Week 14 · Chapter

Chapter 14 — Bias and Decision-Making

Critical Thinking Free sample edition

The module's primary reading · ~40–50 minutes · Objective 8 (with Objective 2 carried forward) · Read it first this week; the lecture, the slides, the tutorial, the quiz, and the Argument Workshop all build on it.

In plain language

This week turns the course's tests on your own thinking. A cognitive bias (a steady tilt in how you gather or weigh evidence) can steer a careful person without their noticing. The best-known one is confirmation bias, which means looking for and favoring evidence that fits what you already believe. You will also meet mental shortcuts that often work and sometimes mislead, such as letting a starting price shape what feels fair. Then you will practice four habits that help you catch these tilts in your own work, including your work on the library question.

Why you already care about this

Think about the last time you looked something up to win an argument. You typed in your side, read the first 5 results, and felt better. Now think about a jacket on a store rack with a tag that reads "was 200 dollars, now 120" (illustrative). The 120 feels like a deal before you have asked what the jacket is worth. Then there is the group project your team swore would take 6 days, and that took 9 (illustrative).

None of these moments involves a bad argument from someone else. Each one is your own judgment, tilted. For most of this course you have tested other people's arguments: their form, their words, their samples, their numbers, and their sources. This week you run the tests on yourself, and you find that the person hardest to check is the one doing the checking.

By the end of this chapter you will be able to:

explain confirmation bias using Wason's rule-discovery task, and say why most people miss the rule;
choose a disconfirming test that could show a guess wrong;
run an evidence audit on your own work on the running question and put its strongest objection in standard form;
spot motivated reasoning with the same-standard test, and check a confident plan against a track record;
name the availability heuristic and anchoring, and say when each shortcut misleads;
apply the four debiasing habits and run the whole bias check on a decision of your own.

Carry this line with you: a test that could fail teaches you something; one that cannot, does not.

Where we are

Already covered: Week 2 taught standard form, the principle of charity, and the argument map. Week 9 taught you to judge a sample, Week 11 to read a base rate, Week 12 to check a source, and Week 13 to ask what would test a claim.
This week: you test your own reasoning for confirmation bias, motivated reasoning, overconfidence, availability, and anchoring, and you practice four habits that push back.
How they connect: Week 13's testable claim was one that some observation could show wrong. This week asks whether you go looking for that observation, or only for the ones that agree with you.
Where we're going next: Week 15 reasons about values and brings the whole course together, and the habit of considering the opposite becomes the method of stating the other side at its strongest.

Two earlier ideas carry this week. From Week 13: a prediction that comes true gives a hypothesis support but never a guarantee, while a prediction that fails counts against it. From Week 2: putting an argument in standard form lists each premise on its own numbered line, and the principle of charity reads the argument as reasonably as its words allow. Both return when you turn the tests on yourself.

1. Catch confirmation bias in the rule-discovery task

Imagine a friend is sure the campus coffee cart is overpriced. Every time they see a high price on the menu, they point it out. Every time they see a fair one, they say nothing. After a month they have a long list of proof, and all of it fits what they believed on the first day.

A cognitive bias is a steady tilt in how people gather or weigh evidence that pulls their judgments the same wrong way, such as favoring what they already believe.

Confirmation bias is the habit of seeking and favoring evidence that fits what one already believes, while overlooking or discounting evidence that could show it wrong.

The classic experiment comes from Wason (1960), "On the failure to eliminate hypotheses in a conceptual task", Quarterly Journal of Experimental Psychology. It is the classic experiment later read as showing confirmation bias. It is not the paper that coined the phrase.

A rule-discovery task is Wason's 1960 experiment in which participants were told that the triple 2, 4, 6 fits a rule and proposed triples of their own to find it; the rule was any ascending sequence.

Worked example — Wason's task (real attribution, taught in words). The experiment ran in three parts.

The setup. Wason (1960) told participants that the triple 2, 4, 6 fits a rule and asked them to find the rule by proposing triples. Each time, they heard whether their triple fit.
The rule. The rule was any ascending sequence: each number larger than the one before it.
What people did. Most participants proposed triples that fit their own guess rather than triples that could disprove it.

In words: participants chose tests that agreed with their guess, so the tests kept saying yes, and many announced a rule that was too narrow. That is the tilt this week names.

Misconception check — confirmation bias is other people's problem. Students often think confirmation bias is something only careless or stubborn people have. That is tempting, because it is easy to see in a friend's coffee-cart list. But Wason's participants were not careless; they were trying to find the rule. The cure is to assume the tilt is in you too, and to check for it on purpose.

Memory hook: agreeable evidence is the easiest evidence to find.

2. Run the disconfirming test

Suppose you think your phone dies early only on days you use the map app. If you check the battery only on map days, every check will agree with you. You will never learn whether it also dies early on days you do not open the map.

A confirming test is a test chosen because it should come out in a guess's favor, so that a pass adds little: a triple that fits a guess about a rule often fits other rules too.

A disconfirming test is a test that could show a guess or belief to be wrong if it is wrong, such as a triple that breaks the guessed rule.

Worked example — a player's run (illustrative). Follow one player through the task.

The guess. A player sees 2, 4, 6 and guesses "add 2 each time".
Three confirming tests. The player tests 3 triples that fit the guess: 8, 10, 12; 20, 22, 24; and 1, 3, 5. All 3 get a "yes", and the player announces "add 2".
Why they could not fail. Every triple that fits "add 2" also fits "any ascending sequence", so none of the 3 tests could have shown the guess wrong.
The disconfirming test. The player tries 1, 5, 20, a triple that breaks the guess. It gets a "yes", so "add 2" is wrong.
One more. Then 6, 4, 2 gets a "no", which points toward the real rule.

In words: 1 test that could fail taught more than 3 that could not. The disconfirming triple did in one try what the confirming triples could never do.

This is Week 13's logic of a test, turned on your own guess. A prediction that comes true supports a guess without proving it. A prediction that fails counts against it. So the most useful test is the one aimed where your guess would break.

Misconception check — more agreeing tests means more proof. Students often think that 10 tests that agree are 10 times better than 1. That is tempting, because each "yes" feels like a point in your favor. But if every one of those tests would also pass under a rival rule, none of them tells the rules apart. One test that could fail is worth more.

Memory hook: aim the test where your guess would break.

3. Audit your own evidence on the running question

Picture yourself writing a short paper on the running question: should the campus library stay open around the clock during finals week? You already lean yes. You open a search engine. What do you type?

An evidence audit is listing the evidence one gathered on a question, with the searches that found it, and checking which side it favors and whether the search itself produced the lean.

Worked example — the evidence list (illustrative, running question). Here is one student's first search.

The first search. A student who favors overnight hours lists the 10 sources found with 3 search phrases, each asking about the benefits of late-night study.
The count. Of the 10 sources, 8 favor overnight hours and 2 support the objection.
The audit. The 3 search phrases all pointed one way, so the lean of 8 to 2 tells you more about the search than about the question.

In words: a search that asks only for one side's reasons finds mostly that side's reasons, so the student cannot yet tell whether the evidence leans or only the search does.

The first repair is to consider the opposite, the habit of stating the strongest case against one's own view, fairly and in full, before deciding how much one's own case is worth. You already have a tool for stating a case fairly: standard form, read with the principle of charity.

Worked example — consider the opposite (illustrative, running question). The student writes the strongest objection in standard form under the principle of charity.

O1: Overnight hours encourage all-night studying.
O2: Losing sleep can hurt exam performance.
So,
OC: Keeping the library open overnight would help students harm themselves.

On the map, the objection is an arrow marked "against" pointing at C, the running question's conclusion. Described in words: O1 and O2 work together to support OC, and OC is an arrow marked "against" aimed at C.

The second repair is a search that could go against you.

Worked example — the disconfirming search (illustrative, running question). The same student searches again.

The new search. The student runs 2 more search phrases that ask about the harms of all-night studying and finds 10 more sources.
The count. Of those 10, 3 favor overnight hours and 7 support the objection.
The full list. The full list is 10 + 10 = 20 sources, with 8 + 3 = 11 favoring overnight hours and 2 + 7 = 9 supporting the objection.
The next step. A count of sources is not a verdict, so the student weighs each one with the source check.

In words: the list went from 8 to 2 to 11 to 9 because the search changed, not because the world did. Counting sources is a check on your search; weighing sources is a check on the question.

Misconception check — the side with more sources wins. Students often think the side with 11 sources beats the side with 9. That is tempting, because a bigger pile looks like better support. But one strong primary study can outweigh several posts that copy one news story. The count tells you whether your search was fair; the source check tells you what the evidence is worth.

A caveat: the audit does not tell you which side of the running question is right. The course takes no side on it. The audit tells you whether your reading was fair to both.

Memory hook: your search phrases are part of your evidence.

4. Spot motivated reasoning and overconfidence

A fan watches a referee's call against their team in slow motion and finds a dozen reasons it was wrong. The same fan sees an identical call against the rival team and accepts it at full speed. The evidence was the same kind; the standard changed.

Motivated reasoning is evaluating evidence more leniently when it supports a conclusion one wants and more strictly when it does not.

The same-standard test is the question of whether one would accept a piece of evidence if it pointed the other way, used to catch evidence held to a looser or stricter standard because of its conclusion.

Worked example — the two posts (illustrative). Two posts land in the same feed.

The posts. Two posts each rest on one student's story. One says overnight study helped a student's grades, and the other says an all-nighter hurt a student's exam.
The uneven standard. The student who favors overnight hours accepts the first and dismisses the second as "only one story".
The same-standard test. Both are single anecdotes, so the same-standard test gives both the same low weight.

In words: the student's complaint about the second post was correct, and it applied to the first post too. Week 9's warning about the generalization from one vivid anecdote cuts both ways.

Overconfidence is holding a belief or a prediction with more certainty than the evidence supports, as when a plan ignores a track record of running late.

Worked example — the project plan (illustrative). A team sets its schedule.

The plan. A team of 4 says it is sure a group project will take 6 days.
The record. Its last 3 projects took 9, 10, and 8 days.
The average. (9 + 10 + 8) ÷ 3 = 27 ÷ 3 = 9 days.
The gap. 9 − 6 = 3 days longer than the plan, and all 3 of the 3 past projects ran past 6 days.

In words: the team's certainty came from how the plan felt, not from its record. A plan of 9 days, the record's average, is the better starting guess, and the team can then ask what would make this project faster.

Misconception check — knowing about a bias protects you from it. Students often think that once they can name motivated reasoning, they are safe from it. That is tempting, because naming feels like seeing. But the team above could define overconfidence and still plan 6 days. The cure is a habit you run, such as checking the record before you commit, not a fact you know.

Memory hook: ask "would I accept this if it went the other way?"

5. Name the shortcuts: availability and anchoring

You guess how many people in your building own a bike by counting the bikes you can picture in the hallway. Most days that works fine. On the day after a bike-theft story in the news, the count you picture might shift for reasons that have nothing to do with bikes.

A heuristic is a mental shortcut for making a judgment quickly, which often works and sometimes misleads.

Tversky and Kahneman (1974), "Judgment under Uncertainty: Heuristics and Biases", Science, described availability and anchoring, along with a third heuristic, representativeness. This chapter teaches the first two.

The availability heuristic is judging how common or likely something is by how easily examples of it come to mind, so that a vivid or recent case can seem more frequent than it is.

Worked example — the plane crash (illustrative). A traveler changes plans.

The judgment. After a news story about a plane crash, a traveler decides that flying is too dangerous and drives instead.
The shortcut. The crash is vivid and easy to recall, and ease of recall is not frequency.
The better question. How many crashes happen out of how many flights? That is a base rate, and the story does not give it.

In words: the story made one crash easy to picture, and the traveler read "easy to picture" as "likely". Week 11 taught you to ask for the base rate; availability is what makes you forget to ask.

Anchoring is the pull of a starting number on an estimate, which keeps the estimate too close to that number even when the number is arbitrary. An anchor is the starting number that an estimate stays too close to, such as a price marked as the old price or the first figure named in a negotiation.

Worked example — the price tag (illustrative). Return to the jacket from the start of the chapter.

The tag. A jacket's tag reads "was 200 dollars, now 120".
The discount. The discount is 200 − 120 = 80 dollars, and 80 ÷ 200 × 100 = 40 percent off.
The anchor. The 200 is the anchor.
The real question. Whether 120 dollars is a fair price is a separate question, and you answer it by comparing prices elsewhere, not by comparing with the old tag.

In words: the arithmetic of the discount is correct, and it still does not tell you what the jacket is worth.

Worked example — a classroom demonstration (illustrative numbers). A class tries the anchor on itself.

The groups. A class of 40 is split into 2 groups of 20.
The starting numbers. Group A is asked whether a used bike costs more or less than 300 dollars, and Group B whether it costs more or less than 30 dollars.
The estimates. Then every student estimates the price. Group A's estimates average 180 dollars and Group B's average 90 dollars.
The gap. 180 − 90 = 90 dollars, from an arbitrary starting number.

In words: the only difference between the groups was a number nobody claimed was the bike's price, and it moved the average estimate by 90 dollars. The same pull works on an opening salary offer at a job interview: the first figure named becomes the anchor for everything after it.

Misconception check — a heuristic is always wrong. Students often think a heuristic is a mistake by definition. That is tempting, because this chapter's examples all go wrong. But a heuristic is a shortcut that often works: things that are common usually are easy to recall, and an old price often does say something about value. The cure is to know when a shortcut misleads, as when a story is vivid or a number is arbitrary, and slow down then.

A limitation worth saying plainly: these numbers are illustrative, made for this chapter. The course does not report the figures from Tversky and Kahneman's own studies.

Memory hook: vivid is not frequent, and first is not fair.

6. Practice the four debiasing habits

A carpenter measures twice before cutting once. The habit does not make the carpenter smarter; it catches errors the carpenter would otherwise make. Debiasing works the same way.

The debiasing habits are four practices that reduce bias in one's own reasoning: consider the opposite, look for the case that would disprove one's view, write the prediction down before seeing the result, and decide in advance what evidence would change one's mind.

Worked example — the habits applied (illustrative, running question). The student from Section 3 uses all four.

Consider the opposite. The student wrote the objection in standard form, O1 and O2 supporting OC (Section 3).
Look for the disconfirming case. The student ran 2 search phrases aimed at the objection's side (Section 3).
Write the prediction down first. Before reading the 10 new sources, the student wrote a guess about which way they would lean.
Decide what would change your mind. The student wrote: evidence that the library's sleep reminders and quiet rest floor do not reduce all-night studying would give the objection more weight.

In words: the student used all 4 debiasing habits, and each one is something done on paper, not something known. The fourth habit matters most, because it is written before the evidence arrives and so cannot bend to fit it.

Misconception check — debiasing means having no view. Students often think that a fair reasoner must stay neutral on everything. That is tempting, because bias sounds like having a side. But the student above still favors overnight hours. The habits do not remove your view; they make sure your view has met the strongest case against it.

Memory hook: write it down before you look.

7. Run the whole bias check

A pilot runs the same checklist before every flight, even after a thousand of them. The list is short, and it works because it does not depend on the pilot feeling careful that day.

The bias check is the set of questions for auditing one's own reasoning: what was searched for and which side it favored, whether the evidence would be accepted if it pointed the other way, what number or vivid story is steering the judgment, what was predicted beforehand, and what would change one's mind.

The bias check: five questions, the bias each one catches, and this week's example of it
Question to ask yourself	The bias it catches	This week's example
What did I search for, and which side did it favor?	confirmation bias	the evidence list's 8 to 2 lean
Would I accept this evidence if it pointed the other way?	motivated reasoning	the two single-story posts
What number or vivid story is steering me?	anchoring and availability	the price tag and the plane crash
What did I predict before I looked?	overconfidence	the 6-day project plan
What would change my mind?	all of them	the sleep-reminder test

Worked example — the verdict on the week's cases. Run the check on everything this chapter met.

The rule-discovery run. The player's 3 confirming tests could not catch the wrong guess, while 1 disconfirming test did.
The evidence list. The 8 to 2 lean came from the search phrases, and a search for the other side changed the list to 11 and 9.
The rest. The posts, the plane crash, the price tag, the bike estimates, and the project plan each show a bias steering a judgment.

In words: the bias check found a tilt in every case, and none of these checks decides the running question.

Misconception check — finding a bias proves the conclusion wrong. Students often think that if a person reached a view through confirmation bias, the view must be false. That is tempting, because a biased route feels like a bad destination. But a biased search can still land on a true conclusion. Finding the bias tells you to recheck the evidence fairly; it does not tell you which way the recheck will come out.

An honesty note: the bias check works best on your own reasoning and on decisions you can still change. Pointing it at someone else's view, as a way to dismiss them, turns a tool for fairness into an attack on the person.

Memory hook: five questions, asked of yourself first.

Key terms

cognitive bias — A steady tilt in how people gather or weigh evidence that pulls their judgments the same wrong way, such as favoring what they already believe. (first used: 1. Catch confirmation bias in the rule-discovery task)
confirmation bias — The habit of seeking and favoring evidence that fits what one already believes, while overlooking or discounting evidence that could show it wrong. (first used: 1. Catch confirmation bias in the rule-discovery task)
rule-discovery task — Wason's 1960 experiment in which participants were told that the triple 2, 4, 6 fits a rule and proposed triples of their own to find it; the rule was any ascending sequence. (first used: 1. Catch confirmation bias in the rule-discovery task)
confirming test — A test chosen because it should come out in a guess's favor, so that a pass adds little: a triple that fits a guess about a rule often fits other rules too. (first used: 2. Run the disconfirming test)
disconfirming test — A test that could show a guess or belief to be wrong if it is wrong, such as a triple that breaks the guessed rule. (first used: 2. Run the disconfirming test)
evidence audit — Listing the evidence one gathered on a question, with the searches that found it, and checking which side it favors and whether the search itself produced the lean. (first used: 3. Audit your own evidence on the running question)
consider the opposite — The habit of stating the strongest case against one's own view, fairly and in full, before deciding how much one's own case is worth. (first used: 3. Audit your own evidence on the running question)
motivated reasoning — Evaluating evidence more leniently when it supports a conclusion one wants and more strictly when it does not. (first used: 4. Spot motivated reasoning and overconfidence)
same-standard test — The question of whether one would accept a piece of evidence if it pointed the other way, used to catch evidence held to a looser or stricter standard because of its conclusion. (first used: 4. Spot motivated reasoning and overconfidence)
overconfidence — Holding a belief or a prediction with more certainty than the evidence supports, as when a plan ignores a track record of running late. (first used: 4. Spot motivated reasoning and overconfidence)
heuristic — A mental shortcut for making a judgment quickly, which often works and sometimes misleads. (first used: 5. Name the shortcuts: availability and anchoring)
availability heuristic — Judging how common or likely something is by how easily examples of it come to mind, so that a vivid or recent case can seem more frequent than it is. (first used: 5. Name the shortcuts: availability and anchoring)
anchoring — The pull of a starting number on an estimate, which keeps the estimate too close to that number even when the number is arbitrary. (first used: 5. Name the shortcuts: availability and anchoring)
anchor — The starting number that an estimate stays too close to, such as a price marked as the old price or the first figure named in a negotiation. (first used: 5. Name the shortcuts: availability and anchoring)
debiasing habits — Four practices that reduce bias in one's own reasoning: consider the opposite, look for the case that would disprove one's view, write the prediction down before seeing the result, and decide in advance what evidence would change one's mind. (first used: 6. Practice the four debiasing habits)
bias check — The set of questions for auditing one's own reasoning: what was searched for and which side it favored, whether the evidence would be accepted if it pointed the other way, what number or vivid story is steering the judgment, what was predicted beforehand, and what would change one's mind. (first used: 7. Run the whole bias check)

Summary

Confirmation bias is seeking and favoring evidence that fits what you already believe.
Wason (1960): most people tested triples that fit their guess, so they missed the rule.
A disconfirming test could show your guess wrong; it teaches more than many confirming tests.
An evidence audit lists your searches and sources and asks whether the search made the lean.
Consider the opposite by writing the strongest objection in standard form, with charity.
Motivated reasoning holds disliked evidence to a stricter standard; the same-standard test catches it.
Overconfidence is more certainty than the evidence supports; check a plan against the record.
Availability reads "easy to recall" as "common"; anchoring keeps estimates near a starting number.
Tversky and Kahneman (1974) described availability and anchoring.
The four debiasing habits are things you do, not things you know.

Check your understanding

Work these before looking at the answers — they are ungraded, and the answers (with reasoning) are at the bottom of the page.

An office manager suspects the printer jams only on double-sided jobs (illustrative). To check, the manager prints 4 double-sided jobs, and all 4 jam. Is the guess now proven? What single test would tell the manager more?
A concert-ticket resale page shows a seat "was 80 dollars, now 60" (illustrative). Two other resale pages list the same kind of seat at 55 dollars and 65 dollars. Work out the discount in dollars and in percent, then say what the other two prices tell you.
A volunteer group plans 2 hours to set up its food drive (illustrative). Its last 4 setups took 3, 4, 3, and 2 hours. What is the average setup time, how many of the past setups ran past 2 hours, and what does that say about the plan?
A student researching whether a town should add a bike-share program (illustrative) searches "bike-share benefits" and "bike-share success stories" and finds 12 sources, 10 of them in favor. What should the student conclude about the 10 to 2 lean, and what should the student do next?

Answers (worked reasoning — read after attempting)

The guess is not proven. All 4 tests were confirming tests: a printer that jams on every kind of job would also jam on 4 double-sided ones. The disconfirming test is a single-sided job. If it jams too, the guess is wrong; if it prints cleanly, the guess has passed a test that could have failed.
The discount is 80 − 60 = 20 dollars, and 20 ÷ 80 × 100 = 25 percent off. The two other prices average (55 + 65) ÷ 2 = 120 ÷ 2 = 60 dollars, so 60 dollars is the going price, not a bargain. The 80 was the anchor; the comparison with other sellers answers the fair-price question.
The average is (3 + 4 + 3 + 2) ÷ 4 = 12 ÷ 4 = 3 hours. Of the 4 past setups, 3 ran past 2 hours. The 2-hour plan is overconfident; the record says to plan about 3 hours, and to write that prediction down before the drive.
Both search phrases asked only for the program's good side, so the 10 to 2 lean may come from the search rather than from the evidence. The student should consider the opposite by stating the strongest objection fairly, then search for it with phrases such as "bike-share problems", and weigh every source with the source check rather than counting them.

Next up in this module: the lecture, the slide deck, the Readings & Resources page, the lecture tutorial, the practice set, and the quiz. Next week, in Reasoning About Values and the Course Synthesis, the habit of considering the opposite becomes steelmanning, the method of stating the other side at its strongest before you reply.

Built with The Course Maker · samples.thecoursemaker.com