Statistics is the science and practice of learning from data while accounting for variation, uncertainty, context, and the way the data were produced. The word also has a second, narrower meaning: a statistic is a numerical value calculated from sample data. These two meanings are related but not identical. A course in statistics teaches a way of reasoning; a statistic is one output within that reasoning.

Introducing Statistics should change how a reader reacts to claims. Instead of asking only, “What number was reported?” a statistically minded reader asks, “What was the question? Who or what did the data describe? How were observations selected? What exactly was measured? How much did the observations differ? Does the reported number describe the sample or claim something about a larger population?” These questions prevent a polished calculation from being mistaken for strong evidence.

Statistics is often divided into two broad activities. Descriptive statistics organize and summarize the data that were actually observed. Tables, graphs, percentages, averages, and measures of spread can all describe a sample or a complete population. Statistical inference uses sample data to learn about a larger population, while acknowledging that another sample could have produced a different result. Topic 1.1 introduces the language required for both activities, even though formal inference is developed later in the course.

Statistics is more than arithmetic

Arithmetic can tell us that the mean of 4, 6, and 8 is 6. Statistics asks whether those three values came from the people or objects relevant to the question, whether the measurement was meaningful, whether 6 is an appropriate summary, how the values varied around 6, and whether the result can be generalized. A correct calculation can coexist with a poor study. Conversely, a thoughtfully designed study may produce a result with unavoidable uncertainty. Statistics makes that uncertainty visible rather than pretending it does not exist.

For example, suppose a student surveys five close friends and finds that all five prefer online homework. The sample proportion is 1.00, or 100 percent. The arithmetic is correct. However, if the intended population is all 1,200 students in the school, the selection process is unlikely to represent the population. The result is a statistic describing those five friends. It is not automatically a trustworthy estimate of the population parameter.

Statistics is a complete investigation cycle

In practice, these stages interact. A preliminary look at data may reveal that a variable was recorded ambiguously. A planned analysis may show that the question needs a clearer population. A conclusion may lead to a better follow-up question. Introducing Statistics teaches the vocabulary needed to move around this cycle without confusing one component with another.

A statistical study can be read as a set of connected components. If one component is missing, the meaning of the study becomes uncertain. Introducing Statistics uses the following map throughout the chapter.

A miniature study

A public library wants to know the typical number of minutes adult visitors spend in the building on weekday afternoons. During randomly selected weekday afternoon periods, staff record the visit length of 120 adult visitors. The sample mean is 46.3 minutes.

Notice that 46.3 is not simply “the average visit length.” A precise statement says it is the sample mean visit length for the 120 observed adult visits. Whether it estimates the population mean well depends on how the sample was selected and how visit length was recorded. Introducing Statistics trains this precision from the beginning.

Context is not decoration

In statistics, “in context” means matching every abstract term and calculation to the real-world component it represents. If a calculation produces 0.37, a contextual interpretation might be “37 percent of the sampled households reported using public transportation at least once during the previous week.” Merely writing “the proportion is 0.37” omits the group, behavior, and time frame.

Context also prevents unit errors. A mean of 12.4 could mean dollars, hours, points, kilograms, or customer complaints. A standard deviation of 3 has the same units as the variable it describes. A percentage must specify the denominator. A sample size must refer to the number of observational units, not automatically the number of measurements if repeated observations are present.

The population is the complete collection of items or individuals relevant to an investigative question. The symbol N commonly represents population size. The sample is the subset from which data are actually obtained, and n commonly represents sample size. A sample is not defined merely by being small; it is defined by being a part of a larger population of interest.

Defining the population precisely

A useful population definition usually includes more than a broad label. “Students” is rarely precise enough. Which students? At which schools? During what academic year? Enrolled at what point in time? Eligible under which conditions? The population might be “all students enrolled in grades 9–12 at North Valley High School on October 1, 2026.” That definition creates a boundary. A student who transferred after October 1 may not belong to the defined population even though the student later attends the school.

Populations can contain people, but they do not have to. A manufacturer may study all light bulbs produced by a machine during a week. An environmental scientist may study all daily air-quality readings in a city during a year. A streaming service may study all viewing sessions started by U.S. subscribers during a month. A botanist may study all trees of a species in a protected forest. Introducing Statistics uses “items or individuals” because the unit depends on the question.

Why use a sample?

A census attempts to collect data from every member of the population. A census can be appropriate when the population is small, every member can be reached, and the required measurements are practical. However, a census can be expensive, slow, or impossible. Some measurements destroy the item being tested. A company cannot measure the breaking strength of every helmet it intends to sell because the test would leave no helmets to sell. Some populations are continuously changing, such as all future customers or all future production units.

Sampling can provide timely information with fewer resources. Yet a sample introduces a central statistical issue: different possible samples can produce different results. That is sampling variability. A sample of 100 residents may report a different average commute time from another sample of 100 residents, even if both are selected carefully from the same city.

Target population and accessible population

The target population is the group the researcher ultimately wants to understand. The accessible population is the portion that can realistically be reached using the available list, place, period, or system. Suppose a university wants to learn about all graduates from the past ten years but only has current email addresses for 60 percent of them. The target population is all graduates in the defined period. The accessible population is the set with usable contact information. The difference can matter because reachable and unreachable graduates may differ.

Population and sample can change with the question

The same data source can support different populations if the investigative question changes. A data set of 500 restaurant orders collected on Fridays could be a sample from all Friday orders at one restaurant, or it might be the complete population of Friday orders during the specific data-collection month. Whether the records form a sample or a population depends on the scope of the question, not on the number of rows alone.

An individual is an object, person, case, or other entity described by the data. The more general term observational unit is useful when the row is not literally a person. In a conventional rectangular data table, each row usually represents one observational unit and each column represents one variable. The word “usually” matters because repeated-measures and hierarchical data can use more complex structures.

The unit must match the question

Suppose a study records the number of emergency-department visits at 40 hospitals for each of 365 days. Is the observational unit a hospital, a day, or a hospital-day combination? The answer depends on the table. If each row is one hospital on one date, the observational unit is a hospital-day. There are up to 40 × 365 = 14,600 rows. If the data are first summarized so that each row contains a hospital’s annual average, the observational unit in the summarized data set is a hospital.

Misidentifying the unit can lead to a false sample size. A study with 50 students measured on 10 occasions has 500 measurements but only 50 student participants. Those measurements are not necessarily 500 independent individuals. Introducing Statistics does not yet develop dependence models, but it establishes the habit of asking what one row or record represents.

Individual, subject, participant, case, and record

Different fields use different words. A subject or participant is usually a person involved in a study. A case may be a person, organization, event, or object. A record is the stored representation of an observational unit. The terms overlap, but the safest practice is to define the unit explicitly: “Each row represents one household interviewed once” or “Each row represents one shipment.”

Units can be nested

Students belong to classrooms, classrooms belong to schools, and schools belong to districts. Employees belong to teams and companies. Measurements may occur within patients over time. At Topic 1.1, the main goal is recognition: a study may contain multiple levels, and the population or sample statement should specify the level to which the question refers. A question about average school size has schools as the units, even though student counts appear in the data.

A datum is one recorded piece of information. Data is the plural form, although in everyday writing the word is often treated as a mass noun. A data set is an organized collection of data. Organization includes labels, units, definitions, and enough structure to connect values to individuals and variables.

Consider the row below from a study of school travel. The student identifier is ST104. The student’s travel mode is bus, travel time is 27 minutes, distance is 8.4 miles, and arrival status is on time. Each cell contains a datum. The complete table of many students and columns is the data set.

Data values need variable definitions

The value 27 is not fully meaningful until the variable is defined. Does travel time begin when the student leaves home or when the bus arrives? Does it end at the school gate, classroom door, or official start time? Are waiting minutes included? Clear operational definitions make values comparable. Without them, two observers may record different numbers for the same journey.

A variable is a characteristic that can take different values across individuals or occasions. Topic 1.2 develops variable types in detail, but Topic 1.1 requires enough understanding to identify what was recorded. In the travel data, travel mode and arrival status are categorical variables; travel time and distance are quantitative variables. The important first step is to name the variable in context rather than simply pointing to a column.

A data set is not just numbers

Names, labels, categories, dates, yes/no responses, text codes, and missing-value indicators can all be part of a data set. Statistical analysis depends on the meanings attached to these entries. A code of 1 might mean “yes,” “male,” “treatment group,” or “first visit.” A data dictionary should explain the code. If a data set uses 999 to mean “not reported,” treating 999 as a genuine measurement could severely distort an average.

Raw data and derived data

Raw data are values as initially recorded, although even “raw” data may already reflect measurement choices. Derived data are created from other variables. Age on an interview date can be derived from date of birth. Body mass index can be derived from height and weight. A pass/fail indicator can be derived from a score and a stated cutoff. Derived values can be useful, but the transformation should be documented.

Data quality begins before analysis

Introducing Statistics encourages basic questions about data quality: Are the intended individuals included? Are there duplicate rows? Are units consistent? Are impossible values present? Are missing values clearly marked? Were definitions applied in the same way? These questions are not advanced software tasks. They are part of understanding what the data set actually says.

A parameter is a numerical characteristic of a population. A statistic is a numerical characteristic calculated from a sample. Parameters and statistics often describe the same kind of feature—such as a mean or proportion—but they refer to different groups.

Example: screen time

A school district wants to know the mean weekday recreational screen time of all 10th-grade students in the district. The population mean, μ, is the parameter. It is fixed for the defined population and measurement rules, although it may be unknown. The district surveys a sample of 180 students and obtains a sample mean of x̄ = 3.6 hours. The value 3.6 hours is a statistic.

The sample statistic is used as an estimate of the population parameter. Another properly selected sample might yield 3.4 or 3.8 hours. This does not mean the population parameter changes every time a sample is drawn. It means the statistic varies from sample to sample.

Parameter does not mean “important number”

Students sometimes label a value a parameter because it appears in the question or because it seems official. The correct distinction is the group described. If 62 percent of the 250 surveyed customers were satisfied, 62 percent is a sample statistic. If company records include every customer in the defined population and show that 62 percent were satisfied, the value is a population parameter for that population.

A statistic can describe a complete observed population

In informal language, any numerical summary may be called a statistic. In introductory inference, however, the parameter-statistic distinction is tied to population and sample. When a school calculates the mean score of every student enrolled in a particular class, that mean is a parameter for the class population. The same class may be treated as a sample from a broader conceptual population, such as future students taught under the same method, but that broader claim requires careful definition.

Estimand, estimate, and estimator

Three related terms deepen the distinction. An estimand is the precise population quantity the study aims to learn, such as the population mean difference in sleep duration between two groups. An estimator is the rule or formula used to estimate it, such as the difference between two sample means. An estimate is the numerical result obtained from the observed data. Topic 1.1 does not require advanced estimator theory, but the language shows why a question must specify the desired population feature.

Variability means that observations are not all identical and that repeated data-collection processes can produce different results. Variation is the reason statistical questions require data. If every individual had exactly the same value and every measurement were perfectly repeatable, many statistical problems would collapse into simple facts.

Variation among individuals

Students differ in commute time, sleep duration, course grades, and preferred learning methods. Manufactured parts differ slightly in length or strength. Daily temperatures change. Customer orders vary in amount. A statistical distribution records not only a typical value but also the pattern of differences across individuals.

Variation in measurement

The same characteristic may be recorded differently because instruments have limited precision, respondents interpret questions differently, observers use slightly different judgments, or conditions change. If a person’s pulse is measured twice, the results may differ because the pulse itself changes and because the measurement process is imperfect.

Variation from sample to sample

Suppose 40 percent of all students at a school use the bus. One random sample of 50 students may contain 18 bus users, giving a sample proportion of 36 percent. Another sample may contain 23 bus users, giving 46 percent. Both statistics can arise from the same population. Later units quantify this sampling variability through sampling distributions and standard errors.

Variation does not automatically mean error

Natural differences are not mistakes. Two batteries from the same production process can have different lifetimes. Two students can respond differently to the same lesson. Statistical reasoning distinguishes natural variation from measurement error, selection bias, recording mistakes, and systematic changes. Introducing Statistics begins with recognition: whenever a question concerns a variable characteristic, expect a distribution of possible values rather than one universal answer.

Variability shapes the wording of conclusions

A sample average does not imply that every individual has that value. If the mean commute time is 24 minutes, some students may travel 5 minutes and others 60 minutes. If 72 percent of a sample supports a policy, the remaining 28 percent does not. Good statistical writing preserves this variation instead of turning a summary into a claim about every case.

An investigative question is a question that guides a statistical study. It identifies what the study seeks to learn and requires data from or about a group. A valid investigative question is answerable, connected to a population or process, and written so that the required variables can be identified.

The official Topic 1.1 skill is to determine a valid investigative question that requires a statistical investigation. That means recognizing more than punctuation. A sentence can end with a question mark and still fail as an investigative question. “Is exercise good?” is too vague. Good for whom, measured how, compared with what, and during what period? Introducing Statistics turns broad interests into questions that can guide evidence collection.

What a strong question reveals

• Population: the group or process about which information is wanted.
• Variable or variables: the characteristics that must be recorded.
• Purpose: description, comparison, association, prediction, or another clearly stated aim.
• Context: place, period, condition, or definition needed for a meaningful answer.
• Variability: an expectation that observations may differ.
• Feasibility: the needed data could reasonably be collected or obtained.

Five common families of investigative questions

Descriptive and inferential versions

“What percentage of the 80 surveyed students bring lunch from home?” is descriptive because it asks about the observed sample. “What proportion of all students at the school bring lunch from home?” is inferential if only a sample is observed. The second question concerns a population parameter and requires a defensible sampling process if the result is to be generalized.

Questions create variable requirements

Consider the question, “Is weekly study time associated with mathematics exam score among first-year college students?” The study needs at least two variables for each student: weekly study time and exam score. It also needs a definition of first-year student, a time period for study hours, and a specified exam or score. If the question compares online and in-person sections, course format becomes another required variable.

A statistical question anticipates variability and is answered by collecting, analyzing, or summarizing data. A nonstatistical question seeks a single fixed fact, definition, or calculation that does not require understanding variation across observations. The distinction depends on the intended question, not simply on whether a number appears in the answer.

Examples and nonexamples

A question can become statistical through repeated observations

“What is the temperature at noon today?” seeks one measurement. “How does the noon temperature vary across the 30 days of April?” is statistical. “What is the battery life of this device?” might seek one test result, while “What distribution of battery life is produced when this model is tested repeatedly under a standard protocol?” expects variability.

A statistical question is not automatically a good question

“What do people think?” anticipates variation, but the population and subject are undefined. “Is there a relationship between everything?” is impossible to operationalize. A valid statistical question must be sufficiently specific to identify data, units, and scope.

A strong investigative question can be built deliberately. Introducing Statistics uses the POP-V-C method: identify the POPulation, name the Variable or variables, and state the Comparison, connection, or contextual boundary.

Step 1: Start with a real purpose

Begin with the decision or understanding the study should support. A school may need to adjust bus schedules, a manufacturer may need to evaluate battery consistency, or a library may need to plan staffing. The purpose should not dictate the result, but it helps identify which information is relevant.

Step 2: Define the population

Replace broad words with an operational boundary. Instead of “teenagers,” write “students ages 14–18 enrolled at the three district high schools on September 15, 2026.” Instead of “customers,” write “customers who completed an online order from the company’s U.S. store during June 2027.”

Step 3: Name measurable variables

Replace abstract ideas with recordable characteristics. “Engagement” might become number of voluntary discussion posts, minutes active in the course platform, or a score from a defined survey. Different operational definitions answer different questions. A variable should include units or categories when necessary.

Step 4: State the intended statistical feature

Decide whether the question asks for a distribution, mean, median, proportion, difference, association, or trend. Topic 1.1 does not require choosing formulas, but a clear target helps the study collect the right data.

Step 5: Include context and time

Conditions can change. Commute time during the morning rush differs from late evening. Customer satisfaction may depend on the product version. A time period, location, or operating condition often belongs in the question.

Step 6: Check that variability is expected

If the answer is a fixed definition or a single known record, the question may not require statistical investigation. A valid statistical question should anticipate differences among individuals, occasions, or repeated samples.

Step 7: Remove loaded wording

“Why do irresponsible drivers speed?” assumes irresponsibility and that the behavior occurs. A more neutral question is, “What proportion of licensed drivers in the county report exceeding the posted limit by at least 10 miles per hour during the previous month?” Neutral wording does not guarantee unbiased measurement, but it avoids building a preferred conclusion into the question.

Question repair workshop

A self-check before data collection

• Can I identify the population without guessing?
• Can I identify each observational unit?
• Can I name the variable or variables and how they are recorded?
• Does the question anticipate variability?
• Is the time, place, or condition clear enough?
• Could data realistically answer the question?
• Does the wording avoid assuming the conclusion?
• Will the planned statistic or graph connect directly to the question?

The following original cases show how Introducing Statistics concepts fit together. Each case begins with a purpose, states an investigative question, and maps the population, sample, individuals, variables, parameter, statistic, and expected sources of variability.

The mean and median describe the sample, not every student. The difference between 31.8 and 27 suggests that longer commutes may pull the mean upward, but a graph would be needed to study the distribution properly. The sample selection method determines whether the statistics can credibly represent the population.

The test conditions are part of the variable definition. The result does not automatically describe battery life at every volume, with every device, or under intermittent use. Introducing Statistics emphasizes that a parameter belongs to a defined population under a defined measurement process.

A contextual interpretation is: “In the sample, 37 percent of the 400 selected registered teen members attended at least one summer program.” Calling 37 percent the population proportion would be unjustified unless every member were included or a later inferential argument supported the estimate.

This question is statistical because plant growth varies. It is comparative because it involves two treatment groups. Later lessons will determine what the design permits the class to conclude. Topic 1.1 focuses on identifying the components before analyzing the results.

The observational unit is a session because one subscriber can contribute multiple sessions. If the question concerned average monthly viewing per subscriber, the unit and data organization would need to change. Introducing Statistics always ties the unit to the question.

These exercises are original and designed specifically for this lesson. They emphasize identification before calculation. For each scenario, write answers in complete contextual phrases. Do not answer “population = everyone” unless the study truly concerns everyone; name the exact group.

Open each solution after attempting the exercise. The wording demonstrates how to keep the answer in context. Equivalent answers may be correct when they define the same population, sample, unit, and variable clearly.

Error 1: Calling the sample the population

A data set contains only the observed cases, so students often call those cases the population. Correct the error by returning to the question. If the study wants to describe all 8,000 district students but surveys 300, the 8,000 students form the population and the 300 surveyed students form the sample.

Error 2: Calling every reported number a parameter

A parameter describes the entire population. A sample mean, sample median, sample proportion, or sample standard deviation is a statistic. The size or authority of the organization reporting the number does not change this definition.

Error 3: Naming a topic instead of a variable

“Health” is a topic, not a sufficiently defined variable. Resting heart rate in beats per minute, number of sick days during the semester, or a score on a named health questionnaire are variables. Introducing Statistics expects a characteristic that can be recorded for each unit.

Error 4: Confusing individuals with variables

In a student survey, students are individuals; grade level and commute time are variables. In a table of schools, schools are individuals; enrollment and graduation rate are variables. A quick check is to ask: Does this label identify a row or a column?

Error 5: Treating repeated measurements as new individuals

Ten readings from one machine are ten observations, but they may belong to one machine unit. The sample size at the machine level is one. A study can analyze readings, machines, or machine-time combinations, but the unit must match the question and design.

Error 6: Writing a question that hides the denominator

“What percent preferred option A?” is incomplete unless the denominator is known. Was the percentage among all invited people, all respondents, all eligible participants, or all people who answered that item? A proportion always compares a count with a defined total.

Error 7: Using a summary as if it describes every individual

If the sample mean is 14.2, it does not follow that every person scored 14.2. If 60 percent answered yes, it does not follow that a typical person is “60 percent yes.” Summaries describe distributions or groups.

Error 8: Assuming more rows guarantee better evidence

A very large convenience sample can be systematically unrepresentative. Sample size influences random variability, but selection methods influence bias and generalizability. Topic 1.1 introduces the distinction; later topics develop sampling design.

Error 9: Forgetting the time frame

A population can change. “All employees” today may differ from all employees last year. A question about monthly spending needs a stated month or typical-month definition. Time is part of the context, not an optional detail.

Error 10: Writing a conclusion without the study noun

“The mean is 7.4” is incomplete. “The sample mean continuous operating time of the 60 tested earbud pairs was 7.42 hours under the standard test” is contextual. Introducing Statistics builds the habit of naming the variable, units, and sample.

A regional transit authority operates 340 buses. It wants to estimate the proportion of buses that require an unscheduled repair during a typical 30-day operating period. Engineers select 70 buses at the beginning of April and follow each bus for 30 days. Fourteen selected buses require at least one unscheduled repair. The data set contains one row per selected bus and columns for route group, bus age, miles traveled, number of repairs, and whether at least one unscheduled repair occurred.

Questions

1. State the investigative question.
2. Identify the population and its size.
3. Identify the sample and its size.
4. Identify the observational units.
5. List the variables named in the story.
6. Identify the sample statistic most directly connected to the question.
7. State the corresponding population parameter.
8. Give two reasons the observed results may vary among buses.
9. Explain why “14” is not the sample proportion.
10. Write a one-sentence contextual description of the sample result.

Why this assessment matters

The story contains several numbers, but only some answer the investigative question. N = 340 and n = 70 describe group sizes. Fourteen is a count. The value 0.20 is the sample proportion. The unknown population proportion is the parameter. Introducing Statistics teaches how to assign each number a role rather than treating every number as interchangeable.

Accessible population

The part of the target population that can realistically be reached through the available list, location, system, or time period.

Census

An attempt to collect data from every member of a defined population.

Context

The real-world meaning that connects statistical terms and calculations to the people, objects, variables, units, place, and time in the study.

Data

Recorded pieces of information about observational units.

Data set

An organized collection of data with labels, definitions, units, and structure.

Datum

One recorded piece of information; the singular form of data.

Descriptive statistics

Methods used to organize, display, and summarize observed data.

Estimate

A numerical value calculated from sample data to approximate a population parameter.

Estimator

A rule or statistic used to estimate a parameter.

Estimand

The precise population quantity a study intends to estimate.

Individual

A person, object, case, or entity described by the data.

Investigative question

A question that guides a statistical study and requires data to answer.

N

A common symbol for population size.

n

A common symbol for sample size.

Nonstatistical question

A question seeking a fixed fact, definition, or single result without requiring analysis of variability across observations.

Observational unit

The entity represented by one record at the level relevant to the question.

Parameter

A numerical characteristic of a population.

Population

The complete group of items or individuals relevant to the investigative question.

Sample

The subset of the population from which data are obtained.

Sampling variability

The natural tendency for statistics calculated from different samples to differ.

Statistic

A numerical characteristic calculated from sample data.

Statistical inference

Using sample data to draw conclusions about a larger population while accounting for uncertainty.

Statistical question

A question that anticipates variability and requires data from a group or repeated process.

Statistical study

A study in which data are collected and analyzed to answer an investigative question about a population or process.

Target population

The full group the researcher ultimately wants to understand.

Variable

A characteristic recorded for each observational unit that can take different values.

Variability

Differences among observations, measurements, occasions, or possible samples.

The final concept map

Every later AP Statistics topic adds tools to this map. Graphs reveal distributions. Probability models random behavior. Confidence intervals estimate parameters. Hypothesis tests evaluate claims. Regression describes relationships. Yet those methods remain meaningful only when the population, sample, unit, variables, and investigative question are clear.

Use this final bank as a rapid conceptual check. Each statement condenses one habit developed in the chapter.

• Introducing Statistics connects the population to the sample before interpreting a number.
• Introducing Statistics asks what one row represents before counting observations.
• Introducing Statistics distinguishes a population parameter from a sample statistic.
• Introducing Statistics treats variability as expected information rather than an inconvenience.
• Introducing Statistics requires every investigative question to identify measurable information.
• Introducing Statistics keeps units, labels, and time frames attached to the data.
• Introducing Statistics uses contextual language so a conclusion answers the original question.
• Introducing Statistics recognizes that a large data set can still come from a weak selection process.
• Introducing Statistics separates a fixed factual question from a question that expects a distribution.
• Introducing Statistics turns broad interests into populations, variables, and answerable comparisons.
• Introducing Statistics connects the population to the sample before interpreting a number.
• Introducing Statistics asks what one row represents before counting observations.
• Introducing Statistics distinguishes a population parameter from a sample statistic.
• Introducing Statistics treats variability as expected information rather than an inconvenience.
• Introducing Statistics requires every investigative question to identify measurable information.
• Introducing Statistics keeps units, labels, and time frames attached to the data.
• Introducing Statistics uses contextual language so a conclusion answers the original question.
• Introducing Statistics recognizes that a large data set can still come from a weak selection process.
• Introducing Statistics separates a fixed factual question from a question that expects a distribution.