View source for Glossary of probability and statistics

{{Short description|none}}
This '''glossary of statistics and probability''' is a list of definitions of terms and concepts used in the [[mathematics|mathematical]] sciences of [[statistics]] and [[probability]], their sub-disciplines, and related fields. For additional related terms, see [[Glossary of mathematics]] and [[Glossary of experimental design]].

{{compact ToC|side=yes|center=yes|nobreak=yes|seealso=yes|refs=yes|}}{{StatsTopicTOC}}
{{ProbabilityTopicsTOC}}

== A ==
{{glossary}}
{{term|[[admissible decision rule]]}}
{{defn|}}

{{term|[[algebra of random variables]]}}
{{defn|}}

{{term|[[alternative hypothesis]]}}
{{defn|}}

{{term|[[analysis of variance]]}}
{{defn|}}

{{term|[[Elementary event|atomic event]]}}
{{defn|Another name for {{gli|elementary event}}.}}
{{glossaryend}}

== B ==
{{glossary}}
{{term|[[bar chart]]}}
{{defn|}}

{{term|[[Bayes' theorem]]}}
{{defn|}}

{{term|[[Bayes estimator]]}}
{{defn|}}

{{term|[[Bayes factor]]}}
{{defn|}}

{{term|[[Bayesian inference]]}}
{{defn|}}

{{term|[[Bias (statistics)|bias]]}}
{{defn|no=1|Any feature of a {{gli|sample}} that is not representative of the larger {{gli|population}}.}}
{{defn|no=2|The difference between the expected value of an {{gli|estimator}} and the true value.}}

{{term|[[binary data]]}}
{{defn|Data that can take only two values, usually represented by the [[binary digit]]s 0 and 1.}}

{{term|[[binomial distribution]]}}
{{defn|}}

{{term|[[bivariate analysis]]}}
{{defn|A type of quantitative statistical analysis in which exactly two {{gli|variable|variables}} are analyzed, for the purpose of determining the empirical relationship between them. Contrast ''{{gli|multivariate analysis}}''.}}

{{term|[[blocking (statistics)|blocking]]}}
{{defn|In experimental design, the arranging of experimental units in groups ("blocks") that are similar to one another. Blocking is often used to manage the problem of {{gli|pseudoreplication}}.}}

{{term|[[Box–Jenkins method]]}}
{{defn|}}

{{term|[[box plot]]}}
{{defn|}}
{{glossaryend}}

== C ==
{{glossary}}
{{term|[[causality#Statistics and economics|causal study]]}}
{{defn|A statistical study in which the objective is to measure the effect of some {{gli|variable}} on the outcome of a different variable. For example, a causal study might ask the question: "How will my headache feel if I take aspirin, versus if I do not take aspirin?" Causal studies may be either experimental or observational.<ref name=Reiter>{{cite journal|last1=Reiter|first1=Jerome|title=Using Statistics to Determine Causal Relationships|journal=American Mathematical Monthly|volume=107|issue=1|pages=24–32|date=January 24, 2000|doi=10.2307/2589374 |jstor=2589374}}</ref>}}
{{defn|}}

{{term|[[central limit theorem]]}}
{{defn|}}

{{term|[[central moment]]}}
{{defn|}}

{{term|[[characteristic function (probability theory)|characteristic function]]}}
{{defn|}}

{{term|[[chi-squared distribution]]}}
{{defn|}}

{{term|[[chi-squared test]]}}
{{defn|}}

{{term|[[cluster analysis]]}}
{{defn|}}

{{term|[[cluster sampling]]}}
{{defn|}}

{{term|[[complementary event]]}}
{{defn|}}

{{term|[[completely randomized design]]}}
{{defn|}}

{{term|[[computational statistics]]}}
{{defn|The study of statistical methods that are enabled by using computational methods, at the interface of statistics and [[computer science]].}}

{{term|concomitants}}
{{defn|In a statistical study, any variables whose values are unaffected by experimental treatments, such as a unit’s age, gender, and cholesterol level before starting an experimental diet.<ref name=Reiter/>}}

{{term|[[conditional distribution]]}}
{{defn|Given two jointly distributed random variables ''X'' and ''Y'', the conditional probability distribution of ''Y'' given ''X'' (written "''Y'' <nowiki>|</nowiki> ''X''") is the probability distribution of ''Y'' when ''X'' is known to be a particular value.}}

{{term|[[conditional probability]]}}
{{defn|The {{gli|probability}} of some event A, assuming the occurrence of event B. In mathematical notation, conditional probability is written P(''A''<nowiki>|</nowiki>''B''), and is read "the probability of ''A'', given ''B''".}}

{{term|[[conditional probability distribution]]}}
{{defn|}}

{{term|[[confidence interval]] (CI)}}
{{defn|In inferential statistics, a range of plausible values for some unknown {{gli|parameter}}, such as a {{gli|population mean}}, defined as an interval with a lower bound and an upper bound.<ref name="Kalinowski">Pav Kalinowski. Understanding Confidence Intervals (CIs) and Effect Size Estimation. Association for Psychological Science Observer April 10, 2010. http://www.psychologicalscience.org/index.php/publications/observer/2010/april-10/understanding-confidence-intervals-cis-and-effect-size-estimation.html</ref> The precise values of these bounds are calculated from a pre-determined {{gli|confidence level}}, chosen by the researcher. The confidence level represents the frequency of intervals that, over the long run, capture the true value of the unknown parameter; i.e. 95% of confidence intervals computed at the 95% confidence level contain the true value, and likewise for other confidence levels. For example, based on a study of sleep habits among a random sample of 100 people, a researcher may estimate at the 95% confidence level that the overall population sleeps somewhere between 5 and 9 hours per night. There is a 95% chance that the true population mean falls within this interval, because 95% of random samples taken from this same population will yield 95% confidence intervals that contain the true mean.}}

{{term|confidence level}}
{{ghat|Also '''confidence coefficient'''.}}
{{defn|A number indicating the probability that the {{gli|confidence interval}} (range) captures the true {{gli|population mean}}. For example, a confidence interval with a 95% confidence level has a 95% chance of capturing the population mean. Technically, this means that, if the experiment were repeated many times, 95% of the CIs computed at this level would contain the true population mean.<ref name="Kalinowski"/>}}

{{term|[[confounder]]}}
{{defn|A {{gli|variable}} that influences both the {{gli|dependent variable}} and the {{gli|independent variable}}, causing a spurious association. The existence of hidden confounding variables is an important quantitative explanation why {{gli|correlation}} does not imply causation: if changes in two variables appear to be correlated, it is risky to presume that one change causes the other because it is possible that one or more unidentified confounders has in fact caused the changes in both variables. A classic example is the correlation between increased consumption of ice cream and increased crime in the summer. It is irrational to assume that eating more ice cream makes people commit more crime, or vice versa; it is more likely that one or more additional variables, e.g. warmer weather, increase both ice cream consumption and crime simultaneously. In this example, warmer weather is the confounder.}}

{{term|[[conjugate prior]]}}
{{defn|}}

{{term|[[continuous variable]]}}
{{defn|}}

{{term|[[convenience sampling]]}}
{{defn|}}

{{term|[[correlation]]}}
{{ghat|Also '''correlation coefficient'''.}}
{{defn|A numeric measure of the strength of a linear relationship between two random variables (one can use it to quantify, for example, how shoe size and height are correlated in the population). An example is the [[Pearson product-moment correlation coefficient]], which is found by dividing the {{gli|covariance}} of the two variables by the product of their {{gli|standard deviation|standard deviations}}. {{gli|independent variable|Independent variables}}, by definition, have a correlation of 0. A population correlation is often represented by the symbol <math>\rho</math>, and a sample correlation by <math>r</math>.}}

{{term|[[count data]]}}
{{defn|Data arising from [[counting]], and which can therefore take only non-negative [[integer]] values.}}

{{term|[[covariance]]}}
{{defn|Given two random variables ''X'' and ''Y'', with expected values <math>E(X)=\mu</math> and <math>E(Y)=\nu</math>, the expected value of random variable <math>(X - \mu) (Y - \nu)</math>, written in statistical notation as <math>\operatorname{cov}(X, Y)</math>. The covariance is used for measuring {{gli|correlation}}; it can be interpreted as the degree to which the two variables change simultaneously with each other or "co-vary".}}
{{glossaryend}}

== D ==
{{glossary}}
{{term|[[data]]}}
{{defn|}}

{{term|[[data analysis]]}}
{{defn|}}

{{term|[[data set]]}}
{{defn|A sample and the associated {{gli|data point|data points}}.}}
{{defn|}}

{{term|[[data point]]}}
{{defn|A typed measurement&nbsp;— it can be a [[Boolean data type|Boolean]] value, a real number, a vector (in which case it is also called a ''data vector''), etc.}}

{{term|[[decision rule]]}}
{{defn|}}

{{term|[[decision theory]]}}
{{defn|}}

{{term|[[Degrees of freedom (statistics)|degrees of freedom]]}}
{{defn|}}

{{term|[[density estimation]]}}
{{defn|}}

{{term|[[correlation and dependence|dependence]]}}
{{defn|}}

{{term|[[dependent variable]]}}
{{defn|}}

{{term|[[descriptive statistics]]}}
{{defn|}}

{{term|[[design of experiments]]}}
{{defn|}}

{{term|[[deviation (statistics)|deviation]]}}
{{defn|}}

{{term|[[discrete variable]]}}
{{defn|}}

{{term|[[dot plot (statistics)|dot plot]]}}
{{defn|}}

{{term|[[double counting (fallacy)|double counting]]}}
{{defn|}}
{{glossaryend}}

== E ==
{{glossary}}
{{term|[[elementary event]]}}
{{defn|An event which contains only a single outcome in the {{gli|sample space}}; in a set of possibilities, a possibility that can occur in precisely one way. For example, when pulling a card from a standard deck of [[playing cards]], 'pulling the jack of spades' is an elementary event (because there is only one jack of spades in the entire deck), while 'pulling a king or an ace' is not (because there are a combined four kings and four aces in the deck).}}

{{term|[[estimation theory]]}}
{{defn|The branch of statistics concerned with estimating the values of {{gli|parameter|parameters}} based on measured empirical data with a {{gli|random}} component. The parameters describe an underlying physical setting in such a way that their values affect the distribution of the measured data; an {{gli|estimator}} attempts to use the measurements to approximate the unknown parameters.}}

{{term|[[estimator]]}}
{{defn|A function of the known data that is used to estimate an unknown {{gli|parameter}}; an estimate is the result of the actual application of the function to a particular set of data. For example, the {{gli|mean}} can be used as an estimator.}}

{{term|[[expected value]]}}
{{ghat|Also '''expectation''', '''mathematical expectation''', '''first moment''', or simply '''mean''' or '''average'''.}}
{{defn|The sum of the probabilities of each possible outcome of an {{gli|experiment}} multiplied by their corresponding {{gli|payoff}} or "value". Thus, it represents the average amount one "expects" to win per bet if bets with identical odds are repeated many times. For example, the expected value of rolling a fair six-sided die is 3.5. The concept is, intuitively, a generalization of the weighted average of all possible outcomes of a particular procedure or experiment, and can be viewed as the [[arithmetic mean]] of a large number of independent realizations of the experiment. The expected value of random variable ''X'' is typically written as E(X) for the ''expectation operator'', and <math>\mu</math> ([[Mu (letter)|mu]]) for the parameter.}}

{{term|[[Experiment (probability theory)|experiment]]}}
{{defn|Any procedure which can be infinitely repeated and which has a well-defined set of outcomes.}}

{{term|[[exponential family]]}}
{{defn|}}

{{term|[[Event (probability theory)|event]]}}
{{defn|A subset of the {{gli|sample space}} of a procedure or {{gli|experiment}} (i.e. a possible outcome) to which a probability can be assigned. For example, on rolling a die, "getting a three" is an event (with a probability of {{frac|1|6}} if the die is fair), as is "getting a five or a six" (with a probability of {{frac|1|3}}).}}
{{glossaryend}}

== F ==
{{glossary}}
{{term|[[factor analysis]]}}

{{term|[[factorial experiment]]}}

{{term|[[frequency (statistics)|frequency]]}}

{{term|[[frequency distribution]]}}

{{term|[[frequency domain]]}}

{{term|[[frequentist inference]]}}
{{glossaryend}}

== G ==
{{glossary}}
{{term|[[general linear model]]}}

{{term|[[generalized linear model]]}}

{{term|[[grouped data]]}}
{{glossaryend}}

== H ==
{{glossary}}
{{term|[[histogram]]}}
{{defn|An approximate graphical representation of the distribution of numerical data. A histogram displays this distribution by dividing the entire {{gli|range}} of values into a series of consecutive, non-overlapping intervals and then counting how many instances of the dataset fall into each interval.}}
[[File:Tips-histogram1.png|thumb|right|A '''{{gli|histogram}}''' showing the distribution of tips given in a restaurant]]
{{glossaryend}}

== I ==
{{glossary}}
{{term|[[Independence (probability theory)|independence]]}}
{{defn|}}

{{term|[[independent variable]]}}
{{defn|}}

{{term|[[interquartile range]] (IQR)}}
{{ghat|Also '''midspread''', '''middle 50%''', and '''H-spread'''.}}
{{defn|defn=A measure of the statistical dispersion or spread of a dataset, defined as the difference between the 25th and 75th {{gli|percentile|percentiles}} of the data. To calculate the IQR, the dataset is divided into four rank-ordered even parts or {{gli|quartile|quartiles}}, the boundaries between which, at the 25th, 50th, and 75th percentiles, are denoted <math>Q</math><sub>1</sub>, <math>Q</math><sub>2</sub>, and <math>Q</math><sub>3</sub>, respectively; the IQR = <math>Q</math><sub>3</sub><math>-</math><math>Q</math><sub>1</sub>.}}
{{glossaryend}}

== J ==
{{glossary}}
{{term|[[joint distribution]]}}
{{defn|Given two random variables ''X'' and ''Y'', the joint distribution of ''X'' and ''Y'' is the {{gli|probability distribution}} of X and Y together.}}

{{term|[[joint probability]]}}
{{defn|The {{gli|probability}} of two {{gli|event|events}} occurring together. The joint probability of ''A'' and ''B'' is written <math>P(A \cap B)</math> or <math>P(A, \ B)</math>.}}
{{glossaryend}}

== K ==
{{glossary}}
{{term|[[Kalman filter]]}}
{{defn|}}

{{term|[[kernel (statistics)|kernel]]}}
{{defn|}}

{{term|[[kernel density estimation]]}}
{{defn|}}

{{term|[[kurtosis]]}}
{{defn|A measure of the "tailedness" of the {{gli|probability distribution}} of a real-valued random variable. There are different ways of quantifying, estimating, and interpreting kurtosis, but a common interpretation is that kurtosis represents the degree to which the shape of the distribution is influenced by infrequent extreme observations ({{gli|outlier|outliers}}); in this case, higher kurtosis means more of the {{gli|variance}} is due to infrequent extreme deviations, as opposed to frequent modestly sized deviations.}}
{{glossaryend}}

== L ==
{{glossary}}
{{term|[[L-moment]]}}
{{defn|}}

{{term|[[law of large numbers]] (LLN)}}
{{defn|A theorem according to which the average of the results obtained from performing the same experiment a large number of times should be close to the experiment's {{gli|expected value}}, and tends to become closer to the expected value as more trials are performed. The law suggests that a sufficiently large number of trials is necessary for the results of any experiment to be considered reliable, and by extension that performing only a small number of trials may produce an incomplete or misleading interpretation of the experiment's outcomes.}}
[[File:Lawoflargenumbers.svg|thumb|right|A graph illustrating the '''{{gli|law of large numbers}}''': the observed results of repeatedly rolling a fair die (green line) quickly approach the theoretical {{gli|expected value}} of a single roll (blue line) as the number of times the die is rolled increases]]

{{term|[[likelihood function]]}}
{{defn|A conditional probability function considered a function of its second argument with its first argument held fixed. For example, imagine pulling a numbered ball with a number ''k'' from a bag of ''n'' balls, numbered 1 to ''n''; a likelihood function for the random variable ''N'' could be described as the probability of pulling ''k'' given that there are ''n'' balls: the likelihood will be 1/''n'' for n greater than or equal to ''k'', and 0 for ''n'' smaller than ''k''. Unlike a probability distribution function, this likelihood function will not sum up to 1 on the {{gli|sample space}}.}}

{{term|[[loss function]]}}
{{defn|}}

{{term|[[likelihood-ratio test]]}}
{{defn|}}
{{glossaryend}}

== M ==
{{glossary}}
{{term|[[M-estimator]]}}
{{defn|}}

{{term|[[marginal distribution]]}}
{{defn|Given two {{gli|joint distribution|jointly distributed}} random variables ''X'' and ''Y'', the marginal distribution of ''X'' is simply the {{gli|probability distribution}} of ''X'' when information about ''Y'' is ignored.}}

{{term|[[marginal likelihood]]}}
{{defn|}}

{{term|[[marginal probability]]}}
{{defn|The probability of a given event, ignoring any information about other events. The marginal probability of ''A'' is written ''P''(''A''). Contrast ''{{gli|conditional probability}}''.}}

{{term|[[Markov chain Monte Carlo]]}}
{{defn|}}

{{term|[[mathematical statistics]]}}
{{defn|}}

{{term|[[maximum likelihood estimation]]}}
{{defn|}}

{{term|[[mean]]}}
{{defn|no=1|The {{gli|expected value}} of a random variable.}}
{{defn|no=2|The [[arithmetic mean]], i.e. the mathematical average of a set of numerical values, calculated by dividing the sum of the values by the number of values.}}

{{term|[[median]]}}
{{defn|}}

{{term|[[median absolute deviation]]}}
{{defn|}}

{{term|[[mode (statistics)|mode]]}}
{{defn|}}

{{term|[[moving average]]}}
{{ghat|Also '''moving mean''' and '''rolling mean'''.}}
{{defn|A series of mathematical averages or {{gli|mean|means}} of different subsets of a larger data set, usually computed so as to understand trends in the data set over time.}}

{{term|[[multimodal distribution]]}}
{{defn|}}

{{term|[[multivariate analysis]]}}
{{defn|}}

{{term|[[multivariate kernel density estimation]]}}
{{defn|}}

{{term|[[multivariate random variable]]}}
{{defn|A vector whose components are random variables on the same probability space.}}

{{term|[[mutual exclusivity]]}}
{{defn|}}

{{term|[[mutual independence]]}}
{{defn|A collection of events is said to be ''mutually independent'' if for any subset of the collection, the joint probability of all events occurring is equal to the product of the joint probabilities of the individual events. Think of the result of a series of coin-flips. This is a stronger condition than pairwise independence.}}
{{glossaryend}}

== N ==
{{glossary}}
{{term|[[nonparametric regression]]}}
{{defn|}}

{{term|[[nonparametric statistics]]}}
{{defn|}}

{{term|[[non-sampling error]]}}
{{defn|}}

{{term|[[normal distribution]]}}
{{defn|}}

{{term|[[normal probability plot]]}}
{{defn|}}

{{term|term=[[null hypothesis]]|content=[[null hypothesis]] (H<sub>0</sub>)}}
{{defn|The statement being tested in a test of {{gli|statistical significance}}; usually a statement of 'no effect' or 'no difference'.<ref name=moore>{{cite book|last1=Moore|first1=David|last2=McCabe|first2=George|title=Introduction to the Practice of Statistics|publisher=W.H. Freeman and Co|location=New York|year=2003|page=438|edition=4|isbn=9780716796572}}</ref> For example, in a test of whether light has an effect on sleep, the null hypothesis would be that light has no effect on sleep (i.e. sleep patterns are the same regardless of the lighting conditions). The null hypothesis is an expression of the expectation that the {{gli|dependent variable}} will not change significantly as the {{gli|independent variable}} is modified; statistical significance is measured and reported according to the degree to which this expectation is met. Contrast ''{{gli|alternative hypothesis}}''.}}
{{glossaryend}}

== O ==
{{glossary}}
{{term|[[opinion poll]]}}
{{defn|}}

{{term|[[optimal decision]]}}
{{defn|}}

{{term|[[optimal design]]}}
{{defn|}}

{{term|[[outlier]]}}
{{defn|}}
{{glossaryend}}

== P ==
{{glossary}}
{{term|[[p-value]]}}
{{defn|}}

{{term|[[pairwise independence]]}}
{{defn|A set of random variables, any two of which are independent.}}

{{term|[[Statistical parameter|parameter]]}}
{{defn|Any measured quantity of a statistical {{gli|population}} that summarizes or describes an aspect of the population, e.g. a {{gli|mean}} or a {{gli|standard deviation}}; often a quantity to be estimated based on the corresponding quantity calculated by drawing {{gli|random sample|random samples}} from the population. Can be a population parameter, a distribution parameter, or an unobserved parameter.}}

{{term|[[particle filter]]}}
{{defn|}}

{{term|[[percentile]]}}
{{defn|}}

{{term|[[pie chart]]}}
{{defn|}}

{{term|[[point estimation]]}}
{{defn|}}

{{term|[[power (statistics)|power]]}}
{{defn|}}

{{term|[[prior probability]]}}
{{defn|In [[Bayesian inference]], prior beliefs or other information that is available before new data or observations are taken into account.}}

{{term|[[population parameter]]}}
{{defn|See ''{{gli|parameter}}''.}}

{{term|[[posterior probability]]}}
{{defn|The result of a [[Bayesian analysis]] that encapsulates the combination of prior beliefs or information (the {{gli|prior probability}}) with observed data.}}

{{term|[[principal component analysis]] (PCA)}}
{{defn|}}

{{term|[[probability]]}}
{{defn|}}

{{term|[[Probability density function|probability density]]}}
{{defn|The probability in a continuous probability distribution. For example, you can't say that the probability of a man being six feet tall is 20%, but you can say he has 20% of chances of being between five and six feet tall. Probability density is given by a {{gli|probability density function}}. Contrast ''{{gli|probability mass}}''.}}

{{term|[[probability density function]]}}
{{defn|The probability distribution for a continuous random variable.}}

{{term|[[probability distribution]]}}
{{defn|A function that gives the probability of all elements in a given space; see [[List of probability distributions]].}}

{{term|[[probability measure]]}}
{{defn|The {{gli|probability}} of events in a {{gli|probability space}}.}}

{{term|[[Probability plot (disambiguation)|probability plot]]}}
{{defn|}}

{{term|[[probability space]]}}
{{defn|A {{gli|sample space}} over which a {{gli|probability measure}} has been defined.}}
{{glossaryend}}

== Q ==
{{glossary}}
{{term|[[quantile]]}}
{{defn|A particular point or value at which the range of a probability distribution is divided into continuous intervals with equal probabilities, or at which the observations in a sample are divided in the same way. The number of groups into which the range is divided is always one greater than the number of quantiles dividing them. Commonly used quantiles include {{gli|quartile|quartiles}} (which divide a range into four groups), {{gli|decile|deciles}} (ten groups), and {{gli|percentile|percentiles}} (one hundred groups). The groups themselves are termed halves, thirds, quarters, etc., though the terms for the quantiles are sometimes used to refer to the groups, rather than to the cut points.}}

{{term|[[quartile]]}}
{{defn|A type of {{gli|quantile}} which divides a range of data points into four groups, termed ''quarters'', of equal size. For any quartile-divided dataset, there are exactly three quartiles or cut points that create the four groups. The first quartile (<math>Q</math><sub>1</sub>) is defined as the middle data point or value that is halfway between the smallest value (minimum) and the {{gli|median}} of the dataset, such that 25 percent of the data lies below this quartile. The second quartile (<math>Q</math><sub>2</sub>) is the median itself, with 50 percent of the data below this point. The third quartile (<math>Q</math><sub>3</sub>) is defined as the middle value halfway between the median and the largest value (maximum) of the dataset, such that 75 percent of the data lies below this quartile. Because the data must be ordered from smallest to largest in order to compute them, quartiles are a type of {{gli|order statistic}}.}}
[[File:Iqr with quantile.png|thumb|right|A {{gli|normal distribution}} divided by three '''{{gli|quartile|quartiles}}''' into four quarters. Note that each of the quarters occupies the same total area under the curve, equal to 25 percent of the distribution.]]

{{term|[[quota sampling]]}}
{{defn|}}
{{glossaryend}}

== R ==
{{glossary}}
{{term|[[random variable]]}}
{{defn|A measurable function on a probability space, often real-valued. The distribution function of a random variable gives the probability of the different values of the variable. The mean and variance of a random variable can also be derived. See also ''{{gli|discrete random variable}}'' and ''{{gli|continuous random variable}}''.}}

{{term|[[Blocking (statistics)#Randomized block design|randomized block design]]}}
{{defn|}}

{{term|[[range (statistics)|range]]}}
{{defn|The length of the smallest interval which contains all of the data in a dataset, calculated as the arithmetic difference between the largest and smallest values contained in the dataset and expressed in the same units used for the data. The range provides a measure of the statistical dispersion of the dataset.}}

{{term|[[recursive Bayesian estimation]]}}
{{defn|}}

{{term|[[regression analysis]]}}
{{defn|A data analysis or statistical model that employs a set of statistical methods which estimate the relationships between a dependent variable and one or more error-free independent variables. The most common example of such a model is a {{gli|linear regression}}, in which the equation of a line that most closely fits a set of data points according to a specific mathematical criterion is calculated and plotted on a graph of the data points; other forms of regression use related methods to estimate alternative {{gli|parameters}} or to estimate conditional expectations from various non-linear models.}}

{{term|[[repeated measures design]]}}
{{defn|}}

{{term|[[response variable]]}}
{{defn|Any {{gli|variable}} whose value is or is expected to be affected by an experimental treatment, or by changes in one or more other variables; e.g. cholesterol levels after following a particular diet for six months. Response variables are those that change or respond to some phenomenon under study. The term is often used interchangeably with {{gli|dependent variable}}.<ref name=Reiter/>}}

{{term|[[restricted randomization]]}}
{{defn|}}

{{term|[[robust statistics]]}}
{{defn|}}

{{term|[[round-off error]]}}
{{defn|}}
{{glossaryend}}

== S ==
{{glossary}}
{{term|[[statistical sample|sample]]}}
{{defn|That part of a {{gli|population}} which is actually observed or measured.}}

{{term|[[sample mean and covariance|sample covariance]]}}
{{defn|}}

{{term|[[sample mean and covariance|sample mean]]}}
{{defn|The {{gli|arithmetic mean}} of a {{gli|sample}} of values drawn from a population, commonly denoted by <math>\overline{x}</math>. An example is the average test score of a subset of 10 students from a class. On the assumption that the sample is representative of the larger population, the sample mean is often used as an estimator of the {{gli|population mean}}, which in this example would be the average test score of all of the students in the class.}}

{{term|[[sample space]]}}
{{defn|The set of possible outcomes of an experiment. For example, the sample space for rolling a six-sided die is <nowiki>{1, 2, 3, 4, 5, 6}</nowiki>.}}

{{term|[[sampling (statistics)|sampling]]}}
{{defn|The process of using a fraction of observations or {{gli|samples}} selected from a larger {{gli|population}} in order to study and estimate characteristics of the larger population. Sampling relies on the assumption that a {{gli|sample size|sufficiently sized}} subset of data points chosen from a larger set of data points, when chosen according to appropriate criteria, is representative of the larger set as a whole – i.e. that the statistical properties of the subset are nearly the same as the properties of the larger set – and therefore that researchers are justified in using the samples to make inferences about the larger population without having to actually observe or measure every single data point in the population, which may be prohibitively expensive, difficult, or impossible to do in reality.}}

{{term|[[sampling bias]]}}
{{defn|}}

{{term|[[sampling distribution]]}}
{{defn|The {{gli|probability distribution}}, obtained by repeated {{gli|sampling}} of the {{gli|population}}, of a given statistic.}}

{{term|[[sampling error]]}}
{{defn|}}

{{term|[[scatter plot]]}}
{{defn|}}

{{term|[[scale parameter]]}}
{{defn|}}

{{term|significance level}}
{{defn|}}

{{term|[[simple random sample]]}}
{{defn|}}

{{term|[[Simpson's paradox]]}}
{{defn|}}

{{term|[[skewness]]}}
{{defn|A measure of the asymmetry of the {{gli|probability distribution}} of a real-valued random variable about its {{gli|mean}}. Roughly speaking, a distribution has positive skew (right-skewed) if the higher tail is longer, and negative skew (left-skewed) if the lower tail is longer. Perfectly symmetrical distributions always have zero skewness, though zero skewness does not necessarily imply a symmetrical distribution.}}
[[File:Relationship between mean and median under different skewness.png|thumb|right|The {{gli|mean}} and {{gli|median}} of a '''{{gli|skewness|skewed}}''' distribution (left and right) may differ substantially from those of a symmetrical distribution (center) with zero skewness.]]

{{term|[[spaghetti plot]]}}
{{defn|}}

{{term|[[spectrum bias]]}}
{{defn|}}

{{term|[[standard deviation]]}}
{{defn|The most commonly used measure of statistical dispersion. It is the [[square root]] of the {{gli|variance}}, and is generally denoted with the lowercase Greek letter <math>\sigma</math> ([[Sigma (letter)|sigma]]).}}

{{term|[[standard error]]}}
{{defn|}}

{{term|[[standard score]]}}
{{defn|}}

{{term|[[statistic]]}}
{{defn|The result of applying a statistical algorithm to a data set. It can also be described as an observable random variable.}}

{{term|[[statistical dispersion]]}}
{{defn|}}

{{term|[[statistical graphics]]}}
{{defn|}}

{{term|[[statistical hypothesis testing]]}}
{{defn|}}

{{term|[[statistical independence]]}}
{{defn|Two events are independent if the outcome of one does not affect that of the other (for example, getting a 1 on a single die roll does not affect the probability of getting a 1 on a second roll). Similarly, when we assert that two random variables are independent, we intuitively mean that knowing something about the value of one of them does not yield any information about the value of the other.}}

{{term|[[statistical inference]]}}
{{defn|Inference about a population based on a random sample drawn from that population or, more generally, about a random process from its observed behavior during a finite period of time.}}

{{term|[[statistical model]]}}
{{defn|}}

{{term|[[statistical population]]}}
{{defn|A set of entities about which statistical inferences are to be drawn, often based on random sampling. One can also talk about a population of measurements or values.}}

{{term|[[statistical dispersion]]}}
{{defn|A measure of the diversity within a set of data, expressed by the {{gli|variance}} or the {{gli|standard deviation}}.}}

{{term|[[statistical parameter]]}}
{{defn|A {{gli|parameter}} that indexes a family of probability distributions.}}

{{term|[[statistical significance]]}}
{{defn|}}

{{term|[[statistics]]}}
{{defn|}}

{{term|[[Student's t-test]]}}
{{defn|}}

{{term|[[stem-and-leaf display]]}}
{{defn|}}

{{term|[[stratified sampling]]}}
{{defn|}}

{{term|[[survey methodology]]}}
{{defn|}}

{{term|[[survival function]]}}
{{defn|}}

{{term|[[survivorship bias]]}}
{{defn|}}

{{term|[[symmetric probability distribution]]}}
{{defn|}}

{{term|[[systematic sampling]]}}
{{defn|}}
{{glossaryend}}

== T ==
{{glossary}}
{{term|[[test statistic]]}}
{{defn|}}

{{term|tidy data}}
{{defn|Standard for structuring data such that "each variable is a column, each observation is a row, and each type of observational unit is a table". It is equivalent to Codd's [[third normal form]].<ref>{{Cite journal|last=Wickham|first=Hadley|authorlink=Hadley Wickham|date=2014|title=Tidy Data|url=https://www.jstatsoft.org/index.php/jss/article/view/v059i10/v59i10.pdf|journal=Journal of Statistical Software|volume=59|issue=10|doi=10.18637/jss.v059.i10 |doi-access=free }}</ref>}}

{{term|[[time domain]]}}
{{defn|}}

{{term|[[time series]]}}
{{defn|}}

{{term|[[time series#time series analysis|time series analysis]]}}
{{defn|}}

{{term|[[time series#time series analysis|time series forecasting]]}}
{{defn|}}

{{term|treatments}}
{{defn|Variables in a statistical study that are conceptually manipulable. For example, in a health study, following a certain diet is a treatment whereas age is not.<ref name=Reiter/>}}

{{term|[[Experiment (probability theory)|trial]]}}
{{defn|Can refer to each individual repetition when talking about an experiment composed of any fixed number of them. As an example, one can think of an experiment being any number from one to ''n'' coin tosses, say 17. In this case, one toss can be called a trial to avoid confusion, since the whole experiment is composed of 17 ones.}}

{{term|[[trimmed estimator]]}}
{{defn|}}

{{term|[[type I and type II errors]]}}
{{defn|}}

{{glossaryend}}

== U ==
{{glossary}}
{{term|[[unimodal probability distribution]]}}
{{defn|}}

{{term|units}}
{{defn|In a statistical study, the objects to which treatments are assigned. For example, in a study examining the effects of smoking cigarettes, the units would be people.<ref name=Reiter/>}}

{{glossaryend}}

== V ==
{{glossary}}
{{term|[[variance]]}}
{{defn|A measure of its statistical dispersion of a random variable, indicating how far from the expected value its values typically are. The variance of random variable ''X'' is typically designated as <math>\operatorname{var}(X)</math>, <math>\sigma_X^2</math>, or simply <math>\sigma^2</math>}}

{{glossaryend}}

== W ==
{{glossary}}
{{term|[[weighted arithmetic mean]]}}
{{defn|}}

{{term|[[weighted median]]}}
{{defn|}}

{{glossaryend}}

== X ==
{{glossary}}
{{term|[[XOR|XOR, exclusive disjunction]]}}
{{defn|}}

{{glossaryend}}

== Y ==

{{glossary}}
{{term|Yates's correction for continuity, yules correction}}
{{defn|}}

{{glossaryend}}

== Z ==
{{glossary}}
{{term|[[z-test]]}}
{{defn|}}

{{glossaryend}}

== See also ==
* [[Notation in probability and statistics]]
* [[Probability axioms]]
* [[Glossary of experimental design]]
* [[List of statistical topics]]
* [[List of probability topics]]
* [[Glossary of areas of mathematics]]
* [[Glossary of calculus]]

== References ==
{{Reflist}}

== External links ==
{{wiktionary|Appendix:Glossary of probability and statistics}}
* {{citation | title = NIST/SEMATECH e-Handbook of Statistical Methods | chapter = A Glossary of DOE Terminology | chapter-url = http://www.itl.nist.gov/div898/handbook/pri/section7/pri7.htm | publisher= [[NIST]] | access-date = 28 February 2009}}
* {{ citation | url = http://www.statistics.com/resources/glossary/ | title = Statistical glossary | publisher = statistics.com | access-date = 28 February 2009}}
* [http://www.economics.soton.ac.uk/staff/aldrich/Probability%20Earliest%20Uses.htm Probability and Statistics on the Earliest Uses Pages (Univ. of Southampton)]

{{Statistics}}
{{Glossaries of science and engineering}}

[[Category:Probability and statistics| Glossary]]
[[Category:Statistics-related lists]]
[[Category:Glossaries of science|Probability and statistics]]
[[Category:Glossaries of mathematics|Probability and statistics]]
[[Category:Wikipedia glossaries using description lists]]