Critical Thinking Basics
Critical thinking is just asking good questions before you believe something. Learn to tell a claim from its evidence, size up a source, check things online the way fact-checkers do, and catch the tricks that weak arguments rely on.
Claims vs evidence
A claim is something someone wants you to believe. Evidence is the reason you should believe it. Most arguments you see online are mostly claim, with very little evidence attached. Your first job is to pull the two apart.
“This app makes kids smarter.”
A statement. On its own it proves nothing, no matter how confident it sounds.
“In a study of 400 kids...”
Something you can check: data, a study, a document, a recording, a direct observation.
Three kinds of claims
- Facts can be checked. “The bridge opened in 1937.” You can look this up.
- Opinions are judgments. “That's the prettiest bridge in America.” Reasonable people can disagree, so look for the reasons behind it.
- Predictions are about the future. “Prices will drop next year.” Ask what the prediction is based on and how often this source has been right before.
Not all evidence is equal
From weaker to stronger, roughly:
- Anecdote: one person's story. Real, but it might be the exception.
- Expert opinion: useful if the expert works in that exact field.
- Survey or observational data: shows patterns, but not always causes.
- Controlled experiment: compares groups that differ in one thing.
- Many studies reviewed together: a systematic review or scientific consensus. The strongest footing we have.
Four questions to ask any claim: What exactly is being claimed? What's the evidence? Who is saying it, and what do they gain? What would I expect to see if it were false?
Big claims need big evidence. “This tea cures cancer” needs far stronger proof than “this tea has caffeine.” The person making the claim has the job of proving it. It's not your job to disprove it.
Judge the source
Before you weigh what's said, figure out who is saying it. A great-looking website, lots of followers, or a professional logo tells you nothing about accuracy.
| Ask | Why it matters |
|---|---|
| Who made this? | A person, a news outlet, a company, an advocacy group? Can you find a real name? |
| Do they know this topic? | A doctor talking about medicine is different from a doctor talking about the economy. |
| What's their motive? | Are they selling something, pushing a cause, or trying to get clicks and shares? |
| Where did they get it? | Do they link to the original study, data or document? Or is it “sources say”? |
| When was it made? | Old news gets reshared as if it just happened. Check the date. |
Primary vs secondary sources
A primary source is the original: the study itself, the full video, the official report, the court filing. A secondary source describes it: a news article about the study, a post about the video. Secondary sources are fine, but details get lost or twisted with each retelling. When it matters, go to the original.
Watch for screenshots. A screenshot of a headline or post is easy to fake or crop. Search for the original before you share it.
Check a claim online with lateral reading
Stanford researchers who watched professional fact-checkers work found they did something most people don't: they left the page almost right away. Instead of reading a site top to bottom to decide if it seems trustworthy, they opened new tabs and looked up what other sources say about it. That's called lateral reading, reading across, not down.
The SIFT method
Mike Caulfield, a digital literacy researcher, boiled this down to four moves:
- Stop. Notice when something makes you feel strongly (angry, scared, smug). That's the moment to slow down, not share.
- Investigate the source. Open a new tab and search the name of the site, group or person. Add words like “funding” or “who runs.” A quick look at Wikipedia's article on an outlet is often a fine start.
- Find better coverage. Search the claim itself. Do trusted news outlets or fact-checking sites report the same thing? If only one site has the story, be careful.
- Trace claims, quotes and media back to the original context. Find the full study, the full speech, the first place the photo appeared. Quotes and clips are often cut to change their meaning.
- Stop: it's exciting news, and it's being shared a lot.
- Investigate: the site sells diet products. Hmm.
- Find better coverage: no major health outlet reports it.
- Trace: the “study” had very few people and wasn't published in a reviewed journal.
- Not convincing. Don't share it.
Photos and videos: use a reverse image search (Google Lens, or TinEye) to see where else an image appeared and when. A “breaking news” photo that's actually from years ago is one of the most common tricks.
Found something you're unsure about? Tick each step as you do it.
Common logical fallacies
A fallacy is a flaw in reasoning. The conclusion might still happen to be true, but the argument doesn't prove it. Fallacies are persuasive because they feel like reasons. Once you can name them, they lose their power.
| Fallacy | What it does |
|---|---|
| Ad hominem | Attacks the person instead of their argument. |
| Straw man | Twists someone's view into a weaker one, then knocks that down. |
| False dilemma | Offers only two options when there are more. |
| Slippery slope | Claims one step will lead to a chain of disasters, without showing why. |
| Bandwagon | “Everyone believes it, so it's true.” |
| Appeal to authority | Relies on someone famous or powerful who isn't an expert on this topic. |
| Post hoc | “B happened after A, so A caused B.” |
| Hasty generalization | Draws a big conclusion from a few cases. |
| Circular reasoning | Uses the conclusion as its own proof. |
| Red herring | Changes the subject to something that sounds related. |
| Appeal to emotion | Uses fear, guilt or pity in place of evidence. |
| Whataboutism (tu quoque) | Dodges criticism by pointing to the critic's own faults. |
| Cherry picking | Shows only the data that helps and hides the rest. |
| Appeal to ignorance | “No one has proven it false, so it's true.” |
Careful: not every mention of an expert is a fallacy. Citing a climate scientist about climate is good evidence. It becomes a fallacy when the authority isn't an expert on the topic, or when “they said so” replaces the actual evidence.
Cognitive biases: the bugs in your own brain
Fallacies are mistakes in an argument. Cognitive biases are mental shortcuts that push everyone's thinking in predictable directions, including yours. Smart people aren't immune. The fix is knowing where to look.
| Bias | What happens | Counter-move |
|---|---|---|
| Confirmation bias | We notice and remember what fits what we already believe. | Search for the best case against your view. |
| Anchoring | The first number we see sticks, like a “was $199” price tag. | Form your own estimate before you look at theirs. |
| Availability | Vivid or recent events feel more common than they are. | Ask for the actual rate, not the memorable story. |
| Survivorship bias | We only see the winners. The dropouts who became billionaires, not the many who didn't. | Ask: who's missing from this picture? |
| Sunk cost | We keep going because of what we've already spent. | Decide based on what happens from here. |
| Overconfidence | We think we know more than we do. | Ask “how would I know if I'm wrong?” |
| In-group bias | We trust people on “our side” more. | Ask if you'd accept the same evidence from the other side. |
“Old buildings were built better. Look how many from 1900 are still standing!” But you only see the ones that survived. The poorly built ones fell down or were torn down long ago.
Make it a habit: when you strongly agree with something, check it twice. That's exactly when your guard is down.
FAQCommon questions
Is critical thinking the same as being negative?
No. It's about matching your confidence to the evidence. Sometimes that means believing something more strongly, not less. A good critical thinker can also say “that's a solid source” or “I was wrong.”
If an argument has a fallacy, is the conclusion false?
Not necessarily. A fallacy means the argument doesn't prove its point. The conclusion could still be true for other reasons. Calling out a fallacy and then assuming the opposite is its own mistake (sometimes called the fallacy fallacy).
Can I trust Wikipedia?
It's a good starting point, especially for checking who runs a website or group. For important facts, follow the footnotes to the original sources it cites.
How do I talk to someone who shared something false?
Be kind and private if you can. Ask questions (“Where did this come from?”) rather than lecturing, and share a better source. People rarely change their minds when they feel attacked.
What about AI-generated images and text?
The same steps work: investigate the source, find other coverage and trace it to the original. Be extra careful with dramatic images that appear in only one place, and check whether reliable outlets have verified them.