A Kid's Guide to Data Tricks
Someone is lying to you with numbers. Let's catch them.
Here's something wild: numbers feel trustworthy. When someone says "Studies show..." or "The data proves..." your brain relaxes a little. Numbers seem objective. Scientific. Real.
But here's the secret that advertisers, politicians, and even your friends already know: numbers can be twisted to tell almost any story you want.
The same data can make a school look amazing or terrible. The same statistic can make a medicine sound like a miracle or a disaster. The same graph can make a tiny change look enormous or a huge change look invisible — just by changing how you draw it.
By the end of this book, you'll know 10 sneaky data tricks and exactly how to see through each one. You'll never look at a graph, a headline, or a "9 out of 10 dentists recommend" claim the same way again. Ready to investigate? Let's go!
"It ain't what you don't know that gets you into trouble. It's what you know for sure that just ain't so." — Mark Twain
"Stretching the Truth"
Graphs are supposed to help us see data clearly. But sneaky graph-makers can manipulate axes, scales, and visual effects to make tiny changes look massive — or huge changes look like nothing. A graph isn't lying with words. It's lying with shapes.
IN THE NEWS
By chopping off the bottom of the axis, a small change fills the whole graph and looks enormous. Always check where the axis starts!
IN A SCHOOL PRESENTATION
If the y-axis started at $0, both bars would look almost identical. The "zoom-in" trick makes small differences look dramatic.
IN AN AD
3D charts look cool but distort the data. They're almost always used to trick you, not to inform you.
ON SOCIAL MEDIA
Choosing the right time range can make any trend look like it's going up or down. A graph that shows only 6 months might tell a very different story than one showing 5 years.
Our brains process images faster than numbers. When we see a bar shooting up or a line plunging down, we feel the change before we read the actual values. Graph-makers exploit this by adjusting the visual scale while the real numbers barely move.
1. Always check: does the y-axis start at zero? If not, the visual scale is exaggerated.
2. Read the actual numbers, not just the shapes.
3. Be suspicious of 3D charts — flat charts are almost always more honest.
4. Ask: what time range is being shown? Would a longer or shorter range change the story?
Bonus: look for 3D graphs and ask yourself why they used 3D instead of flat. The answer is almost always "to make it look more impressive than it is."
"Only Showing the Good Stuff"
Cherry-picking is when someone selects only the data points that support their argument and conveniently ignores everything else. It's like showing your parents your three best test scores and hiding the other seven.
IN A PRODUCT REVIEW
Three hand-picked reviews can make any place look amazing. The full picture — all 200 reviews — tells a very different story.
IN SPORTS
Zooming in on a small winning streak hides the bigger losing picture. Always ask: what's the FULL record?
IN THE NEWS
Picking one data point (a cold week) and ignoring the trend (decades of warming) is textbook cherry-picking.
AT SCHOOL
One good day doesn't represent a whole week of study habits. The full picture matters.
We naturally pay more attention to evidence that confirms what we already believe. Cherry-pickers exploit this by giving us only the "confirming" data. Since we never see the contradicting data, everything feels like it lines up perfectly.
1. Always ask: "What data am I NOT seeing?"
2. Check the full dataset, not just the highlighted examples.
3. Be suspicious when someone shows only a few data points from a much larger set.
4. Look for the overall trend, not just individual moments.
That's the power (and the danger) of cherry-picking.
"The 'Average' Lie"
The word "average" sounds simple, but it hides a sneaky trap. There are different kinds of averages, and picking the wrong one — accidentally or on purpose — can paint a wildly misleading picture. The mean (add everything up and divide) gets wrecked by extreme values. The median (the middle number) often tells a much more honest story.
THE BILLIONAIRE IN THE COFFEE SHOP
One extreme value dragged the mean up by a ridiculous amount. The median didn't budge. When you hear "average," always ask: mean or median?
SCHOOL FUNDRAISER
Reporting the mean makes the fundraiser sound like a massive success. Reporting the median shows what the typical student actually did.
IN A JOB AD
The company used the mean to make salaries sound much higher than what most employees actually earn.
VIDEO GAME STATS
One lucky game doesn't make you an elite player. The median score tells you what to actually expect.
Most people hear "average" and picture "typical." But the mean can be wildly different from what's typical when extreme values are involved. People who want to mislead you will always pick whichever type of average makes their story sound better.
1. When you hear "average," ask: is this the mean or the median?
2. Look for outliers — extreme high or low values that could skew the mean.
3. Ask: what does the TYPICAL person/student/player actually experience?
4. If the mean and median are very different, that's a red flag.
Mean = $30.63. Median = $5. The mean is 6x higher than what most kids actually get!
"Sounds Scary, But Is It?"
Percentages sound impressive, but they're meaningless without knowing the base number — the number you started with. "200% increase!" sounds terrifying. But if it went from 1 to 3... that's just 3. The base number is everything.
IN A SCARY HEADLINE
A 100% increase from a tiny number is still a tiny number. Always ask: 100% of WHAT?
IN AN AD
Percentages from tiny samples sound more impressive than the raw numbers. "8 out of 11 women" doesn't quite have the same ring as "73%."
AT SCHOOL
The percentage makes the improvement sound massive. The actual numbers tell you it's a D either way.
IN A YOUTUBE VIDEO
Huge percentages applied to tiny numbers create fake drama. And mathematically, nothing can decrease by more than 100%!
Our brains react to big-sounding numbers. "200% increase!" triggers alarm bells before we even think about what 200% of a tiny number actually is. Advertisers and headline writers know this and exploit it constantly.
1. Always ask: "Percent of WHAT? What's the base number?"
2. Convert the percentage back into real numbers.
3. Remember: nothing can decrease by more than 100%.
4. Watch for percentages based on tiny samples — they sound scientific but aren't.
"Where Are the Ones Who Didn't Make It?"
Survivorship bias happens when we only look at the people or things that "survived" a process (the winners, the successes, the famous ones) and ignore all the ones that didn't. This creates a completely warped picture of reality because we're drawing conclusions from an incomplete dataset.
IN A MOTIVATIONAL VIDEO
We see the one who made it. We don't see the millions who took the same risk and lost. The survivors get the interviews — the failures stay invisible.
IN MUSIC
The success stories get all the attention. The identical-effort failures are invisible, making success look more common than it is.
IN HISTORY CLASS
Old things seem better because only the best examples survived. All the terrible old stuff is already dust.
IN THE APP STORE
Top charts show survivors. The graveyard of failed apps is massive but invisible.
Failures are invisible. We literally can't see them because they don't get coverage, don't get interviewed, and don't make headlines. Our brain builds its model of reality from what it can see — and what it sees is only the winners.
1. Always ask: "What happened to the ones who DIDN'T make it?"
2. Look for the failure rate, not just the success stories.
3. Remember: the news covers the exception, not the rule.
4. Be suspicious of advice based on a single success story.
"Who Did You Actually Ask?"
When someone makes a big claim based on data, two questions matter more than anything: How many people did they ask? And who did they ask? A tiny or biased sample can produce results that mean absolutely nothing — but they'll be presented as if they're facts.
AT LUNCH
Three friends isn't a survey. It's a chat. And if your friends all like tacos, that says more about your friend group than about the whole school.
IN AN AD
"4 out of 5" sounds scientific until you realize we know nothing about how those dentists were selected or what question they were actually asked.
ON A YOUTUBE CHANNEL
Self-selected samples (people who choose to respond) are almost always biased. The people who disagree simply didn't bother answering.
IN THE NEWS
A tiny, non-diverse sample tells you almost nothing about the general population. Twelve college kids aren't "everyone."
Numbers sound official. "4 out of 5" feels like a scientific result. But the quality of the sample determines the quality of the data. A perfect study with the wrong sample is still wrong.
1. Ask: "How many people were in the sample?"
2. Ask: "Who chose to participate? Could they be biased?"
3. Be suspicious of surveys where people self-select (online polls, fan votes).
4. Good research uses large, random, diverse samples — look for those words.
"Rare Things Stay Rare"
The base rate trap happens when you forget how rare something is in the first place. Even a very accurate test can give you mostly wrong results if the thing it's testing for is extremely rare. This is one of the trickiest concepts in data — and even doctors get it wrong!
THE DISEASE TEST
Even with a 99% accurate test, when the disease is rare, most positive results are false positives. The BASE RATE (how common the disease is) matters enormously.
SCHOOL SECURITY
When the thing you're looking for is rare, false alarms outnumber real catches — even with a very accurate system.
IN AIRPORT SECURITY
High accuracy + rare events = lots of false alarms. This is why your water bottle gets confiscated but actual threats are extremely rarely caught this way.
"99% accurate" sounds nearly perfect. Our brains hear that and think "only 1% chance of being wrong." But when the thing you're looking for is super rare, that 1% error rate creates way more mistakes than real catches. We forget to consider how common or rare the thing is in the first place.
1. Always ask: "How common is this thing in the first place?"
2. If something is very rare, even very accurate tests will produce lots of false positives.
3. Don't be dazzled by accuracy percentages without knowing the base rate.
4. When in doubt, think about 10,000 people and walk through the math.
"Same Number, Different Spin"
Framing is when the exact same fact is presented in a way that makes you feel completely different about it. "90% fat-free" and "10% fat" are the same thing — but one sounds healthy and the other sounds gross. The data hasn't changed. Only your feelings about it have.
IN THE GROCERY STORE
Marketers always choose the frame that sounds better. "90% fat-free" is the positive frame of a fact that's less appealing when stated as "10% fat."
IN A NEWS REPORT
Whether you frame it as "25% failed" or "75% passed" completely changes the emotional reaction — even though the fact is identical.
AT THE DOCTOR
Loss framing (focusing on the bad outcome) feels scarier than gain framing (focusing on the good outcome), even with identical numbers.
ON A REPORT CARD
We naturally choose the frame that supports our position. The kid emphasizes the 70% right; the parent emphasizes the 30% wrong.
Our brains are wired to react differently to gains vs. losses, even when they're the same thing. Framing exploits this emotional wiring. We don't process "90% fat-free" and "10% fat" as the same fact — we FEEL them differently.
1. Flip the frame. If something is presented positively, restate it negatively (and vice versa).
2. Ask: "Would I feel differently if this same number was presented the other way?"
3. Focus on the actual number, not the spin around it.
4. Be suspicious when only one frame is given — someone chose that frame for a reason.
"Garbage In, Garbage Out"
Surveys and polls can look scientific, but the way questions are written can control the answers. Leading questions, loaded words, limited choices, and biased samples can make any survey produce any result you want. If you put garbage questions in, you get garbage data out.
LEADING QUESTIONS
Leading questions are designed to push you toward a specific answer. A fair question is neutral and lets you answer honestly.
LIMITED CHOICES
If the options don't include all reasonable answers, the results are rigged from the start.
LOADED SCALES
When every answer option is positive, you can't express a negative opinion. The results will always look good.
ONLINE POLLS
Self-selected online polls are popularity contests, not real research. The people who bother to vote are NOT representative of the general population.
"A survey found that..." sounds like science. But the quality of a survey depends entirely on how the questions are written, who is asked, and what answer options are given. A badly designed survey can "prove" literally anything.
1. Read the actual question that was asked — was it neutral or leading?
2. Check the answer options — are all reasonable viewpoints represented?
3. Ask who was surveyed — random sample or self-selected group?
4. Look for the word "online poll" — that's almost always meaningless data.
"When Data Tells a Fake Story"
Just because two things happen at the same time doesn't mean one caused the other. This is called correlation without causation, and it's one of the most common data tricks in the world. Two data lines can move together for completely unrelated reasons — and the results can be absolutely hilarious.
THE CLASSIC: ICE CREAM AND DROWNING
The hidden third variable (summer heat) causes BOTH. This is called a confounding variable — an invisible factor that drives both trends.
PIRATES AND GLOBAL WARMING
With enough data, you can find correlations between almost ANYTHING. That doesn't make any of them meaningful.
ON SOCIAL MEDIA
Wealth, education, and nutrition all go together in developed countries. Attributing Nobel Prizes to cheese is hilarious but wrong.
IN SCHOOL
There could be many hidden variables. Families that ensure breakfast may also ensure homework time, bedtimes, and school attendance. Correlation isn't causation.
Our brains are pattern-matching machines. When we see two things moving together, we instinctively create a story connecting them. "A went up AND B went up, so A must cause B!" It feels logical. But the universe is full of coincidences, hidden variables, and trends that move together for no meaningful reason at all.
1. The golden rule: correlation does NOT equal causation.
2. Ask: "Is there a hidden third variable causing BOTH?"
3. Ask: "Could this be a coincidence?"
4. Ask: "Does it even make logical sense that A would cause B?"
5. For fun, visit tylervigen.com — a whole website of absurd spurious correlations!
Your 5 questions to crack any number's case
Learning to question numbers isn't about becoming suspicious of everything. It's about becoming smarter about the information you consume.
When you can spot data tricks, you:
Numbers are everywhere — in your grades, your game stats, the news, your doctor's advice, and every ad you see. Learning to read them critically is one of the most powerful skills you can develop.
There's an important difference. A skeptic asks good questions and looks for real evidence. A cynic assumes everything is a lie and trusts nothing. Don't become a cynic.
Most people sharing data aren't trying to trick you. But some are. And even well-meaning people can accidentally mislead with numbers. Your job isn't to assume the worst — it's to ask the right questions so you can tell the difference.
Data, when used honestly, is one of the most powerful tools in the world. It cures diseases, solves problems, and helps us understand reality. The tricks in this book are the exceptions — but they're common enough that you need to know how to spot them.
You've graduated from Data Detective Academy. You know the 10 tricks. You have the 5 questions. You're ready.
The next time someone throws a number at you, don't just accept it. Ask questions. Check the source. Flip the frame. Look for what's missing. You've got this.