# [ST-APP] Summary of probability distributions

Miguel-Angel Canela
Department of Managerial Decision Analysis

Binomial distribution
Type: Discrete.
Applications: Modelling the distribution of the number of successes in a series of independent
trials.
µ ¶
n x¡ ¢n−x
Probability density: f (x) = π 1−π , x = 0, 1, . . . , n.
x
Parameters: n, π.
Notation: B(n, π).
Mean/Variance: µ = nπ, σ 2 = nπ(1 − π).
Stata functions: binomialp, binomial, binomialtail, invbinomial, invbinomialtail.

Poisson distribution
Type: Discrete.
Applications: Modelling the distribution of counts of defective items (quality control), customers
in a time interval (queuing theory), innovations (entrepeneurship), etc.
λx
Probability density: f (x) = e−λ , x = 0, 1, 2, . . ..
x!
Parameters: λ.
Mean/Variance: µ = σ 2 = λ.
Notation: P(λ).
Stata functions: poissonp, poisson, poissontail,invpoisson, invpoissontail.

Exponential distribution
Type: Continuous.
Applications: Modelling the distribution of the time to repair of a TV set (engineering), heart
stroke (biostatistics), quitting a business (strategic analysis), getting a job (labor stats), etc.
Probability density: f (x) = λ e−λx .
Parameter: λ.
Notation: E(λ).
Mean/Variance: µ = 1/λ, σ 2 = 1/λ.

Normal distribution
Type: Continuous.
Applications: Modelling the distribution of a continuous variable, as far as we do not have evidence
that it is not appropriate. In testing, it is used as an (asymptotic) model for the distribution of
many statistics.

[ST-APP] Summary of probability distrubutions / 1 2014–09–01

invnormal. invttail. σ 2 ). Chi square distribution Type: Continuous. σ 2 . Student’s t distribution Type: Continuous. 1 2 Probability density: f (x) = √ e−(x−µ) /2 . invchi2. Applications: Modelling the distribution of the F statistics used in ANOVA and multiple linear regression.n2 ). σ 2 . Mean/Variance: µ. [ST-APP] Summary of probability distrubutions / 2 2014–09–01 . Stata functions: normalden. ttail. Applications: Modelling the distribution of the t statistics used for testing means and regression coefficients. Stata functions: chi2. Notation: F (n1 . invF. such as the Pearson χ2 statistic of contingency tables and the likelihood ratio statistics in many types of regression. ¡ ¢ n /2 n /2 Γ n1 /2 + n2 /2 n1 1 n2 2 xn1 /2−1 Probability density: f (x) = ¡ ¢ ¡ ¢ ¡ ¢(n1 +n2 )/2 . invchi2tail. (nπ)1/2 Γ(n/2) n Parameters: n (degrees of freedom). invFtail. in particular the χ2 statistics of structural equation modelling. Stata functions: Fden. Notation: N (µ. σ 2 = 2n. F distribution Type: Continuous. n2 (degrees of freedom). σ 2 = n/(n − 2) (for n > 2). Applications: Modelling the (asymptotic) distribution of many test statistics. Mean/Variance: µ = 0. Mean/Variance: µ = 0. Notation: χ2 (n). Notation: t(n). Stata functions: tden. Ftail. F. x > 0. σ 2 = n/(n − 2) (for n > 2). ¡ ¢ µ ¶(n+1)/2 Γ (n + 1)/2 x2 Probability density: f (x) = 1+ . 1 Probability density: f (x) = n/2 xn/2−1 ex/2 . Γ n1 /2 Γ n2 /2 n2 + n1 x Parameters: n1 . chi2tail. 2 Γ(n/2) Parameter: n (degrees of freedom). 2π Parameters: µ. normal. Mean/Variance: µ = n.