What Is Statistics?
Statistics is the science of learning from data when you can't see everything. You almost never get to measure an entire population, so you take a manageable sample, summarize it honestly, and use it to make a careful, uncertainty-aware guess about the whole. This course is about doing that well.
Two jobs: describe, then infer
- Descriptive statistics summarizes the data you actually have: averages, spreads, charts. It makes no claims beyond the data in front of you.
- Inferential statistics takes the leap: from the sample in hand to a statement about the population you didn't measure. This is where uncertainty comes in, and with it most of the machinery of the rest of the course.
The vocabulary that trips everyone up
Two pairs of words. Get these straight now and everything later is easier (and when a later term ambushes you, the glossary defines every one the site uses):
- A population is everyone/everything you care about; a sample is the subset you actually measure.
- A parameter is a true (usually unknown) number about the population, such as its mean μ (mu). A statistic is the matching number computed from your sample, such as the sample mean x̄. We use the statistic to estimate the parameter.
🎮 Population → Sample → Estimate
The faint cloud is a whole population of 2,000 exam scores; its true mean μ (teal) is the parameter. Take a sample and your sample mean x̄ (orange) is the estimate. Bigger samples land closer.
Notice the honest tension: your estimate is almost never exactly right, but with a bigger sample it's reliably close. Quantifying "how close, how confident" is precisely what sampling distributions, the Central Limit Theorem, and confidence intervals are for.
Why not just measure everyone?
Usually you can't. The "population" might be all potential customers, every possible repetition of an experiment, or all humans with a condition: too big, too expensive, or not yet existing. Sampling is not a compromise to be embarrassed about. Done right, a few hundred well-chosen observations can pin down a population astonishingly well. The word doing the work there is well-chosen: no amount of statistics later can repair a sample that systematically missed part of the population.
Why it matters: every technique ahead is a variation on this one move — sample, summarize, generalize. Keeping "parameter vs statistic" and "population vs sample" straight is the foundation the whole course is built on.
Common questions
What is the difference between a parameter and a statistic?
A parameter is a number that describes the whole population (like the true mean μ), and it's usually unknown. A statistic is the matching number computed from your sample (like x̄), which you use to estimate the parameter. Memory hook: parameter–population, statistic–sample.
What is the difference between descriptive and inferential statistics?
Descriptive statistics summarizes the data you actually collected (means, standard deviations, charts) and claims nothing beyond it. Inferential statistics uses that sample to draw conclusions about the wider population it came from, which is why it always comes with uncertainty attached (confidence intervals, p-values).
Do I need to be good at math to do statistics?
Less than most people fear. The arithmetic in an applied course is squaring, adding, dividing and reading a value off a table, and software does all of it for you. What the subject genuinely demands is comfort with an argument: what a number is evidence for, what could have produced it by chance, and where a conclusion starts to outrun the data. People who found algebra a struggle often do well here, and people who found it easy are not automatically safe from the reasoning traps.