Difficulty: Beginner | Prerequisites: None
Statistics is the backbone of data-driven decision-making across nearly every field, from business forecasting to medical research to government policy. This opening material lays the groundwork for the entire course: what statistics is, how we talk about data, why we use samples instead of studying whole populations, and how to collect those samples properly. If you understand these foundations, every later topic (probability, distributions, hypothesis testing) will make more sense. If you skip them, the vocabulary alone will trip you up for the rest of the semester.
Statistics is the science of collecting, organising, and drawing conclusions from data. Because studying an entire population is usually impractical, we draw samples and use them to make inferences. How you draw that sample (random, stratified, cluster, systematic) determines whether your conclusions are trustworthy.
Statistics
The study of methods and procedures for collecting, classifying, summarising, and analysing data, and for making scientific inferences from that data. In simple terms, it is the toolkit for turning raw numbers into useful knowledge.
Descriptive statistics
Methods of organising, summarising, and presenting data in an informative way. Think of it as the "here is what the data looks like" side of statistics: charts, averages, tables.
Inferential statistics
Methods used to determine something about a population on the basis of a sample, including making decisions and drawing conclusions. In simple terms, this is where you go beyond describing what you see and start making predictions or generalisations.
Population
A collection, or set, of all units (people, individuals, objects, or events) that we are interested in studying. It can be finite or infinite. Think of it as the entire group you care about.
Sample
A subset of the units of a population. In simple terms, it is a smaller chunk drawn from the population so you do not have to study every single member.
Parameter
A numerical value summarising all the data of an entire population. Think of it as the "true" number you would get if you could measure everyone.
Statistic
A numerical value summarising the sample data. In simple terms, it is the estimate you calculate from your sample, which stands in for the parameter you cannot easily measure.
Variable
A characteristic about each individual element of a population or sample. Think of it as the question you are asking about each unit, such as "What is your height?" or "Which brand do you prefer?"
Data (singular, datum)
The value of the variable associated with one element of a population or sample. This value may be a number, word, or symbol.
Data (plural)
The set of values collected for the variable from each of the elements belonging to the sample.
Experiment
A planned activity whose results yield a set of data.
There are three practical reasons why researchers almost always work with samples rather than full populations:
The entire population may be difficult or impossible to locate. If you want to study every person with a certain medical condition worldwide, finding them all is not realistic.
Resources are limited. Time, money, and labour constrain what you can measure.
Some measurements destroy the item being tested. Quality-control testing of light bulbs, crash-testing cars, or taste-testing food means you cannot test every unit and still have product left to sell.
A well-chosen sample should have the same characteristics as the population it represents. When it does, you can generalise your findings with confidence.
With replacement: a member of the population may be chosen more than once. After each selection, the individual goes "back into the pool."
Without replacement: a member may be chosen only once. Once selected, they are removed from the pool for all future draws.
Most real-world surveys use sampling without replacement, but some probability calculations assume replacement to keep the maths simpler.
Sampling methods split into two broad families:
Random sampling: every member of the population has an equal chance of being selected. This is the gold standard for avoiding bias.
Non-random sampling: selection is not governed by chance. Results may be biased and are harder to generalise.
Simple random sample (SRS)
A sample of size n is selected so that every possible sample of that size has the same chance of being chosen. Drawing names from a hat (assuming thorough mixing) is the classic illustration.
Stratified random sample
The population is first divided into non-overlapping subgroups called strata (e.g., by age bracket, income level, or region). Then a separate simple random sample is drawn from each stratum. This guarantees representation from every subgroup.
Cluster sampling
The population is divided into non-overlapping groups called clusters (e.g., schools, city blocks, factories). Clusters are then selected at random, and every member within each chosen cluster is included. This differs from stratified sampling: in stratified sampling you sample within every group; in cluster sampling you sample entire groups but only some of them.
Systematic sampling (1-in-k)
You pick a random starting point among the first k individuals in an ordered list, then select every kth individual after that. For example, if k = 10 and your random start is person 7, you take persons 7, 17, 27, 37, and so on.
Convenience sampling
Selecting individuals or items simply because they are easy to reach. Surveying only your friends, or only people who walk past a certain spot, are common examples. Results are quick to gather but prone to bias.
Sampling methods show up everywhere. Political pollsters use stratified random sampling to ensure their surveys reflect the electorate by region, age, and party. Quality engineers use systematic sampling on a production line, pulling every 50th widget for inspection. Understanding which method fits which situation is a skill that carries well beyond this course.
Students often confuse a parameter with a statistic. A parameter describes the whole population; a statistic describes only the sample. If you measured the average height of every student at UCF, that is a parameter. If you measured a random group of 200 students, the average you calculate is a statistic.
Students sometimes think cluster sampling and stratified sampling are the same because both divide the population into groups. They are not. In stratified sampling you take a sample from every group. In cluster sampling you randomly pick some groups and take everyone in them.
Convenience sampling feels "good enough" to many students, but it almost always introduces bias. The fact that a sample is large does not fix the problem if the selection method was non-random.
⚠️ Be able to identify which sampling method is being described in a scenario. Exam questions typically give you a short story and ask you to name the method.
⚠️ Know the difference between a parameter and a statistic cold. This distinction underpins the entire course.
⚠️ Understand why sampling is necessary (the three reasons). Short-answer or multiple-choice questions on this are common early exam material.
⚠️ Stratified vs cluster sampling is a classic exam comparison. If the question says "a sample is taken from each subgroup," it is stratified. If it says "entire subgroups are selected," it is cluster.
True or false: A statistic describes an entire population.
Fill in the blank: In _______ sampling, the population is divided into strata and a random sample is drawn from each one.
True or false: Convenience sampling is a type of random sampling.
Fill in the blank: A _______ is a numerical value that summarises all the data of an entire population.
True or false: In cluster sampling, you sample some members from every cluster.
Answers: 1. False (that is a parameter). 2. Stratified. 3. False (it is non-random). 4. Parameter. 5. False (you select entire clusters, not some members from each).
Q: A researcher numbers all 5,000 employees at a company, uses a random number generator to pick 200, and surveys them. What sampling method is this?
A: Simple random sample, because every possible group of 200 employees had the same chance of being selected.
Q: A polling firm divides voters into age brackets (18–29, 30–44, 45–64, 65+) and randomly selects 100 voters from each bracket. What method is this?
A: Stratified random sampling. The age brackets are the strata, and a separate random sample is drawn from each.
Q: A factory inspector picks a random number between 1 and 20, gets 14, and then inspects every 20th item on the assembly line starting from item 14. Name the method.
A: Systematic sampling (1-in-k, where k = 20).
Q: Explain why a large convenience sample can still produce misleading results.
A: Convenience sampling selects units based on ease of access rather than randomness, so certain segments of the population are likely over- or under-represented. A larger sample size does not correct this selection bias.
Q: What is the key difference between a parameter and a statistic?
A: A parameter summarises data for an entire population; a statistic summarises data for a sample drawn from that population.
This material connects directly to probability (Lectures 3–4 in most syllabi), because the validity of probability-based inference depends on how the sample was collected. It also sets the stage for hypothesis testing later in the course, where you will use sample statistics to make claims about population parameters. If you go on to study regression or ANOVA, the assumption of a properly drawn sample remains central.
statistics, descriptive statistics, inferential statistics, population, sample, parameter, statistic, variable, data, datum, experiment, sampling, simple random sample, SRS, stratified sampling, strata, cluster sampling, systematic sampling, 1-in-k sampling, convenience sampling, random sampling, non-random sampling, sampling with replacement, sampling without replacement, STA 2023, Principles of Statistics, UCF