Research & Statistics Terms, Explained
News stories about studies rarely explain what the study actually showed. This section covers the vocabulary of research itself, including what a p-value means, what a control group is for, and why two studies can disagree without either of them being wrong. Inside a paper, the Clicked extension does the same job — a research paper assistant that explains any term you highlight, in place. These pages explain how research works, and are not medical or scientific advice.
Featured
Correlation vs. Causation: What's the Difference?
Two things moving together is not one causing the other, and the gap is where bad headlines live.
What is a p-value?
How easily luck alone could fake a result — and what 0.05 actually means.
What is Peer Review?
Who checks a study before publication, what that catches, and what it does not.
All 38 terms
What is a Confidence Interval?
The range a study reports around its result, and why a wide one means nobody knows yet.
What is a Control Group?
The group that gets nothing, and why a study without one proves almost nothing.
Correlation vs. Causation: What's the Difference?
Two things moving together is not one causing the other, and the gap is where bad headlines live.
What is a p-value?
How easily luck alone could fake a result — and what 0.05 actually means.
What is Peer Review?
Who checks a study before publication, what that catches, and what it does not.
How to actually read a research paper
Skim smart, break the jargon wall, and keep what you learned — one session, done right.
What is Sample Size?
How many were actually studied, why small numbers mislead, and why big ones are not automatically better.
What is Standard Deviation?
How far the numbers wander from their average — the spread two identical averages can hide.
What does statistically significant mean?
It means probably not luck. It does not mean big, useful, or true.
What is a Cohort Study vs a Case-Control Study?
Two ways to study a cause nobody can assign. Cohort follows people forward from the habit for years. Case-control starts with the ill and looks back. Only one gives a rate.
What is a double-blind study?
A trial where neither the participants nor the researchers know who got the real treatment, so nobody's expectations can shape the result.
What is a meta-analysis?
Many small studies merged into one answer, and why the merge can still mislead.
What is a preprint?
A research paper posted publicly before peer review. Fast, free to read, and not yet checked.
What is a randomized controlled trial (RCT)?
Chance picks who gets the pill, so the pill explains the result.
What is effect size?
How big a difference actually is, separate from whether it's real. The number news stories skip.
What is Number Needed to Treat (NNT)?
How many people must take a treatment, for a stated time, for one of them to be spared. Strokes 2.9 to 1.9 in 100: 100 take the pill so that one avoids a stroke.
What is regression to the mean?
An extreme result is usually followed by an ordinary one, and something else takes the credit.
What is selection bias?
When how you picked your sample decides the answer before the study starts. Big samples don't fix it.
What is Simpson's Paradox?
A comparison that flips once you split the data into groups. Berkeley took 44% of men and 35% of women, yet no department did.
What is Statistical Power?
A study's chance of finding an effect that is really there. Trials accept missing 1 in 5 by design, so \"no effect found\" is a question.
What is survivorship bias?
Judging by the winners because the losers aren't around to be counted. The most common way data lies.
What is the Base Rate Fallacy?
Judging a test result while forgetting how rare the thing is. A 1-in-1,000 disease, a 5% false alarm rate: 51 positives, one real. Count the crowd first.
What is the placebo effect?
The pill is sugar and the relief is real, which is why every drug has to beat it.
What is the Null Hypothesis?
The "nothing is going on" claim every study has to beat, and why beating it proves less than you think.
What are False Positives and False Negatives?
The two ways a test can be wrong, why fixing one makes the other worse, and how to choose.
What is a Confounding Variable?
The hidden third factor that makes two unrelated things look like cause and effect.
What is Publication Bias?
Why the studies that found nothing never get published, and what that does to everything you read.
What is a Funnel Plot?
The chart that shows when studies are missing from the record, and why it is not the sales funnel.
What is Anecdotal Evidence?
Why one person's story is not proof, why it feels like proof anyway, and what it is actually good for.
What is an Outlier?
The one number far from the rest, how it bends the average, and the three things it can turn out to be.
What is a Forest Plot?
Each study gets a row and a range. A range crossing the no-difference line means that study alone could not tell. The diamond at the foot is what they say together.
What is P-Hacking?
Nothing is faked and every number is real. The analysis was simply chosen after seeing the answers, from many quiet tries, and only the winning version was reported.
What is the Replication Crisis?
The discovery that many published studies fail when other researchers rerun them. Psychology redid 100 of its own in 2015 and 36 came back. One run is never proof.
What's the Difference Between an Odds Ratio and a Relative Risk?
A relative risk of 2 means twice as likely. An odds ratio of 2 means something close to that only when the outcome is rare, and headlines skip the difference.
What's the Difference Between Absolute Risk and Relative Risk?
A percentage change compares two chances without showing you either one. The same 50% rise fits 2 in a million and 2 in 10. Ask for the counts.
What's the Difference Between Percentage Points and Percent?
The gap when you subtract one percentage from another. Percent change divides that gap by the start. One move, two sizes.
What's the Difference Between Sensitivity and Specificity?
A yes-or-no test has two scores: the share of real cases it catches, and the share of non-cases it leaves alone. One number alone can hide a useless test.
What's the Difference Between an Observational Study and an Experiment?
Every study is one of two kinds: researchers assigned something, or they only watched. The kind decides whether a finding can say causes or only linked.
No matches in this category. Search the whole gallery →