rate neglect
Base-rate neglect grabs a vivid likelihood and ignores how common the thing was to begin with. A test can be quite accurate and still be wrong most of the time it raises an alarm, if the condition is rare.
Status-quo bias
Status-quo bias prefers the current arrangement because it is current, and treats a change as a loss even when the change is the better expected option. There can be real costs to changing. The bias is counting “it is already so” as a reason beyond those costs.
Often confused
Straw man vs red herring
One replaces the claim. The other walks away from it.
Ad hominem vs tu quoque
Both aim at the person. Only one of them says “you do it too.”
Slippery slope vs false dilemma
One invents a slide. The other hides the middle.
Appeal to authority vs appeal to popularity
A famous name and a large crowd are different substitutes for a reason.
Confirmation bias vs availability
One keeps the evidence you already wanted. The other trusts whatever comes to mind first.
Anchoring vs framing
A first number pulls the estimate. A wording changes the choice.
Post hoc vs correlation
Sequence is not a cause. Moving together is not a cause either. This library files the first mistake as post hoc. It does not file a separate correlation card.
Sunk cost vs status quo
One clings to what you already spent. The other clings to what you already have.