Home / Confusing words / parameter vs statisticNo. 0511

parameter vs statistic

A parameter describes a whole population; a statistic describes a sample. The average height of all adults in a country is a parameter — one fixed, true value, almost never actually measurable. The average height of the 1,000 adults you surveyed is a statistic — computable, and a little different with every new sample. Statistics are how we estimate parameters we cannot reach.

Quick rule
describes the whole population, fixed but usually unmeasurableparameter
computed from a sample, varies from sample to samplestatistic
parameter

Fig. 1aparameter.

/pəˈræmɪtər//pəˈræmɪtə/·noun
vs
statistic

Fig. 1bstatistic.

/stəˈtɪstɪk//stəˈtɪstɪk/·noun

In everyday English a parameter is a limit — the edges within which something may vary, like the stops on the mixing desk in the scene above. Statistics borrowed the word for the fixed truths of a whole population: the real mean, the real proportion — values that exist whether or not anyone can afford to measure every member. A statistic is the working substitute: compute the same quantity on a sample and use it to estimate the parameter. That is the entire relationship — one number is the target, the other is the shot at it — and the notation keeps them apart: Greek letters (μ, σ) for parameters, Latin (x̄, s) for statistics.

01What each means

parameter

A parameter comes from Greek para- (beside) and metron (measure): a measure set alongside, a limit you agree to work within. In the plural it usually means the boundaries of a task — 'within the parameters of this study' — the agreed scope no one should overstep. In its technical sense a parameter is a variable factor whose value you can set, like a constraint you choose rather than one imposed. Either way it defines a space: fix the parameters and you fix where the work may go.

statistic

A statistic is one number distilled from many — an average, a percentage, a count — that stands in for a whole mass of measurements. The word shares its root with 'state', since such figures were first gathered to describe nations: their populations, harvests, and taxes. A single statistic can clarify or mislead, which is why careful readers ask how it was collected before they let it settle an argument. The plural, statistics, also names the entire discipline of analysing such data.

02At a glance
parameterstatistic
Describesthe whole populationone sample
Valuefixed — one true numbervaries from sample to sample
Computable?rarely — you would need everyoneyes — that is its job
SymbolsGreek: μ, σ, πLatin: x̄, s, p̂
Rolethe target being estimatedthe estimate doing the work
Examplemean height of all adultsmean height of 1,000 surveyed
03How to remember

The stops and the summary. A parameter is the pair of stops on the desk: a fixed fact about the whole system, set whether or not you ever look at it. A statistic is the glowing number in the corner of the chart: computed from the readings you actually collected, and different the next time you collect. Ask one question — was every member measured, or just a sample? Whole population: parameter. Sample: statistic.

04In use

parameter

  • iThe true unemployment rate for the whole country is a parameter no survey ever sees directly.
  • iiMean body temperature across all living adults is a parameter; no thermometer will ever reach it.
  • iiiA census aims at parameters: it tries to measure everyone, not a sample.

statistic

  • iThe poll's 43% approval figure is a statistic drawn from 1,200 voters.
  • iiEach new sample of bulbs gives a slightly different failure-rate statistic.
  • iiiThe clinic reports a sample statistic: average recovery time among its own patients.

Two housekeeping notes keep the pair tidy. First, everyday English uses parameter loosely for any limit or factor ('within the parameters of the budget') — that sense, the one this site's parameter page animates, is separate from the statistical one, and in essays it is almost always the one you want. Second, the pair travels with two companions: population and sample. Hold all four together and the topic is essentially held — which is why the one-sentence summary in the first question below is worth memorising outright.

05TOEFL & IELTS

The stats sense surfaces in TOEFL reading passages on research methods: when a passage says a statistic was computed, it is flagging sample-based evidence, with all the uncertainty that implies — a favourite setup for inference questions. In your own writing, keep the register straight: essay 'parameters' are scopes and limits, not population values, and IELTS Task 1 numbers are statistics only in the loose sense of figures — no sampling theory required.

06Asked
What is the difference between a parameter and a statistic?
One sentence holds it: a statistic from a sample estimates a parameter of a population. The parameter is the true, fixed value across everyone — every voter, every bulb, every fish in the lake — and is almost never directly measurable. The statistic is the same quantity computed on the members you actually measured. The number you have is virtually always the statistic; the number you want is the parameter.
How do you know whether a number is a parameter or a statistic?
Ask two questions. Does the number claim to describe every member of the group? And could its makers really have measured them all? 'Sixty per cent of this school's 400 students walk to school' can be a parameter — a school can count everyone. 'Sixty per cent of the country's students walk' must be a statistic: nobody asked them all, so the number came from a sample and merely estimates the truth.
Is a parameter fixed while a statistic changes?
Yes — and the scenes above stage the contrast. The parameter is the pair of stops on the fader: one setting, holding whether or not anyone measures it. The statistic is the number distilled from whichever readings you happened to gather — survey a fresh thousand people and it shifts a little every time. Statisticians call that shifting sampling variability, and margins of error exist precisely to say how far a statistic may plausibly sit from its parameter.
What symbols are used for parameters and statistics?
Two alphabets, deliberately: Greek for parameters, Latin for statistics. The population mean is μ, its standard deviation σ; the sample versions are x̄ ('x-bar') and s. Proportions are the standing exception — textbooks write the population proportion as π or plain p, and the sample proportion as p̂ ('p-hat'). The hat is the reliable giveaway: it marks an estimate, so a hatted symbol is always the statistic.
Is a percentage a parameter or a statistic?
Either — the percentage sign carries no answer. 'Fifty-two per cent of this class passed' is a parameter if the whole class was counted. 'Fifty-two per cent of viewers approved' is a statistic if a ratings panel stood in for the audience. The form of the number never decides; the reach of the measurement does. That is the trap exam questions set with percentages, and the two-question test above disarms it.
How do you remember which is which?
The alliteration is standard for a reason: Population Parameter, Sample Statistic — the first letters pair off. If you want a second anchor, count the alphabets in: Greek letters for the grand truth, Latin letters for the local estimate. Between the initial-letter pairing and the two alphabets, most students never mix them again.
Related