An interactive introduction — no formulas, no prior knowledge needed. Learn through a story, step by step.
For students with no statistics background
Imagine you are a junior doctor. Your supervisor calls you in and says:
"We need an answer to one question: what is the average blood pressure of adult hypertensive patients in this country?"
You live in a country with 3,000,000 adult hypertensive patients.
Each one of them has a blood pressure value. That number is real — it exists, even if we do not know it.
Question for you
What would be the most practical way to get an answer?
Step 1 of 8
Exactly! Measuring everyone is not feasible. That is why in medicine — as in all sciences — we take a sample.
Below you can see all the patients (each dot = one patient). Press the button and watch what happens.
Patients not in the study Patients in your study
Population = all patients we are interested in (500 dots here, millions in reality). Sample = the small, randomly chosen group we actually measure.
Step 2 of 8
You have now measured blood pressure in 50 patients.
You have 50 different numbers. How do you summarise them into one?
Add up all values and divide by the number of patients — you get the mean (average).
We write it as x̄ (pronounced "x bar").
This is your estimate of the true average blood pressure. You do not know the true value — but this is your best approximation.
Question
You measured 5 patients and got: 130, 145, 128, 152, 135 mmHg. What is the mean?
Step 3 of 8
But wait — your colleague ran the same study, with the same method, in the same country. They got a slightly different mean. How is that possible?
Every time you pick a different group of 50 patients, you get a slightly different mean.
This is called sampling error — and it is completely normal. It is not a mistake by the researcher. It is the natural variability between samples.
Step 4 of 8
So every sample gives a slightly different result. But what if we take a larger sample? Will the estimate be more accurate?
Try it yourself — move the slider and watch how the means from different studies cluster closer to the true value.
20
Larger sample = more accurate estimate. This is one of the most important principles in clinical research.
A study with 10 patients gives a much less reliable conclusion than one with 200.
Question
Two studies investigate the same drug. Study A has 15 patients, Study B has 150. Which gives a more reliable conclusion?
Step 5 of 8
Good. Now let us look at another important concept — how spread out the data are.
Imagine two hospital wards:
Ward A
Patients' blood pressure: 138, 140, 139, 141, 137
Mean: 139 mmHg
Ward B
Patients' blood pressure: 110, 130, 139, 155, 181
Mean: 139 mmHg — the same!
Both wards have the same mean — but the patients in Ward B are far more varied.
Standard deviation (SD) measures this "spread".
Small SD = patients are similar to one another.
Large SD = patients differ a great deal.
Question
Which ward has the larger standard deviation?
Step 6 of 8
Great! Now we reach something you will read in every clinical paper:
"Drug X reduces blood pressure by 12 mmHg (95% CI: 8 – 16)"
What does "95% CI" actually mean?
When you measure one sample, you get one estimate (e.g. "12 mmHg"). But you know that a different sample would give a slightly different estimate.
Instead of reporting just one number, we calculate a confidence interval — a range within which the true value most likely lies.
Each line = one study. Green lines "capture" the true value. Red ones miss it.
A 95% CI means: if you repeated the same experiment 100 times, about 95 of the intervals would contain the true value.
Narrower interval = more precise study (usually larger n).
Step 7 of 8
And finally — the p-value. Probably the most famous and most misused term in medicine.
You give a drug to 40 patients and a placebo to 40 patients. After 3 months:
• Drug group: mean blood pressure drop of 8.5 mmHg
• Placebo group: mean blood pressure drop of 3.2 mmHg
Difference = 5.3 mmHg. But — is this a real difference, or just chance?
The p-value answers the question: "How likely is it that I would see a difference this large by chance alone, if the drug actually did nothing?"
p < 0.05 = less than 5% chance this is a fluke → we say "statistically significant" p > 0.05 = not enough evidence to conclude the drug works
Critical thinking
A drug reduces blood pressure by 1.2 mmHg. p = 0.001. What do you conclude?
Step 8 of 8
Congratulations — you have been through the full foundation of statistics! Let us recap what we learned:
1
Population and sample
We cannot measure everyone — we take a representative, random subset.
2
Mean (x̄)
Summarises a group of patients into one number — but it is not the whole story.
3
Sampling error
Every sample gives a slightly different result — this is normal, not a researcher mistake.
4
Larger n = more accurate estimate
More patients means less sampling error and a more reliable study.
5
Standard deviation (SD)
Measures how varied patients are — the "spread" of the data.
6
Confidence interval (CI)
The range within which the true value most likely lies. Narrower = more precise study.
7
p-value
Measures the probability of a chance finding — but p < 0.05 does not mean clinically important!
The most important lesson: Statistically significant ≠ clinically important.
Always ask: how large is the effect? How wide is the CI? How many patients were in the study?
You are now ready for more advanced concepts: normal distribution, statistical tests (t-test, Mann-Whitney), and how to read statistical tables in clinical papers.