1 Critical Thinking IIa

Critical Thinking — Intermediate

Introduction

In the Bare Essentials level, you built a real foundation. You learned to separate the objective from the subjective (SOS), to recognize the structure of an argument, to spot common logical fallacies and cognitive biases, and to start applying those skills in your daily life. That’s not nothing — most people go through life without any of those tools.

But the Bare Essentials toolkit is a starting point, not a finish line. The world you’re navigating is saturated with confident-sounding claims that don’t hold up to scrutiny, emotionally compelling stories that feel true but aren’t, and systems — sometimes deliberately designed, sometimes not — that can bypass your reasoning and influence your behaviour. The basic toolkit helps. But modern life keeps raising the bar.

The good news is that critical thinking is a skill, and skills develop with practice and deeper understanding. This Intermediate level is about sharpening the tools you already have, adding new ones, and building the habits that make them feel natural rather than effortful.

What You’ll Cover

Evidence, Probability, and Trust — How to evaluate claims and sources, how much evidence different claims actually require, how to think in probabilities rather than certainties, and why confident-sounding people are not necessarily reliable ones
SOS: The Subjective Trap — A deeper look at how subjective narratives — romantic ideas, cultural sayings, emotionally resonant stories — can quietly slip into objective decision-making if you’re not watching for them
Media Literacy — Applying critical thinking to the information environments you live in every day: news, social media, advertising, and more
Intellectual Virtues — The character side of critical thinking: intellectual honesty, courage, and humility as learnable skills, not fixed personality traits
More Tools — An expanded look at logical fallacies and cognitive biases, with resources to keep building beyond this program
Scam and Manipulation Defense — How scams and cons work, how they bypass your reasoning, and how to defend against them

Why This Matters for What’s Ahead

In the horse, carriage, and driver framework (see Level 1, Topic 6: Overcoming Barriers), critical thinking is the core driver skill — it’s what navigates. The tools in this Intermediate level are useful immediately, at every level of this program and in everyday life: evaluating a news story, deciding whether to trust a source, recognising when an emotionally compelling argument is leading you somewhere you wouldn’t otherwise choose to go, or simply being honest with yourself about whether your beliefs are based on evidence.

They also connect deeply to other Level 2 topics. Understanding how personal certainty can mislead you links directly to Psychology and Emotion Management. Intellectual honesty and humility connect to Communication Skills. Media literacy touches Technology & Society. These aren’t isolated skills — they reinforce each other.

And when you reach Level 3 — systems thinking, social change, institutional change — they become especially critical. Those topics involve complex cause-and-effect relationships, competing claims from parties with interests at stake, and decisions that affect many people over long time periods. Weak reasoning does its most damage there. But you don’t have to be headed for Level 3 for any of this to be worth your time.

Evidence, Probability, and Trust

Why Certainty Is the Wrong Goal

Most of us were never taught to think in probabilities. We learned to sort things into true or false, right or wrong, fact or opinion. That’s a reasonable starting point, but it breaks down quickly when you try to apply it to the messy, complicated claims that make up real life.

Consider: how certain are you that your memory of an argument you had last year is accurate? That the news story you just read is complete and unbiased? That the supplement your friend swears by actually does what they say it does? In each case, demanding a simple yes or no forces you into a false position. The honest answer in most of these situations is somewhere in between — and the more precisely you can locate where in between, the better your decisions will be.

The goal of critical thinking is not certainty. It’s calibration — having beliefs that accurately reflect the actual evidence available to you.

Thinking in Probabilities

A probability mindset means asking “how likely is this, and based on what?” instead of “is this true or false?” It’s the difference between:

“That’s not proven, so I don’t believe it” — which treats absence of certainty as evidence of falsehood
“The evidence so far suggests this is probably true / probably false / genuinely unclear” — which reflects the actual state of knowledge

This is not relativism. It doesn’t mean all claims are equally valid, or that you can never be confident in anything. Some things are so thoroughly supported by evidence that treating them as uncertain would itself be a distortion — the Earth is approximately 4.5 billion years old, vaccines do not cause autism, evolution happened. Other things are genuinely uncertain, and pretending otherwise is its own kind of error.

The skill is matching your level of confidence to the actual strength of the evidence — no more, and no less.

Why This Feels Uncomfortable

Humans are not naturally comfortable with uncertainty. We tend to prefer a confident wrong answer over an honest “I’m not sure” — partly because uncertainty feels like weakness, and partly because our minds are wired to reach conclusions quickly. This is part of why the cognitive biases covered in the Bare Essentials level exist: they are often mental shortcuts that get us to the feeling of certainty faster, even at the cost of accuracy.

Sitting with uncertainty is a learnable skill, and a genuinely useful one. As you’ll see in the Intellectual Virtues section later in this level, the ability to say “I don’t know yet” is one of the markers of a strong thinker, not a weak one.

Proportional Evidence

Not all claims need the same amount of evidence. This sounds obvious when you say it out loud, but in practice people often apply evidence standards inconsistently — demanding rigorous proof for things they’re skeptical of, while accepting things they already believe with almost no scrutiny at all.

A useful rule of thumb: the strength of evidence required should match the size of the claim.

Calibrating by Claim Size

Think of it as a sliding scale. A small, ordinary claim needs very little evidence — sometimes none at all:

Your colleague mentions they had a sandwich for lunch. You don’t need documentation.
A friend tells you the bus was late this morning. You take their word for it.

These claims are low-stakes, easily verifiable if needed, and consistent with normal everyday experience. Demanding evidence for claims like these would be exhausting and socially absurd.

Now consider a larger claim:

A news article says a new medication cures a common disease. The stakes are higher, the claim is more significant, and you’d reasonably want to know: what kind of study? How many people? Peer reviewed? Who funded it?
Someone tells you that a well-known public figure committed a serious crime. Before repeating that, you’d want a credible source, not just a social media post.

And larger still:

A claim contradicts well-established scientific consensus — that climate change isn’t happening, or that a particular group of people is genetically inferior. Claims this large require extraordinary evidence, because they’re not just asserting a new fact — they’re asserting that an enormous body of existing evidence is wrong.

The famous formulation, often attributed to Carl Sagan, puts it simply: “Extraordinary claims require extraordinary evidence.”

The Asymmetry Trap

One of the most common errors in reasoning is applying this scale asymmetrically — holding claims you dislike to a high standard while waving through claims you find agreeable with almost no scrutiny. This is closely related to confirmation bias, which you encountered in the Bare Essentials level.

A well-calibrated thinker applies the same proportionality standard regardless of whether they want the claim to be true. That’s harder than it sounds, and it’s one reason intellectual honesty comes up later in this level as something that requires active effort, not just good intentions.

Absence of Evidence

One more important point: absence of evidence is not the same as evidence of absence — but it’s not meaningless either.

If a claim is the kind that should produce detectable evidence if true, and no such evidence exists despite people looking for it, that’s relevant information. It shifts the probability. It doesn’t prove the claim is false, but it does make it less likely. The weight you give to that depends on how thoroughly people have looked, and how detectable the evidence would be if the claim were true.

For example: scientists have been actively searching for signs of large-scale alien civilisations for decades using increasingly sensitive instruments. The absence of a signal doesn’t prove they don’t exist — but it does meaningfully update the probability.

Confidence vs. Reliability

Here is something that most people know in the abstract but consistently underestimate in practice: confidence and reliability do not correlate. A person can be completely, sincerely, convincingly certain about something they are completely wrong about. And a person can be cautious, hesitant, and full of caveats about something they are entirely correct about.

This matters because we are wired to treat confidence as a signal of competence. In many social situations throughout human history, that was a reasonable heuristic — someone who had experience and knowledge often did carry themselves with assurance. But it was never a reliable rule, and in the modern information environment it has become a significant liability.

(For more on how our emotions evolved in an environment very different than our modern life, see Emotion Management: Bare Essentials, How Emotions Work, up to and including “The Modern Mismatch”.)

How This Gets Exploited

Skilled deceivers — whether deliberate fraudsters, manipulative politicians, or simply people who have never learned to doubt themselves — rely heavily on projected confidence. The performance of certainty is often more persuasive than actual evidence, because it triggers that deep social instinct to trust the person who seems to know what they’re doing.

This shows up in predictable patterns:

The financial advisor who speaks with absolute authority about guaranteed returns
The pundit who delivers a confident take on a complex geopolitical situation as though it were obvious
The person in a meeting who dominates discussion not because they know more, but because they hesitate less

None of this means confident people are untrustworthy. Many knowledgeable, reliable people are also confident. The point is that confidence alone tells you nothing either way — it is simply not useful information about whether someone’s claims are accurate.

A Note on AI

This is worth connecting explicitly to something covered in the Technology & Society Intermediate level: large language models (LLMs) — the AI systems behind tools like chatbots — are remarkably good at sounding confident and authoritative. They produce fluent, well-structured, assured-sounding text even when they are wrong, and they do not hedge or express uncertainty the way a careful human expert might.

This is not a flaw exactly — it is a product of how they are built. But it means the confidence-as-reliability instinct is particularly dangerous when interacting with AI tools. The output can feel trustworthy in a way that has nothing to do with its actual accuracy. Treat AI-generated information the way you would treat any other confident-sounding source: evaluate the claim, not the delivery.

(For more about LLMs and AI systems, see Technology & Society Intermediate level.)

The Practical Takeaway

When evaluating a claim, consciously separate two questions:

How confident does this person seem?
What is the actual evidence for what they’re saying?

The first question is almost always easier to answer — and almost always less relevant. Training yourself to reach for the second question, even when the first one feels satisfying, is one of the more valuable habits this program can help you build.

Personal Certainty Is Not Evidence

So far we’ve looked at why other people’s confidence isn’t a reliable guide to accuracy. Now for the harder version: the same applies to you.

This is genuinely difficult to sit with. When you are absolutely certain of something — when you know what you saw, or heard, or experienced — it feels categorically different from being merely convinced by an argument. The subjective experience of certainty feels like its own kind of proof. It isn’t.

The Unreliability of Human Perception and Memory

Human perception is not a recording. It is a construction — your brain takes incomplete sensory data and fills in the gaps using context, expectations, prior experience, and pattern recognition. Most of the time this works well enough. But it means that what you perceive is always, to some degree, an interpretation — and interpretations can be wrong.

Memory compounds this. Rather than storing events like files on a hard drive, memory reconstructs them each time they are recalled — and each reconstruction is subtly shaped by what you’ve learned, felt, and experienced since. Memories feel vivid and real even when they have drifted significantly from what actually happened. This is not a character flaw. It is how human memory works for everyone.

This is covered in more depth in the Psychology topic, but the practical implication here is straightforward: sincere, vivid, detailed personal certainty is not the same as evidence. You can be completely, genuinely convinced of something that did not happen the way you remember it — or did not happen at all.

“I Know What I Saw”

A classic example: someone reports seeing something they can’t explain — an unusual light in the sky, a figure in the dark, an event that seemed impossible. They are not lying. They are not confused in any simple sense. They are certain. And yet certainty alone cannot establish what actually happened, because the human mind can generate the experience of certainty from incomplete, ambiguous, or misinterpreted information.

This is not a reason to dismiss personal experience entirely — personal experience is real and valuable data. But it is a reason to hold even your own most vivid experiences with some degree of humility, particularly when the claim they seem to support is an unusual one. Remember the proportionality principle from the previous section: the more extraordinary the claim, the more evidence it requires — and “I’m certain I saw it” is a single data point from a perceptual system that is known to make errors.

This is, incidentally, exactly why the scientific method was developed. Not because scientists distrust people, but because everyone’s perception and memory are fallible, and systematic methods of verification exist precisely to catch the errors that individual certainty cannot.

The same goes for objectivity: in order for something to be considered objective, more than one person must be able to experience the same thing in the same way, and the more the better. A person may have the most vivid, detailed memory of their time on an alien space ship, but without anyone else to corroborate it, and no physical evidence, other explanations must be considered, and it may very well remain a mystery.

This Is Not About Doubting Yourself

It is worth being clear about what this section is not saying. It is not saying you should distrust yourself, second-guess every experience, or refuse to act on your own judgment. Practical life requires trusting your perceptions most of the time, and excessive self-doubt is its own kind of problem.

The point is narrower: when a claim based on personal experience is significant, unusual, or consequential — especially when you’re asking others to act on it — it deserves the same proportional evidence standard as any other claim. Your inner certainty is real. It just isn’t, by itself, transferable to someone else as proof.

And if someone doesn’t take your word for something extraordinary or important, that isn’t an insult. It’s critical thinking — the same tool you’re building right now. Good mental habits should be applauded.

How to Evaluate a Source

Knowing that not all sources are equally reliable is easy. Actually evaluating a specific source in the moment — quickly, without a research team — is harder. What follows is a practical framework: a set of questions you can ask about almost any source to get a reasonable sense of how much weight to give it.

None of these questions are individually decisive. Think of them as evidence that accumulates — the more a source checks out across these dimensions, the more confidence is warranted.

Questions to Ask

1. Who is this, and what is their relevant expertise? Credentials matter, but they need to be the right credentials. A cardiologist is an expert on heart disease, not necessarily on climate science. A historian who specialises in medieval Europe may not be a reliable guide to contemporary geopolitics. Look for expertise that is actually relevant to the specific claim being made.

2. What is their track record? Has this source been reliable in the past? Have they made significant errors, and if so, did they acknowledge and correct them? A source that has been wrong repeatedly without acknowledgment, or that has a history of sensationalism, earns less trust — regardless of how authoritative they sound right now.

3. Do they have a conflict of interest? Who funds this source? What do they stand to gain if you believe what they’re saying? A conflict of interest doesn’t automatically disqualify a source — people with relevant expertise often also have financial or ideological stakes in a topic — but it is important information that should factor into your assessment.

4. Is the claim independently corroborated? Does the same claim appear in other credible sources that are genuinely independent of this one — meaning they didn’t all get it from the same original source? Convergence from multiple independent sources is one of the strongest signals of reliability. A claim that exists in only one place deserves more scrutiny.

5. Do they distinguish between what they know and what they think? Reliable sources tend to hedge appropriately — they signal the difference between established fact, reasonable inference, and speculation. A source that presents everything with equal certainty, never acknowledges uncertainty, and never says “we don’t know yet” is likely overstating its confidence. That’s a warning sign, not a sign of authority.

6. Are they engaging with opposing evidence? Trustworthy sources generally acknowledge the strongest counterarguments or contradictory evidence and explain why they weigh the evidence as they do. Sources that ignore or dismiss contradictory evidence without engaging it are doing something closer to advocacy than honest inquiry.

Red Flags

Some patterns are worth treating as immediate prompts for more scrutiny:

The source cannot be identified or verified
All citations lead back to the same source or affiliated sources
Critics are attacked personally rather than their arguments being addressed (you may recognise this as the ad hominem fallacy from the Bare Essentials level)
You are discouraged from checking other sources
Urgency is used to pressure you into accepting a claim before you’ve had time to think

flowchart TD
    A[You encounter a claim] --> B{How significant\nor unusual is it?}
    B -->|Small and ordinary| C[Basic awareness is enough\nNo deep evaluation needed]
    B -->|Significant or unusual| D[Evaluate the source]
    D --> E[Who are they?\nRelevant expertise?]
    E --> F[Track record?\nConflicts of interest?]
    F --> G[Independently corroborated?]
    G --> H[Do they distinguish\nknowing from thinking?]
    H --> I[Do they engage with\nopposing evidence?]
    I --> J{How does it\nadd up?}
    J -->|Strong on most counts| K[Higher confidence\nappropriate]
    J -->|Mixed| L[Proceed with caution\nSeek more sources]
    J -->|Weak on most counts| M[Significant skepticism\nwarranted]

A Note on Expertise and Consensus

One particular application of source evaluation is worth naming directly: when a strong scientific consensus exists on a topic, that consensus deserves significant weight. Not because scientists are infallible — they aren’t, and the Bare Essentials level of Science covers this honestly — but because scientific consensus represents the accumulated, independently reviewed judgment of large numbers of experts actively working to find errors in each other’s reasoning. A single dissenting expert, however credentialed and confident, is a much weaker signal.

This connects back to proportionality: overturning a scientific consensus is an extraordinary claim, and it requires extraordinary evidence.

Going Further

If you want to develop this skill further, media literacy researchers have developed several practical frameworks for rapid source evaluation. The SIFT method (Stop, Investigate the source, Find better coverage, Trace claims to their origin) is one of the more practical and well-regarded approaches for online information specifically — and it connects naturally to what you’ll find in the Media Literacy section coming up next, and is described in more detail there.

Putting It Together: A Working Habit

The concepts in this section are not difficult to understand. The hard part is applying them consistently, in real time, when you’re tired or busy or emotionally invested in a particular outcome. That gap — between knowing how to think critically and actually doing it — is what separates a toolkit from a skill.

The goal is not to become a suspicious, exhausting person who demands citations before believing anything. It’s to build a set of reflexes that engage automatically when they’re needed, so that proportional, evidence-based thinking becomes your default rather than something you have to consciously force.

A Simple Mental Checklist

When you encounter a claim that matters — significant enough to act on, share, or build other beliefs on top of — run through these questions quickly:

How likely is this, and based on what? (Probability, not certainty)
How big is this claim, and how much evidence does that size require? (Proportionality)
How confident does the source seem, and is that confidence actually informative? (Confidence ≠ reliability)
Am I certain of this myself — and is my certainty actually evidence? (Personal certainty)
Who is the source, and how do they hold up? (Source evaluation)

You won’t always have time for all five. And many claims won’t need all five — a small, ordinary claim needs very little scrutiny, as discussed earlier. But for claims that are significant or unusual, running through even a few of these questions quickly is far better than accepting or rejecting based on gut feeling alone.

Habits Take Time

Don’t expect to do this perfectly at first. These reflexes run counter to some deeply ingrained human tendencies — the comfort of certainty, the persuasive pull of confidence, the special weight we give our own experiences. Knowing about a bias or a flaw in reasoning does not automatically neutralise it. That’s one of the more humbling findings of cognitive psychology, and it’s worth being honest about.

What helps is practice — which is what the exercises at the end of this topic are designed to provide. The more you deliberately apply these questions in low-stakes situations, the more naturally they’ll engage when the stakes are higher.

A Note on Intellectual Generosity

One final point: thinking probabilistically and evaluating sources carefully does not mean assuming everyone is lying or wrong. Most people most of the time are doing their best to convey what they genuinely believe. The tools in this section are not for detecting deception — they’re for accurately calibrating your confidence in a world where even honest, intelligent, well-meaning people can be wrong, including you.

Approach other people’s claims — and your own — with curiosity rather than suspicion. The question is never “is this person trying to fool me?” It’s “what does the evidence actually support?”

How It Connects

The principles in this section — proportional evidence, probabilistic thinking, source evaluation, and the limits of personal certainty — run through almost every other topic in this program. Evidence-based thinking is not a standalone skill; it is the foundation on which most of the other tools here are built.

SOS: Separation of Objective from Subjective (Critical Thinking - Bare Essentials): SOS and evidence evaluation work together closely. SOS identifies which claims are objective — and therefore require evidence — and which are subjective — and therefore don’t. The evidence evaluation framework in this section is what you apply once SOS has identified a claim as objective or mixed. The two tools are most powerful used together.

Media Literacy (this topic): The source evaluation framework and the proportionality principle from this section are the analytical backbone of media literacy. SIFT is essentially a fast, practical version of the same process. Media literacy applies these principles to a specific and important environment; this section provides the underlying reasoning.

Psychology: The discussion of personal certainty and the limits of human perception and memory in this section connects directly to Level 2: Psychology, where the mechanisms behind those limitations are examined in depth. Understanding why memory is reconstructive rather than reproductive, and why perception is interpretive rather than recording, gives the “I know what I saw” principle its full grounding.

Emotion Management: Emotional investment in a belief is one of the most reliable obstacles to applying proportional evidence standards honestly. When we want something to be true — or fear it might be — the evidence threshold we apply tends to shift without our noticing. Level 2: Emotion Management covers this dynamic and provides tools for recognising and managing it.

Science: The scientific method is a formalised, institutional version of the proportional evidence principle. Scientific consensus represents the strongest available form of independent corroboration — many experts, working independently, attempting to disprove each other’s findings. Level 2: Science covers both the method itself and how to evaluate scientific claims, which makes it a direct companion to this section.

Technology & Society Intermediate: The connection between confidence and reliability — specifically the observation that AI language models produce fluent, assured-sounding output regardless of accuracy — is made explicit in this section and covered in depth in Level 2: Technology & Society Intermediate. The two sections reinforce each other directly.

Long-term Thinking: Probabilistic thinking requires patience — the willingness to hold a question open rather than forcing a premature conclusion. Level 2: Long-term Thinking covers the broader skill of deferring immediate certainty for better outcomes over time, which applies directly to evidence evaluation.

Communication Skills: How you communicate your level of certainty matters. Saying “the evidence suggests” rather than “it’s a fact that” is not weakness — it is accurate representation of epistemic status. Level 2: Communication Skills covers how to express calibrated confidence in ways that are both honest and credible.

Level 3 — Systems Thinking and Social Change Strategies: Complex systems rarely permit certainty. Systems Thinking at Level 3 requires comfort with probabilistic reasoning and incomplete evidence — the same disposition developed here. Social Change Strategies covers how false certainty and manufactured consensus are used as tools of manipulation at scale, which is the Level 3 application of the principles in this section.

Practice Exercises

Comprehension

In your own words, explain the difference between seeking certainty and seeking calibration. Why does the distinction matter practically?
Explain the proportional evidence principle using an example of your own — one small claim that needs little evidence, and one large claim that needs considerably more.
What does “absence of evidence is not the same as evidence of absence” mean? Give an example of a situation where this distinction matters.
List the six questions from the source evaluation framework. For each one, explain in a sentence why that question is relevant to assessing reliability.

Reflection

Identify a belief you currently hold with strong confidence. Examine it honestly: what is that confidence actually based on? Is it proportional to the evidence you have, or has certainty crept in beyond what the evidence strictly supports?
Think of a time you trusted someone primarily because they seemed confident and authoritative. Looking back, was that confidence a reliable indicator of accuracy? What would source evaluation have revealed that the confidence alone didn’t?
Think of something you were once certain about that turned out to be wrong or significantly more complicated than you thought. What gave you that certainty at the time? What changed it? What does that experience tell you about personal certainty as evidence?
Consider a topic where you have strong feelings — political, social, personal. Honestly examine whether you apply the same evidence standard to claims that support your position as to claims that challenge it. If not, what would it take to apply the same standard to both?

Application

Take five claims you encounter in a single day — from news, conversation, social media, or advertising — and assign each one a rough probability rather than a true/false verdict. What information would you need to refine each estimate? Notice how this shifts your relationship to the claims compared to simply accepting or rejecting them.
Find a source making a significant claim in an area you care about. Apply the full source evaluation framework to it: credentials, track record, conflicts of interest, independent corroboration, whether it distinguishes knowing from thinking, and whether it engages with opposing evidence. What do you find? Does your assessment of the source change?
Find an example of an extraordinary claim being made with insufficient evidence — in media, advertising, politics, or everyday conversation. Identify specifically what evidence would actually be required to support the claim proportionally, and what evidence is actually being offered. What is the gap between the two?
For one week, practice expressing your beliefs with calibrated language — “the evidence suggests,” “I think it’s likely that,” “I’m not certain but,” “as far as I know” — rather than stating them as flat certainties. Notice how others respond, and notice how it feels. Does it feel like weakness, or precision?

Discussion

Share a belief you hold with high confidence with a partner. Have them apply the source evaluation questions to the evidence you offer for it — not to attack the belief, but to examine the quality of its foundations together. Then swap roles. What does each of you discover about the difference between feeling certain and having strong evidence?
As a group, discuss the “I know what I saw” principle using a historical or hypothetical example — a reported supernatural event, a disputed eyewitness account, a case where multiple sincere witnesses described the same event differently. What does the example reveal about the relationship between personal certainty and reliable evidence? How do you think about personal experience differently after this discussion?
Discuss with a partner: is it possible to be too skeptical? Where is the line between healthy evidence evaluation and paralysing doubt? How do you make practical decisions when the evidence is genuinely incomplete — which it almost always is?

Key Sources & Further Reading

Practical Tools

Calibration practice — Several free online tools allow you to practice probabilistic thinking by answering trivia-style questions and assigning confidence percentages, then tracking how well-calibrated your estimates are over time. Search “calibration training quiz” or “credence calibration” to find current options. Regular practice builds the probabilistic thinking habit more effectively than reading about it.
Metaculus (metaculus.com) — A forecasting platform where users make probability estimates on real-world questions and track their accuracy over time. Engaging with it directly builds calibrated thinking as a lived skill.

Accessible Reading

Carl Sagan — The Demon-Haunted World (1995) — One of the most important books on evidence-based thinking ever written for a general audience. Sagan’s “baloney detection kit” chapter is essentially a warmly written, deeply humane version of this entire section. Essential reading for anyone serious about these skills.
Nate Silver — The Signal and the Noise (2012) — A thorough and readable exploration of probabilistic thinking applied to forecasting across domains — sports, politics, economics, natural disasters. Excellent for building intuition about calibration and the difference between signal and noise in evidence.
Philip Tetlock & Dan Gardner — Superforecasting (2015) — Based on decades of research into what makes some people significantly better at predicting outcomes than others. The answer is consistently about calibration and epistemic humility rather than expertise or confidence — directly relevant to the principles in this section.
Ben Goldacre — Bad Science (2008) — An entertaining and sharp examination of how scientific evidence is misrepresented in media, advertising, and medicine. Particularly good on the gap between what research actually shows and how it gets reported — connects directly to the proportional evidence and source evaluation material here.

On Human Memory and Perception

Elizabeth Loftus — A psychologist whose decades of research on false memories is among the most important evidence for why personal certainty is not the same as accurate recollection. Her work is covered in academic papers but also in accessible talks and interviews — searching her name will surface both. Her TED talk is a good starting point.
Daniel Schacter — The Seven Sins of Memory (2001) — An accessible account of how and why human memory fails in predictable ways. Provides the psychological grounding for the personal certainty discussion in this section.

Deeper Reading

Daniel Kahneman — Thinking, Fast and Slow (2011) — A comprehensive account of the two systems of human thinking — fast, intuitive, and prone to bias versus slow, deliberate, and more reliable — by one of the founders of behavioural economics. Dense but foundational; much of what is discussed in the cognitive biases sections of this program traces back to Kahneman’s research.
Sam Wineburg — Why Learn History (When It’s Already on Your Phone) (2018) — A media literacy researcher examines how professional fact-checkers evaluate sources differently from students and experts, and what “lateral reading” — quickly checking what others say about a source before reading it — actually looks like in practice. Directly relevant to the source evaluation framework in this section.

Continue to Critical Thinking Intermediate Part 2 →

Return to the main page.