BMS 1110 · Common Core (Tronc commun)Section 6.1 · Sampling, estimation and inference

Section 27 of 30

Population, sample, parameter and statistic

Read a study description and name its population and sample, say which number quoted is a parameter and which is a statistic, and match each Greek or Latin symbol to the quantity it denotes.

New to this course? How each section works

Learn gives you one short idea at a time. Each idea starts with a single sentence, then explains it, gives an example, and asks you a small question you can answer straight away.
Explore shows a diagram of the same ideas. Click the numbered parts to read what each one is.
Check asks a few questions. Answer them all correctly to finish the section and unlock the next one.
“Go deeper” is optional. It adds mechanism and extra detail for when you are ready — nothing you need for the checks is hidden there.

Why this matters

Almost everything you will read about health after this course is based on a sample, not on every patient. When a report says that 82 % of observed hand-hygiene moments were correct, somebody measured 1,200 moments out of millions and used that number to speak about the whole hospital. If you cannot tell the population from the sample, and the parameter from the statistic, you cannot judge how strong that claim really is — or whether it applies to your own ward at all. This short section gives you the four words and the symbol convention that the rest of unit 6 stands on.

What you need to know first

You need only the mean, the proportion, the variance and the standard deviation from unit 5, and you should be comfortable with the idea that a proportion is a fraction of a whole. No new arithmetic is introduced here: this section is vocabulary and notation, and the numbers we use are small.

Words to know (5)

Estimation: Using information from a sample to make an approximate numerical statement about an unknown characteristic of the whole population. Measuring 500 adults and concluding that the average systolic pressure of the city’s adults is about 122 mmHg is an estimation.
Population: The complete group of people or items about which the researcher wants information — every one of them, not just the ones studied. For a study of pressure ulcers in one hospital, the population is every patient admitted to that hospital during the study year.
Sample: The smaller group actually selected from the population and measured by the researcher. A sample of 200 nursing notes drawn from the 4,800 notes written in one month.
Parameter: A numerical characteristic of the population. It has one fixed value, but that value is usually unknown and written with a Greek letter. The true mean age μ of all diabetic patients registered at a clinic is a parameter; nobody knows it exactly.
Statistic: The corresponding numerical value calculated from the sample. It is known with certainty, but it changes from one sample to the next and is written with a Latin letter. The mean age x̄ = 54 years of the 120 patients actually measured is a statistic.

One idea at a timeIdea 1 / 5

In one sentence The population is the complete group the researcher wants information about; the sample is the smaller group actually studied.

A population is the whole group; a sample is the part studied

In biostatistics, population does not mean only people. It means the complete collection of individuals or records about which you want to draw a conclusion: every patient admitted to a hospital in a year, every blood pressure reading taken on one ward, every vial of a batch of vaccine. The sample is the smaller group selected from that population and actually measured. The notes put the contrast in one line each: the population is the complete group about which the researcher wants information, and the sample is a smaller group selected from the population for study. Everything in unit 6 depends on keeping both in view at once. The sample is what you can touch and count; the population is what you want to say something about. A nurse who confuses the two will report a fact about 500 people as though it were a fact about a city.

For example

A hospital wants to audit hand hygiene. The population is all 12,000 patients discharged in the year; the sample is 600 randomly selected records. The finding describes the 600 with certainty and the 12,000 only approximately.

Pause and check yourself

A study of needle-stick injuries surveys every nurse who works in a 90-bed hospital. Is that group a population, a sample, or both?

Show the answer

It is the population of that hospital, because no nurse was left out and no smaller group was selected. If the researcher had taken 40 of those nurses, those 40 would be a sample drawn from the same population. The word depends on the researcher’s intention and on whether anyone was left out.

Read detailed explanations

The notation convention in full, and why nurses meet it on every ward: The Greek-for-population, Latin-for-sample convention is not a cosmetic habit; it is a compression that lets a reader see the structure of a study in a single symbol. Those four pairs were set out in the idea above; here the point is what the convention buys. A caret adds one more layer of meaning: it marks a value obtained by estimation from data rather than measured on the whole group, so x̂, p̂ and even ŝ appear when a quantity has been inferred rather than observed. Two adjacent symbols carry almost the whole logic of a report: when an author writes x̄ ± 1.96 s/√n, the Latin letters announce that everything inside the expression was computed from the sample, the square root of n shows how the sample size enters, and the result is an interval around a sample value. Reading that formula as “a sample mean, widened by an amount that depends on the sample size”, rather than as an incantation, is the beginning of statistical literacy. Nursing practice meets the convention constantly. A journal club article states the proportion of catheter-associated infections and marks it p̂. A quality dashboard quotes the hospital-wide caesarean rate as a parameter because the hospital’s own data are complete for that metric. An infection-control nurse who understands that P is unknowable while p̂ is measurable will phrase her report honestly, and an audit committee will not mistake a sample result for a system-wide truth. There is one frequent point of confusion worth naming: the letter p alone, without a caret or a Greek counterpart, is used in hypothesis testing to mean the p-value, an entirely different quantity from a proportion. Context decides, and the caret is the reliable signal that a proportion is meant. Finally, sample size has its own convention, capital N for a population and lower-case n for a sample, so σ² = Σ(x − μ)² / N uses N because the denominator counts the whole population, while s² = Σ(x − x̄)² / (n − 1) uses n and a correction that exists only because a sample was used. Two alphabets, one caret, one capital letter: between them they encode who was measured and who was only spoken about.
What a census would actually cost, and when it is the right choice: The three reasons for sampling — cost, time and necessity — become convincing when you put numbers on them. Suppose a region wants the true mean systolic blood pressure of its two million adults. A single standardised measurement, with consent, seating, rest and recording, takes about fifteen minutes of trained staff time, and each participant costs roughly 8 currency units in consumables, travel and data entry. Two million measurements therefore demand about 500,000 hours of work, which is close to 240 person-years of full-time effort, and 2,000,000 × 8, about 16 million in direct cost. At that pace the survey would take a decade, during which the population would age and migrate, so the parameter being chased would have moved before it was finally measured. Now compare the sampling alternative. A well-drawn sample of 1,000 adults, measured in two months by four nurses, gives an estimate of the population mean with a margin of error of about 1 mmHg, taking a standard deviation of about 15 mmHg as the course does elsewhere, because the precision of an estimate depends mainly on the size of the sample and hardly at all on the size of the population. That last point surprises beginners and is worth stating plainly: for a sample of a given size, the precision of the estimate is almost the same whether the population has 20,000 members or 20 million, provided the sample was drawn without bias. The population size does not need to be counted accurately to interpret the result. Sampling is nevertheless not always right. A census is the correct choice when the population is small and the answer must be exact rather than approximate, or when the very act of counting carries its own duty. A ward counting its own nursing staff for a rota needs every name, not an estimate. A national death register counts every death, because each one matters legally and individually, and a hospital antibiotic audit may review all prescriptions in a month precisely because the aim is to change every prescriber’s practice, not to publish an average. The rule of thumb is therefore: use a sample when you need a generalisable estimate at tolerable cost, and use a census when the number itself is the deliverable, when the group is small enough to complete, or when leaving anyone out would be unjust — and never use a census when the measurement destroys what it measures.
Estimation in practice: the point estimate, the margin of error, and what sampling cannot repair: Estimation has a standard shape in practice, and it is worth learning now because unit 6 builds on it. The single number computed from the sample — x̄ = 122 mmHg, or p̂ = 0.46 — is called the point estimate, because it points to one value. On its own it hides uncertainty, so it is normally reported with a margin of error, which is the quantity added and subtracted to produce an interval estimate such as 122 ± 3 mmHg. That margin is computed from the variability of the data and the size of the sample, and it shrinks as the sample grows, but slowly: for a mean, the standard error is s ÷ √n, so quadrupling the sample only halves the margin. A researcher who wants to halve the uncertainty must pay four times as much, and a beginner who expects a bigger sample to remove uncertainty altogether will be disappointed. For a proportion the same logic gives a standard error of about √(p̂(1 − p̂) ÷ n); with p̂ = 0.5 and n = 100 the standard error is √(0.25 ÷ 100) = 0.05, giving a margin of about 10 percentage points, whereas n = 400 halves it to 0.025 and a margin of about 5 points. Those numbers are the arithmetic behind the phrase “margin of error of three points” in every election poll. The deeper distinction, and the one that matters most on a ward, is between sampling error and bias. Sampling error is the ordinary wobble caused by having measured only part of the population; it is unavoidable, it is quantified by the margin of error, and it shrinks as the sample grows. Bias is a systematic distortion in one direction, caused by who was selected or who answered; it does not shrink as the sample grows, and a million badly chosen records can be worse than a hundred well chosen ones. A sampling error can therefore be repaired by more data, while bias cannot be repaired by size at all. This is why a nurse reading a staff survey should ask two separate questions: how many responded, which sets the wobble, and whether the people who did not respond are systematically different, which sets the bias. A satisfaction survey answered only by day-shift staff who were on duty the day the form was circulated may show 90 % satisfaction with a tight margin of error and still be badly wrong about the night shift. Recognising that the second problem is not a sample-size problem is the practical payoff of this section’s vocabulary.

Explore the diagram

Look at the whole drawing first. Then select one numbered part and connect it to the idea you just read.

Population, sample, parameter and statistic

Three panels. The top panel shows a population of 12,000 discharges with a sample of 600 records drawn from it. The middle panel shows the estimation link: the sample statistic p̂ estimates the unknown population parameter π. The bottom panel gives the notation convention, pairing each Greek population symbol with its Latin sample symbol and stating the unit of each quantity.

New to diagrams? Read this first

Look at the whole drawing once without clicking anything. Ask yourself only: what kind of picture is this (a comparison, a route, a sequence, a table)?
Press Play, or use the arrows, to move through the numbered parts in order. The drawing stays still; only the highlight moves.
Click a number, or its name in the list below, to read what that part is and why it matters. Use “Hide labels” when you want to test your memory.

More controls

100%

01 / 3Population and sample are different groups

If the drawing is wider than this panel, scroll it sideways. Labels stay readable; the numbered explanations below show the same ideas.

Population and sample are different groups

The outer population is every patient discharged from the 12 regional hospitals in one year: 12,000 discharges, and the study wants to say something about all of them. The inner blue box is the sample of 600 files actually drawn and read. The 600 can be counted exactly; the 12,000 cannot, which is the whole reason a sample was taken. Confusing the two turns a fact about 600 files into a false claim about 12,000 discharges.

Schematic teaching plate. The sample box is drawn small on purpose so the size gap between 12,000 discharges and 600 files is visible; it is not an area-to-scale drawing, and the numbers are the worked figures used in the section.

Key points to remember

The population is the complete group the researcher wants information about; the sample is the smaller group actually studied.
A parameter is a numerical characteristic of the population, fixed but usually unknown; a statistic is the corresponding value computed from the sample.
Greek letters denote population parameters (μ, σ, σ², π); Latin letters and the caret denote sample statistics (x̄, s, s², p̂).
Estimation is using a sample to make an approximate statement about an unknown population value; the estimate is never presented as exact.
Studying the whole population is usually too expensive, too slow, impossible, or unnecessary because the measurement would destroy or harm what is measured.

Go deeper

Worked example

Naming the four parts of a published infection-control study. A regional infection-control report states: “Across the 8 hospitals of the region, 1,200 patient observations were reviewed and hand hygiene was correct in 82 % of them. The regional compliance rate is estimated at 82 %.” Before believing or repeating that sentence, a nurse must name its population, its sample, its parameter and its statistic, and must say which symbol belongs to which.

Find the population from the purpose of the study, not from the number measured. The report speaks about the region, so the population is every hand-hygiene opportunity in the 8 regional hospitals during the study period — an enormous number nobody counted.
Find the sample. It is the 1,200 patient observations that were actually reviewed. That is the group the researchers could see and count in full.
Classify each quoted number. The 82 % is computed from the 1,200 observations, so it is a statistic: a sample proportion, written p̂ in our notation, because it measures a proportion and it comes from a Latin letter.
Name the parameter the authors are chasing. It is the true regional compliance proportion across all opportunities in the 8 hospitals, written π. It is a single fixed value, and it stays unknown unless every opportunity is observed.
Read the sentence again with the four words in place. “The regional compliance rate is estimated at 82 %” means: the statistic p̂ = 0.82 is used as an estimate of the parameter π. The sentence is honest only because it says “estimated”, and a careful reader should expect a margin around that 82 %.

Population: all hand-hygiene opportunities in the 8 hospitals. Sample: the 1,200 observations. Statistic: p̂ = 0.82. Parameter: the true regional proportion π. Check the arithmetic of the statistic before accepting it: 82 % of 1,200 is 984 correct observations, and 984 ÷ 1,200 = 0.82 exactly, so the reported figure is internally consistent.

Tutorial 1 — reading a study description for the four parts

A regional health authority wants to know how many patients return to hospital within 30 days of discharge. Counting every discharge in the region would mean reviewing 12,000 patient files, so the team draws 600 files at random from the 12 hospitals. In those 600 files, 276 patients had returned within 30 days, which is 46 %. A newspaper headline announces: “41 % of patients return within 30 days.” Identify the population, the sample, the parameter and the statistic, and decide which of the two numbers the headline should have used.

Rule used: Population = the complete group the researcher wants information about. Sample = the smaller group actually studied. Parameter = a numerical characteristic of the population, fixed but usually unknown, written with a Greek letter. Statistic = the value computed from the sample, known exactly, written with a Latin letter. A statistic is used as an estimate of a parameter.

Write down the population in one sentence, using the intention of the study rather than the size of the number measured.
Reveal this step
The population is every patient discharged from the 12 hospitals of the region during the study period — all 12,000 discharges, whether or not their file was opened. The number 12,000 is the size of the population only because the team counted the files available; the population is the discharges themselves.
The sample is the 600 files. Why is it wrong to call those 600 files “the population of the study”?
Reveal the answer
Because 11,400 discharges were deliberately left out, so the group studied is not complete. The researcher measured a part in order to speak about the whole. Calling the part the population erases the very gap the study has to bridge, and it hides the fact that a different set of 600 files would have given a different percentage.
Compute the sample proportion and write it with its symbol.
Reveal this step
The sample proportion is 276 ÷ 600 = 0.46, that is 46 %. Written with the convention, p̂ = 0.46: the “hat” marks an estimated value computed from the sample, and the Latin letter p marks a proportion.
The headline says 41 %. Is that number a parameter or a statistic, and what does that make the 46 %?
Reveal the answer
The 41 % is nothing but the newspaper’s unsupported claim about the parameter; it is not a measurement, and the true regional return proportion π remains unknown. The 46 % is the statistic, p̂, computed on the 600 sampled files, and it is the only number here that came from data. Reporting a statistic as though it were the parameter — as the headline does — is the error the tutorial is there to expose.
State the study’s estimate in one correct sentence that names the statistic and the parameter.
Reveal this step
The 600 sampled files give a return proportion of 0.46; this statistic is used to estimate the unknown population return proportion, so the honest sentence is “the regional 30-day return rate is estimated at about 46 %”.
Why was it impossible in practice to measure the parameter, and what would have happened to the study if the team had insisted on it?
Reveal the answer
Reviewing all 12,000 files would have needed far more staff time, money and months of work, and some files would be incomplete or the patients untraceable. The result would arrive late and still contain missing data, so a complete census would have been slower and more expensive without being perfectly accurate.

Answer: Population: all 12,000 discharges in the region’s 12 hospitals. Sample: the 600 files drawn at random. Statistic: p̂ = 0.46 (276 ÷ 600). Parameter: the true regional proportion π, unknown, which the newspaper appears to place at 41 %.

Keep this habit: Every study description contains these four elements, and they must be named in order: intention gives the population, method gives the sample, arithmetic gives the statistic and the target gives the parameter. Before you repeat any health number, ask which of the four you are holding — a fact about the group measured, or a claim about the group meant.

Tutorial 2 — matching each Greek and Latin symbol to its quantity

A nursing student is reading a journal article for an assignment and meets eight symbols in two paragraphs: μ, x̄, σ², s², σ, s, P and p̂. The student can pronounce none of them and cannot tell which ones describe the study sample and which ones describe the whole population. Sort all eight symbols into the two families, attach each one to the quantity it denotes, and decide what the article is claiming when it writes x̄ = 4.2 days.

Rule used: Alphabet rule: a Greek letter (μ, σ, σ², π) denotes a population parameter — a fixed but usually unknown number. A Latin letter (x̄, s, s², p̂) denotes a sample statistic, computed from the data in hand. A caret is an extra mark meaning “estimated from a sample”. Quantity rule: mean, proportion, variance and standard deviation each have one Greek and one Latin symbol; variance is in squared units (days²), while the mean and the standard deviation stay in the original unit (days).

Sort the eight symbols into the Greek family and the Latin family, naming the letter in each case.
Reveal this step
Greek family: μ (mu), σ (sigma), σ² (sigma squared) and π (pi). Latin family: x̄ (x bar), s, s² (s squared) and p̂ (p hat). The alphabet alone tells you whether the article is talking about a population value or a value measured on the sample.
Pair the mean symbols with their quantities and explain why x̄ is readable and μ is not.
Reveal the answer
μ is the population mean and x̄ is the sample mean. x̄ can be calculated at once, because the sample is in front of the researcher; μ cannot be calculated, because the whole population was never measured. The article writes x̄ = 4.2 days meaning “in the patients we studied, the mean length of stay was 4.2 days”.
Pair the standard deviation symbols, then the variance symbols, and check the units of each.
Reveal this step
σ is the population standard deviation and s is the sample standard deviation; both carry the unit of the data, here days. σ² is the population variance and s² the sample variance; both carry squared units, days². A value in days² cannot be compared with a value in days, which is why the standard deviation is quoted far more often.
Now the proportion pair. What does the caret add to p̂, and why does the population proportion carry the Greek letter P rather than a Latin p?
Reveal the answer
The caret marks a value estimated from the sample, so p̂ is the proportion observed among the people actually measured, and it is known exactly. π denotes the population proportion, a fixed value for everyone concerned that no one has measured. It shares the caret’s family, not its Latin letters: π is the parameter and p̂ the statistic. The pair follows the same alphabet rule as the mean.
Decide what the article claims when it writes x̄ = 4.2 days, and say which symbol its claim is aimed at.
Reveal this step
The claim is about the patients actually studied: their mean stay was 4.2 days. The number is offered as an estimate of μ, the unknown mean stay for all patients in the hospital, so the honest reading is “all patients stay about 4.2 days on average, as far as this sample shows”.

Answer: Greek for the population: μ (mean), σ (standard deviation), σ² (variance), π (proportion). Latin for the sample: x̄ (mean), s (standard deviation), s² (variance), p̂ (proportion). The article’s x̄ = 4.2 days describes the sample and estimates the unknown μ.

Keep this habit: Read the letter before you read the number. A Greek letter means the value belongs to the population and is usually a target, not a measurement; a Latin letter or a caret means the value was computed from the sample in hand. Then check the unit: mean, standard deviation and proportion stay in the original unit, while variance is always in squared units.

Now you try

Attempt each one on paper first. Write the formula, substitute the numbers, then check the unit of your answer.

A clinic audits the blood pressure of its patients. The nurse selects 150 files at random from the 2,400 files in the clinic’s register and finds a mean systolic pressure of 131 mmHg. Name the population, the sample, the parameter and the statistic, with the correct symbol for the statistic.
Hint
Ask what the clinic wants to say something about, then ask what it actually looked at. The symbol depends on the alphabet: Greek for the population, Latin for the sample.
Show the worked answer
Population: all 2,400 patients in the clinic’s register. Sample: the 150 files drawn at random. Parameter: the true mean systolic pressure μ of all 2,400 patients, a fixed but unknown value. Statistic: the sample mean x̄ = 131 mmHg, computed from the 150 files, and used as the estimate of μ.
In one article you read “s = 12.4 mL” and, three lines later, “σ = 12.4 mL”. Both numbers are identical. Explain what each symbol is describing, which of the two a researcher can defend with data, and which one is a claim about people who were never measured.
Hint
The size of a number says nothing about where it came from; the letter does. Look at who was measured in each case.
Show the worked answer
s = 12.4 mL is the sample standard deviation: it is computed from the measurements the researchers actually made, so it can be checked against the data and defended exactly. σ = 12.4 mL denotes the population standard deviation, the spread among every patient concerned, including those never measured and those who refused; nobody can compute it unless the whole population is measured. Reporting the same value for both is acceptable only as an estimate of σ from s, never as a measurement of σ.
A student writes: “The population proportion is 0.35. I know this because p̂ = 0.35 in the study.” Explain the mistake and rewrite the sentence correctly, using the right symbols.
Hint
A caret is not decoration: it announces where the number came from. Which alphabet belongs to a value nobody measured?
Show the worked answer
The student treats an estimated value as though it were the truth. p̂ = 0.35 is a statistic computed on the sample, so it is known exactly for that sample and only estimates the population proportion π, which remains unknown. The correct sentence is: “The sample proportion is p̂ = 0.35; this estimates the population proportion π, which has not been measured.”
A laboratory receives a batch of 20,000 vaccine vials. Testing one vial for sterility destroys it. The quality officer tests 250 vials and all pass. State the population, the sample and the statistic, and explain why a complete census is not merely expensive here but logically impossible to use.
Hint
Think about what would be left of the batch if every vial were tested. The population can be judged only from a part that survives.
Show the worked answer
Population: all 20,000 vials in the batch, and more precisely their sterility. Sample: the 250 vials actually tested. Statistic: p̂ = 250 ÷ 250 = 1.0, that is 100 % sterile in the sample. A census would require destroying all 20,000 vials, leaving no vaccine to distribute, so the true population parameter can never be measured and must be estimated from the surviving sample.

Common misconception

Beginners often treat the sample value as though it were the population value with a smaller sample size — “we measured 46 %, so the region is 46 %”. That is the error the notation exists to prevent. A statistic is a fact about the people you measured; the parameter is a fixed but unknown fact about the people you did not measure. The two are linked by estimation, which is always approximate and always carries uncertainty. Repeating a statistic as though it were a parameter is exactly how a survey of 600 files becomes a false claim about 12,000 patients on a newspaper front page.

Check your reasoning

You are asked to explain to a patient why a hospital reports “our infection rate is 6.2 %” although only 500 of the year’s 9,000 operations were reviewed. Write the explanation in plain language, using the words sample, statistic and parameter, and say what the patient should and should not conclude.

Reveal an answer

The 500 reviewed operations are a sample drawn from the population of all 9,000 operations of the year. The 6.2 % is a statistic: it describes those 500 operations exactly as measured. The hospital’s true infection rate for the whole year is a parameter, written P, and it remains unknown because 8,500 operations were not reviewed. The patient can conclude that the hospital is actively measuring its infection rate and that about 6 % is a reasonable estimate. The patient cannot conclude that exactly 6.2 % of all 9,000 operations became infected, and should ask how much the estimate could vary if another 500 operations had been drawn.

Where this comes from

BMS 1110 Applied Mathematics and Biostatistics — supplied lecture notes, Unit 6 biostatistics block · Lines 1029–1043: estimation defined as using a sample to make an approximate statement about an unknown population characteristic (1031); the population parameter usually unknown because studying every individual is impossible or impractical (1032); the city blood-pressure example with a sample of 500 (1033–1034); population as the complete group about which the researcher wants information and sample as a smaller group selected from it (1036–1037); the hospitals example of all patients versus 500 selected (1039–1040); the notation table giving μ/x̄, p/p̂, σ²/s² and σ/s (1042). The notes' table writes p for the population proportion, a letter later reused in the same unit for the p-value. This course therefore writes P for the population proportion and p for the p-value; the four parameter/statistic pairs themselves are as the notes give them.
Course architecture decisions for BMS 1110 · the notation decision to write P for a population proportion and p for a p-value, because the notes reuse p for both; and the course code BMS 1110, which the curriculum catalogue records while the supplied docx cover reads 'BMS 110'.
Standard biostatistics practice for the deeper explanations · the caret as the general mark of an estimated quantity (x̂, p̂, ŝ); capital N for a population size against lower-case n for a sample size; the point estimate, margin of error and interval estimate (x̄ ± 1.96 s/√n); the standard errors s ÷ √n and √(p̂(1 − p̂) ÷ n) with the worked values 0.05 at n = 100 and 0.025 at n = 400; the cost and time arithmetic of a two-million-person census; the distinction between sampling error, which shrinks with sample size, and bias, which does not.

Lesson progress is saved on this device for the current learner. Practice answers stay separate and are not stored.

About this course and its sources

Bilingual lessons organised from the supplied BMS 1110 Applied Mathematics and Medical Biostatistics lecture notes (Mr. Ebangan Prosper Achuanoh). Diagrams are original educational schematics; sections that go beyond the supplied notes say so in their sources. Mathematics, clinical and French-language review are pending.