A statistic having a multivariate distribution which is the analogue of the univariate t-distribution. The statistic was introduced in 1931 by Hotelling. A sample of size n is drawn from a multivariate normal distribution with mean vector μ and variance–covariance matrix estimated as S. If the column vector of sample means is denoted by x̆ the statistic T2 is given by
In the case of p variables, {(n−p)T2}/{(n−1)p} has an F-distribution with p and (n−p) degrees of freedom. See also Mahalanobis D2.
- two-by-two table
- two-clock paradox
- two-colour diagram
- two-colour photometry
- two-dimensional array
- two-dimensional flow
- two-dimensional infrared spectroscopy
- two-dimensional NMR
- two-envelope paradox
- two-gap model
- two-hybrid screening
- two-level grammars
- two-level memory
- Two-Micron All-Sky Survey
- two-norm
- two-part tariff
- two-person zero-sum game
- two-phase commit protocol
- two-phase multiplier
- two-phase system
- two-place predicate
- two-plus-one address
- two-port analysis
- two-port network
- two-port parameters