Glossary

The sum of observations divided by how many there are.

An attempt to measure every member of a population.

A sample drawn by randomly selecting a few whole groups and including every unit inside them.

A range of plausible values for an unknown population parameter, built from sample data.

An outside variable linked to both the cause and the effect being studied, which can make an unrelated pair look connected.

A variable that can, in principle, take any value in an interval.

A sample of whichever units are easiest to reach, with no randomness.

How strongly and in what direction two quantitative variables move together.

A single number, between -1 and 1, summarizing the strength and direction of a linear relationship.

The rectangular arrangement of a dataset -- one row per observation, one column per variable.

Summarizing and organizing the data actually collected, with no claim beyond it.

A question with one correct answer fixed by known inputs, with no variability to account for.

How far a single value sits from the mean, computed as the value minus the mean.

A variable that can only take specific, separated values, with nothing possible between two consecutive ones.

The pattern of how often each value or range of values occurs in a dataset or random variable.

A subset of the sample space -- a described collection of outcomes.

The long-run average value of a random variable, weighted by probability.

A study that deliberately assigns subjects to treatments and observes the outcome.

A bar chart of a quantitative variable where bar heights show how many values fall in each interval.

A claim about a population parameter that a statistical test is built to weigh evidence against.

One event or draw gives no information about another -- knowing one outcome doesn't change the odds of the other.

Using a sample's statistics to draw a conclusion about the population, with acknowledged uncertainty.

A level of measurement with equal numeric gaps, but an arbitrary zero.

The middle value of ordered data -- half the values fall above it, half below.

The most frequently occurring value in a dataset.

A level of measurement with categories and no order at all.

Bias from selected units failing to respond, when responders differ systematically from non-responders.

One row of a data matrix -- one unit the data was collected about.

A study that records variables as they naturally occur, with no intervention.

A level of measurement with ordered categories but unequal or unknown gaps.

A value far enough from the rest of the data that it stands apart.

A fixed number describing a population, usually unknown.

A proportion scaled to a base of 100.

The plain-subtraction difference between two percentages, distinct from percent change.

The value below which a given percentage of the data falls.

The entire group a statistical question is actually about.

A variable used to predict or explain another variable in a regression model.

A number between 0 and 1 describing how likely an event is, in the long run.

A part-to-whole ratio.

A variable recording a category or label, not a meaningful quantity.

A variable recording a numeric quantity that arithmetic applies to meaningfully.

One of the three values that split ordered data into four equal-sized groups.

A number assigned to the outcome of a random process.

The largest value minus the smallest value in a dataset.

A ratio comparing two different kinds of quantities, not a part to its own whole.

A level of measurement with equal numeric gaps and a true zero.

The gap between an observed value and the value a model predicted for it.

Bias from responders answering inaccurately, often due to leading question wording.

The subset of a population actually observed.

The set of every possible outcome of a random process.

A systematic tendency for a sampling method to favor some population members over others.

A sample drawn so every possible group of that size has an equal chance of selection.

A distribution with a longer tail on one side, pulling the mean away from the median.

The square root of the variance, a measure of spread in the same units as the data.

A number computed from a sample; it changes if a different sample is drawn.

A sample drawn by taking a separate SRS from within every stratum of the population.

A sample drawn by taking every k-th unit after one random starting point.

When part of a population has little or no chance of being selected because the sampling frame never included it.

How much individual values in a group differ from each other, rather than all matching a single number.

One column of a data matrix -- one fact recorded about every observation.

The average of the squared deviations from the mean, a measure of spread.

A sample where units opt themselves in, like an online poll.

A mean where some observations count more than others.