Discrete Range of X Expected Variance

Distributions Value

Bernoulli 𝑝(𝑥) = 𝑝 𝑥 (1 − 𝑝)1−𝑥 𝑥 = 0,1 𝑝 𝑝(1 − 𝑝)

Binomial 𝑛 𝑥 = 0,1, … , 𝑛 𝑛𝑝 𝑛𝑝𝑞

𝑝(𝑥) = ( ) 𝑝 𝑥 𝑞 𝑛−𝑥
𝑥
𝑞 =1−𝑝

Geometric 𝑝(𝑥) = (1 − 𝑝)𝑥−1 𝑝 𝑥 = 1,2, … 1 1−𝑝

𝑝 𝑝2

Hypergeometric 𝑁1 𝑁 𝑥 ≤ 𝑛, 𝑥 ≤ 𝑁1 , 𝑁1 𝑁1 𝑁2 𝑁 − 𝑛
()( 2 ) 𝑛( ) 𝑛( )( )( )
𝑝(𝑥) = 𝑥 𝑛 − 𝑥 𝑁 𝑁 𝑁 𝑁−1
𝑁 𝑛 − 𝑥 ≤ 𝑁2
( )
𝑛
Negative 𝑥−1 𝑟 𝑥 = 𝑟, 𝑟 + 1, 𝑟 + 2, . .. 1 𝑟(1 − 𝑝)
𝑝(𝑥) = ( ) 𝑝 (1 − 𝑝)𝑥−𝑟 𝑟( )
Binomial 𝑟−1 𝑝 𝑝2

Poisson 𝑒 −𝜆 𝜆𝑥 𝑥 = 0,1,2, … 𝜆 𝜆
𝑝(𝑥) =
𝑥!
Uniform 1 𝑥 = 1,2, … , 𝑚 𝑚+1 𝑚2 − 1
𝑝(𝑥) =
𝑚 2 12

Continuous Range of X Expected Value Variance

Distributions

Exponential 𝑓(𝑥) = 𝜆𝑒 −𝜆𝑥 𝑥≥0 1 1 2

( )
𝐹(𝑥) = 1 − 𝑒 −𝜆𝑥  𝜆

Normal 1 2 /2𝜎 2 −∞ < 𝑥 < ∞  𝜎2

𝑒 −(𝑥−𝜇)
√2𝜋𝜎 2

Uniform 1 𝑎≤𝑥≤𝑏 ( a + b) (𝑏 − 𝑎)2

f ( x) =
b−a 2 12
Parameter Assumption Confidence interval

𝜇 𝑁(𝜇, 𝜎 2 ) or n large 𝜎
𝑥̄ ± 𝑧𝛼
2 √𝑛
 2 known

𝜇 𝑁(𝜇, 𝜎 2 ) 𝑠
𝑥̄ ± 𝑡𝑛−1,𝛼
2 √𝑛
2
𝜎 unknown

𝜇1 − 𝜇2 𝑁(𝜇1 , 𝜎1 2 )
𝜎1 2 𝜎2 2
(𝑥̄ 1 − 𝑥̄ 2 ) ± 𝑧𝛼 √ +
2 𝑛1 𝑛2
𝑁(𝜇2 , 𝜎2 2 )

𝜎1 2 and 𝜎2 2 known

𝜇1 − 𝜇2 𝑁(𝜇1 , 𝜎1 2 ) 1 1
(𝑥̄ 1 − 𝑥̄ 2 ) ± 𝑡𝑛 𝛼 𝑠𝑝 √ +
1 +𝑛2 −2, 2 𝑛1 𝑛2
N ( 2 ,  2 2 )

where
𝜎1 2 =𝜎2 2 , unknown

(𝑛1 − 1)𝑠1 2 + (𝑛2 − 1)𝑠2 2

𝑠𝑝 = √
𝑛1 + 𝑛2 − 2

𝜇1 − 𝜇2 𝑋1 and 𝑋2normal but dependent 𝑑̄ ± 𝑡 𝑠𝑑

𝛼
𝑛−1,
2 √𝑛

𝑝 𝑏(𝑛, 𝑝)
(𝑝̂ )(1 − 𝑝̂ )
𝑝̂ ± 𝑧𝛼 √
2 𝑛
𝑛 is large
𝑥
𝑝̂ =
𝑛

𝑝1 − 𝑝2 𝑏(𝑛1 , 𝑝1 )
(𝑝̂1 )(1 − 𝑝̂1 ) (𝑝̂2 )(1 − 𝑝̂ 2 )
𝑏(𝑛2 , 𝑝2 ) (𝑝̂1 − 𝑝̂2 ) ± 𝑧𝛼 √ +
2 𝑛1 𝑛2
𝑥1 𝑥2
𝑝̂1 = , 𝑝̂ 2 =
𝑛1 𝑛2
CORRELATION AND REGRESSION

2 2
𝑆𝑃XY
𝑛 𝑛 𝑛 𝑛
1 1 𝑛 𝑛 𝑛
𝑆𝑆𝑋 = ∑ 𝑥𝑖2 − (∑ 𝑥𝑖 ) 𝑆𝑆𝑌 = ∑ 𝑦𝑖2 − (∑ 𝑦𝑖 ) 1
𝑛 𝑛 = ∑ 𝑥𝑖 𝑦𝑖 − (∑ 𝑥𝑖 ) (∑ 𝑦𝑖 )
𝑖=1 𝑖=1 𝑖=1 𝑖=1 𝑛
𝑖=1 𝑖=1 𝑖=1

a = y − bx
𝑆𝑃XY 𝑆𝑃XY
𝑏= 1 n 1 n  𝑟=
𝑆𝑆𝑋 =  yi − b   xi  √(𝑆𝑆𝑋 )(𝑆𝑆𝑌 )
n i =1  n i =1 

𝑏−𝛽 𝑆𝑆𝐸 𝑠
𝑡𝑛−2 = 𝑠 where 𝑠 2 = 𝑛−2 and 𝑆𝑆𝐸 = 𝑆𝑆𝑌 − 𝑏𝑆𝑃𝑋𝑌 𝑏 ± 𝑡𝑛−2,𝛼
2 √𝑆𝑆𝑥
√𝑆𝑆𝑥

1 (𝑥𝑝 − 𝑥̄ )2
𝑦̂ ± 𝑡𝑛−2,𝛼 𝑠 √ +
2 𝑛 𝑆𝑆𝑥

1 (𝑥𝑝 − 𝑥̄ )2
𝑦̂ ± 𝑡𝑛−2,𝛼 𝑠 √1 + +
2 𝑛 𝑆𝑆𝑥