1

Moments and Absolute Moments of the

Normal Distribution
Andreas Winkelbauer
Institute of Telecommunications, Vienna University of Technology
Gusshausstrasse 25/389, 1040 Vienna, Austria
arXiv:1209.4340v2 [math.ST] 15 Jul 2014

email: andreas.winkelbauer@nt.tuwien.ac.at

Abstract

We present formulas for the (raw and central) moments and absolute moments of the normal distribution. We
note that these results are not new, yet many textbooks miss out on at least some of them. Hence, we believe that it
is worthwhile to collect these formulas and their derivations in these notes.

I. I NTRODUCTION

Let X ∼ N (µ, σ 2 ) be a normal (Gaussian) random variable (RV) with mean µ = E{X} and variance σ 2 =
E{X 2 } − µ2 (here, E{·} denotes expectation). In what follows, we give formulas and derivations for E X ν ,


E (X − µ)ν , E |X|ν , and E |X − µ|ν , i.e., for the (raw) moments, the central moments, the (raw) absolute
  

moments, and the central absolute moments. We note that the formulas we present hold for real-valued ν > −1.
The remainder of this text is structured as follows: Section II deals with preliminaries and introduces notation,
particularly regarding some special functions. In Section III we present the results; the corresponding derivations
are given in Section IV.

II. P RELIMINARIES
√ √
We denote the standard deviation by σ = σ 2 . The imaginary unit is j = −1 and z ∗ denotes the complex
conjugate of z. The nonnegative integers are denoted by N0 = N∪{0}. Next, we give the definitions of subsequently
used special functions (cf. [1]).
• Gamma function: Z ∞
Γ(z) , tz−1 e−t dt. (1)
0

• Rising factorial:
Γ(z + n)
zn , (2)
Γ(z)
= z(z + 1) · · · (z + n − 1), n ∈ N0 . (3)
 2

• Double factorial:
r
2z+1  z 
z!! , Γ +1 (4)
π 2
= z · (z − 2) · . . . · 3 · 1, z ∈ N odd. (5)

• Kummer’s confluent hypergeometric functions:

∞
X αn z n
Φ(α, γ; z) , M (α, γ, z) = 1 F1 (α; γ; z) = . (6)
n=0
γ n n!
• Tricomi’s confluent hypergeometric functions:
Γ(1 − γ) Γ(γ − 1) 1−γ
Ψ(α, γ; z) , U (α, γ, z) = Φ(α, γ; z) + z Φ(α − γ + 1, 2 − γ; z). (7)
Γ(α − γ + 1) Γ(α)
• Parabolic cylinder functions:
" √ √ #
ν 1 z2 1 − ν 3 z2
  
ν/2 −z 2 /4 π 2πz
Dν (z) , 2 e
Φ − , ; −
Φ , ; . (8)
Γ 1−ν
2
2 2 2 Γ − ν2 2 2 2

III. R ESULTS

In this section we give formulas for the raw/central (absolute) moments of a normal RV. If not noted otherwise,
these results hold for ν > −1.
• Raw moments:
µ2
  
 ν ν µ
E X = (jσ) exp − 2 Dν −j (9)
4σ σ
" √  2
 √  2
#
π ν 1 µ µ 2π 1 − ν 3 µ
= (jσ)ν 2ν/2
Φ − , ; − 2 + j
Φ , ;− 2 (10)
Γ 1−ν2
2 2 2σ σ Γ − ν2 2 2 2σ
  
 Ψ − ν , 1 ; − µ22 , µ ≤ 0
ν ν/2
= (jσ) 2 ·  2 2 2σ  (11)
 Ψ∗ − ν , 1 ; − µ2 , µ > 0
2 2 2σ2

 σ ν 2ν/2 Γ(√2 ) Φ − ν , 1 ; − µ22 ,

 ν+1  
π 2 2 2σ ν ∈ N0 even
= Γ ( ν
+1)
 2
 . (12)
 µσ ν−1 2(ν+1)/2 √ 1−ν 3 µ
2
π
Φ 2 , 2 ; − 2σ 2 , ν ∈ N 0 odd
• Central moments:
√
π
E (X − µ)ν = (jσ)ν 2ν/2

1−ν

(13)
Γ 2
ν+1


ν ν/2 Γ 2
= (jσ) 2 cos(πν/2) √ (14)
π

ν ν/2−1 Γ ν+1


ν 2
= 1 + (−1) σ 2 √ (15)
π

 σ ν (ν − 1)!!, ν ∈ N even
0
= . (16)
 0, ν ∈ N0 odd
• Raw absolute moments:
ν+1


ν ν/2 Γ µ2

 ν 2 ν 1
E |X| = σ 2 √ Φ − , ;− 2 . (17)
π 2 2 2σ
 3

• Central absolute moments:

ν+1


ν ν ν/2 Γ 2

E |X − µ| = σ 2 √ . (18)
π

IV. D ERIVATIONS

In this section we give derivations for the previously presented results. Below we use the following two identities
which hold for γ ∈ R, ν > −1 (cf. [2, Sec. 3.462]):
Z ∞ √ 
γ

ν −x2 +jxγ −γ 2 /8
(−jx) e dx = 2 πe −ν Dν √ , (19)
−∞ 2
Z ∞  
ν −x2 −xγ γ 2 /8 γ
x e dx = 2 −(ν+1)/2
Γ(ν + 1)e D−ν−1 √ . (20)
0 2
• Raw moments (9):
Z ∞  
 ν 1 ν 1 2
E X = √ x exp − 2 (x − µ) dx (21)
2πσ 2 −∞ 2σ
r
ν 2ν 2
µ√ 
 Z ∞
2 σ µ 
= exp − 2 xν exp −x2 + x 2 dx (22)
π 2σ −∞ σ
µ2
   µ
(19)
= (jσ)ν exp − 2 Dν −j . (23)
4σ σ
• Central moments (13): Follows directly from (9) with Φ(α, γ; 0) = 1 and, hence,
√
π
Dν (0) = 2ν/2 1−ν
. (24)
Γ 2
To obtain (14) from (13) we use the identity [2, Sec. 8.334]
  
1−ν

1+ν π
Γ Γ = . (25)
2 2 cos(πν/2)
Then (15) follows from (14) by noting that
1 + exp(jπν) 1 + (−1)ν
cos(πν/2) = = . (26)
2 exp(jπν/2) 2j ν
• Raw absolute moments (17):
Z ∞  
1 1
E |X|ν = √ |x|ν exp − 2 (x − µ)2 dx

(27)
2πσ 2 −∞ 2σ
r
2ν σ 2ν µ2 µ√  µ√ 
  Z ∞  Z ∞  
ν 2 ν 2
= exp − 2 x exp −x − x 2 dx + x exp −x + x 2 dx
π 2σ 0 σ 0 σ
(28)
r
2ν σ 2ν µ2
 
(20)

= exp − 2 2−(ν+1)/2 Γ(ν + 1) D−ν−1 (µ/σ) + D−ν−1 (−µ/σ) (29)
π 4σ
r
2
2ν ν + 1 1 µ2
   
σ µ Γ(ν + 1)
= exp − 2 Φ , ; 2 (30)
2ν 2σ Γ(ν/2 + 1) 2 2 2σ
r
2
ν 2ν ν + 1 1 µ2
     
2 σ µ ν +1
= exp − 2 Γ Φ , ; 2 (31)
π 2σ 2 2 2 2σ
ν+1


Γ 2
 
ν 1 µ
= σ ν 2ν/2 √2 Φ − , ; − 2 , (32)
π 2 2 2σ
 4

where we have used Kummer’s transformation [2, Sec. 9.212], i.e.,

Φ(α, γ; z) = ez Φ(γ − α, γ; −z) , (33)

in the last step.

• Central absolute moments (18): Follows directly from (17) with Φ(α, γ; 0) = 1.

R EFERENCES
[1] W. Magnus, F. Oberhettinger, and R. P. Soni, Formulas and Theorems for the Special Functions of Mathematical Physics, 3rd ed., B. Eckmann
and B. van der Waerden, Eds. Springer-Verlag, 1966.
[2] I. Gradshteyn and I. Ryzhik, Table of integrals, series, and products, 6th ed., A. Jeffrey and D. Zwillinger, Eds. Academic Press, July
2000.