Cognitive bias test

How to take the test

1. Press Start the test. You get 10 scenarios, picked at random from the 14 listed below.
2. For each one, pick the answer that fits: usually the name of the bias at work, sometimes the decision or effect it leads to. The answer, an explanation and the lesson’s source appear straight away.
3. After the tenth scenario, part 2 has two short puzzles. Answer with your first thought. If you’ve met a puzzle before, tick the box under it and it won’t count.
4. Press Check my answers, or Skip to my results if you’d rather not do the puzzles.
5. Your results list the biases you spotted and the ones you missed, each linked to the lessons that explain them.

Take it again and the scenarios you haven’t seen yet come first. There’s no timer.

How it’s scored

Part 1 gives one point per scenario, so the score runs from 0 to 10. Most scenarios ask you to name the bias. Three ask what the sensible decision is, or what a price tactic is for (the sunk cost, contrast effect and decoy effect scenarios); a right answer there still counts as spotting that bias.

Say you get 8 of 10 and miss the scenarios on the availability heuristic and hindsight bias. Your score reads 8/10, both biases appear under “Biases you missed”, and each links to the cognitive biases lessons.

Part 2 is scored on its own, out of the puzzles you hadn’t seen before. Tick “seen it before” on both and part 2 shows “Not scored”. The answer options in part 1 are shuffled each time, so the right answer moves around.

What cognitive biases are

A cognitive bias is a predictable way judgment drifts from what the evidence supports. Many come from mental shortcuts that work well enough most of the time. Judging how common something is by how easily examples come to mind gives a quick answer about everyday things, and a skewed one after a plane crash has been on the news all week. So a bias describes when a shortcut tends to go wrong; the shortcut itself is often useful.

The 14 biases in this test

Each scenario is the quiz question from a lesson in Scroll’s Psychology category, shown as written in the app. The sources are the ones the lessons name.

Bias	What it looks like	Lesson’s source
Confirmation bias	Seeking and remembering evidence that fits what you already believe	Wason (1960)
Halo effect	One good trait, such as looks, colors judgments of unrelated ones	Thorndike (1920)
Sunk cost fallacy	Sticking with a plan because of what you’ve already put into it	Arkes and Blumer (1985)
Availability heuristic	Judging how likely something is by how easily examples come to mind	Tversky and Kahneman (1973)
Bandwagon effect	Following what other people do, especially when unsure	Leibenstein (1950)
Fundamental attribution error	Putting other people’s behavior down to character and overlooking their situation	Ross (1977)
Self-serving bias	Taking credit for successes and blaming failures on circumstances	Miller and Ross (1975)
Status quo bias	Keeping whatever option is the default	Samuelson and Zeckhauser (1988)
Contrast effect	Judging something against whatever you saw just before it	Kenrick and Gutierres (1980)
Decoy effect	A third option, not meant to be chosen, that makes another look like good value	Huber and colleagues (1982)
Hindsight bias	Feeling an outcome was predictable once you know it	Fischhoff (1975)
Planning fallacy	Underestimating time and cost, even after past overruns	Buehler and colleagues (1994)
Ostrich effect	Avoiding information you expect to be bad	Karlsson and colleagues (2009)
Choice-supportive bias	Remembering what you chose as better than it was	Mather and Johnson (2000)

Knowing a bias and avoiding it

Part 1 checks recognition: can you name the bias in a short story about someone else? That’s the easier half. Part 2 tries the harder one, whether you make the error yourself when nothing warns you it’s coming.

The first puzzle is a conjunction problem. A description of Amara makes her sound like an environmental volunteer, and you’re asked whether “works in a bank” or “works in a bank and volunteers with a local climate group” is more likely. Everyone in the second group is also in the first, so the second statement can never be more likely, however well it fits the story. Amos Tversky and Daniel Kahneman named this error the conjunction fallacy in 1983. Their best-known version describes a woman called Linda, and because so many people have seen it, we wrote a new story for this page.

The second puzzle comes from Shane Frederick’s Cognitive Reflection Test (2005): a bat and a ball cost $1.10 in total, and the bat costs $1.00 more than the ball. The answer that comes to mind first is 10 cents. But then the bat would cost $1.10 and the pair $1.20. Call the ball’s price x: the bat is x plus $1.00, so 2x plus $1.00 makes $1.10, and the ball costs 5 cents.

Famous findings that didn’t hold up

Psychology has its own record of confident claims that later work cut down to size, which is one reason we list the source under every scenario.

• Ego depletion, the idea that self-control runs down like a battery, was tested in 23 labs with 2,141 people in a preregistered replication published in 2016. The effect came out at d = 0.04, with a 95% confidence interval from minus 0.07 to 0.15, which includes zero.
• In a 2018 replication of the marshmallow test by Watts and colleagues, the link between a child’s ability to wait for a treat and later school achievement was half the size reported in the original studies, and shrank by two thirds once family background was taken into account.
• Le Texier’s 2019 study of the Stanford Prison Experiment’s archives found that the guards had been given precise instructions on how to treat the prisoners.
• A famous story says 38 neighbors watched Kitty Genovese’s murder and did nothing. Manning and colleagues found no evidence in 2007 that 38 witnesses were present, that witnesses saw the murder, or that they stayed inactive.

Scroll’s app has lessons on all four. We’ve kept them out of this test and the daily quiz until they’re corrected in the app.

What your result doesn’t tell you

Scenarios are short and written to have one best answer, while real decisions rarely come labeled. A few of the answer options are close cousins (hindsight bias and confirmation bias, for instance), so a wrong pick can mean you knew the idea but not the name. If a scenario looks wrong to you, use the “Report a problem with this question” link under it.