Normal Distribution

Junzhe Bao
Department of Biostatistics and Epidemiology

Email: [email protected]
1
 Outline

1 Probability density curve

2 Normal distribution and it’s general features

3 Standard normal distribution

The applications of normal distributions

4

2
 Probability density curve

EXAMPLE：The weight of 1402 pregnant women was measured.

The distribution of weight from 1402 pregnant women f: frequency

i: interval width
（4）

n: sample size（
1402）

3
Probability density curve
Probability density

weight
4
Probability density curve

With the increase of

sample size and decrease
of interval width, the peak
line of histogram will turn
into a smooth curve . The
y-axis represents
probability density, so the
figure is called probability
density curve.

5
 Normal distribution and it’s general features

When the data values are evenly distributed about the mean, high in the middle and low on
both sides, showing a bell shape, which is similar to a normal curve, such distribution could
be called a normal distribution.

The area under normal curve is 1.

6
 Normal distribution

The characteristics of normal curve were revealed by Abraham de Moivre

and were developed by karl Friedrich Gauss. The normal distribution is also
called Gaussian distribution.

Abraham de Moivre Karl Friedrich Gauss

(1777 - 1855)
(1667-1754) 7
 Normal distribution and it’s general features

Function expression of normal distribution:

1 −( X −  ) 2 / 2 2
f (X ) = e ,−  X  +
 2
f(X): the probability density.
σ : standard deviation.
π ≈ 3.14 (≈ means “is approximately equal to”)
e ≈ 2.718
μ : mean value.
σ2 ：variance

X ~ N(μ, σ2 )
8
 Normal distribution and it’s general features

The shapes and locations of normal distributions with different values of μ and σ

μ: location; σ: shape

9
 Normal distribution and it’s general features

a1 a2
x
a μ

The calculation of the probability of the values falling into the area smaller than a specific value
(such as “a”) or falling into a range (such as “a1~a2”) is equal to calculate the corresponding
area under the normal distribution curve.

10
 Normal distribution and it’s general features

General features of normal distribution:

➢ The distribution is centered at mean, bell-shaped.

➢ The distribution is unimodal and symmetric.

➢ The distribution has two parameters. One is μ ( location parameter), the other is σ
(variability parameter).

➢ The curve never touches the x axis, and the total area under a normal distribution
curve is equal to 1.
11
 Outline

1 Probability density curve

2 Normal distribution and it’s general features

3 Standard normal distribution

The applications of normal distributions

4

12
 Standard normal distribution

➢ Since each normally distributed variable has its own mean and standard
deviation, as stated earlier, the shape and location of these curves will vary.

➢ In practical applications, then, you would have to have a table of areas under
the curve for each variable, so as to predict the probability of a specific value
or range.

➢ To simplify this situation, statisticians use what is called the standard normal
distribution, with a mean of 0 and a standard deviation of 1.

13
 Standard normal distribution

All normally distributed variables can be transformed into the standard normally
distributed variable :
1 −( X −  ) 2 / 2 2
f (X ) = e ,−  X  +
 2
X −
Z=


1 −Z 2 / 2
 (Z ) = e ,−  Z  +
2
14
 Standard normal distribution

X −
Normal
Z = Standard normal
distribution  distribution

 =1

 Z
x =
15
 Standard normal distribution
Example 1：
Suppose that X~N(μ, σ2 ), calculate the probability of the values of X falling into
the range between μ - 1.96σ and μ + 1.96σ.

Step 1: Standard normally distributed transformation:

(μ − 1.96σ)−μ (μ + 1.96σ)−μ
Z1= = −1.96; Z2= = 1.96.
σ σ
P(μ − 1.96σ < X < μ + 1.96σ)=P(Z<1.96) - P(Z<-1.96) =[1- P(Z<-1.96)]- P(Z<-
1.96) =1-2* P(Z<-1.96)=1-2*0.025=0.95=95%

16
The area
for
Negative
infinity to
Z (-∞ ~ Z)

17
 Standard normal distribution

Example 2：
Supposed that X~N(μ, σ2 ), please calculate the probability of the values of X
falling into the range between μ – 2.58σ and μ + 2.58σ, as well as between μ –
1.64σ and μ + 1.64σ.

18
Standard normal distribution

the standard normal curve and it’s area distribution

P(μ−1.96σ <X< μ+1.96σ)=P(-1.96<Z<1.96)=0.95=95.00%
P(μ−1.64σ <X< μ+1.64σ)=P(-1.64<Z<1.64)=0.90=90.00%
P(μ−2.58σ <X< μ+2.58σ)=P(-2.58<Z<2.58)=0.99=99.00% 19
 Standard normal distribution

Example 3：
Suppose that diastolic blood pressure (X) in hypertensive
women centers is normally distributed, with the mean value
of 100mmHg and standard deviation of 16mmHg. Please
calculate P(X<90mmHg), P(X>124mmHg), as well as
P(90mmHg < X <124mmHg).

20
 Standard normal distribution

➢ Transform X into a standard normal variable. μ =100 and σ=16.

(X1)−μ 90−100 (X2)−μ 124−100

= = −0.625 σ =
Z1= ≈ -0.63; Z2= = 1.50
σ 16 16

➢ We wish to find the area when Z<-0.63, Z>1.50, as well as -0.63<Z<1.50

-0.63 0 1.50 21
The area
for
Negative
infinity to
Z (-∞ ~ Z)
P(Z>1.5)=
1-P(Z<1.5)
=
1-0.9332
=
0.0668=
6.68%=
P(Z<-1.5)

22
 Standard normal distribution

Example 3：

P(X<90mmHg)=P(Z<-0.63)=0.2643=26.43%

P(X>124mmHg)=P(Z>1.50)=P(Z<-1.50)=0.0668=6.68%

P(90mmHg < X <124mmHg)=P(-0.63<Z<1.5)=P(Z<1.5)-P(Z<-0.63)=0.9332-

0.2643=0.6689=66.89%

23
 Outline

1 Probability density curve

2 Normal distribution and it’s general features

3 Standard normal distribution

The applications of normal distributions

4

24
 The applications of normal distributions

➢ Estimate reference range

➢ Estimate distribution of frequency

25
 The applications of normal distributions

➢ Estimate reference range

Reference range usually describes the variations of a

measurement or value in healthy individuals. It is a basis
for a physician or other health professional to interpret a set
of results for a particular patient.

26
The applications of normal distributions

27
 The applications of normal distributions

➢ Definition of reference range

----The prediction interval between which 95%

(90%, 99%) of values of a reference group fall
into, in such a way that 5% (10%, 1%) of the time
a sample value will be beyond this range.

28
 The applications of normal distributions

Two-sided values

2.5% of the time a sample value will be less than the lower limit
of this interval, and 2.5% of the time it will be larger than the
upper limit of this interval.

Examples:

SBP, DBP, Weight……

29
 The applications of normal distributions

➢ One-sided values
In some cases, only one side of the range is usually of interest, such
as with markers of pathology including cancer antigen, where it is
generally without any clinical significance to have a value below
what is usual in the general population.

30
 The applications of normal distributions

➢ Example 4
Suppose the concentration of hemoglobin in 120 health women is
normally distributed with a mean of 117.4g/L and standard
deviation 10.2g/L. What is the 95% reference range of hemoglobin?

31
 The applications of normal distributions
Because it is abnormal whether hemoglobin is too higher or too lower. We should use
two-sided range.

In practical applications, the sample mean(𝑥)ҧ is often used to represent the population
mean (μ), and the sample standard deviation (s) is the population standard deviation (σ).

x + 1.96 s = 117 .4 + 1.96  10 .2 = 137 .39 ( g / l )

x − 1.96 s = 117 .4 − 1.96  10 .2 = 97 .41( g / l )
So 95% reference range of hemoglobin is from 97.41 to 137.39g/L
32
The applications of normal distributions

the standard normal curve and it’s area distribution

P(μ−1.96σ <X< μ+1.96σ)=P(-1.96<Z<1.96)=0.95=95.00%
P(μ−1.64σ <X< μ+1.64σ)=P(-1.64<Z<1.64)=0.90=90.00%
P(μ−2.58σ <X< μ+2.58σ)=P(-2.58<Z<2.58)=0.99=99.00% 33
 The applications of normal distributions

➢ Example 5
Measure the vital capacity of 110 health adult men and get a mean
of 4.2/L and standard deviation of 0.7/L.
What is 95% reference range of vital capacity ?

34
 The applications of normal distributions

Because it is abnormal when vital capacity is too lower.

We should use one-sided range.

x − 1.64 s = 4.2 − 1.64  0.7 = 3.052（L)

So 95% reference range of vital capacity is not less than

3.052 L

35
 The applications of normal distributions

X + 1.64S
One-sided
normal distribution X −1.64S
Two-sided X  1.96S
Distribution Plot Distribution Plot
Normal, Mean=0, StDev=1 Normal, Mean=0, StDev=1

0.4 0.4

0.3 0.3
Density

Density
0.2 0.2

95% 95%
0.1 0.1

0.025 0.025 0.05

0.0 0.0
-1.960 0 1.960 -1.645 0
X X
36
 The applications of normal distributions

Reference ranges can also be established directly from the percentile

method, such as the 2.5th and 97.5th percentile (<95th or >5th for
one-sided) of the measurements in the reference group.

If the distribution is asymmetric or the type of distribution is

unknown, the percentile method should be used to calculate the
reference value range; if the distribution is a normal distribution, the
95% reference value range of the normal distribution should be used.

37
 The applications of normal distributions

One-sided X + 1.64S
X −1.64S
normal distribution
two-sided X  1.96S
One-sided <P Or >P
95 5
Percentile method

two-sided P 2.5~P 97.5

38
 The applications of normal distributions

➢ Estimate frequency distribution

Example 6
Suppose that the scores on CET-4 are normally distributed with a mean of 70 and

standard deviation of 6.

What is the probability that a randomly selected score is larger than 80?
 The applications of normal distributions

Transform X into a standard normal variable.

X − 80 − 70
Z= = = 1.67
 6
P(X>80)=P(Z>1.67)=0.0475
So the probability that a randomly selected score is larger than 80
is 0.0475. that is to say, the scores of 4.75% people is higher than
80
 The applications of normal distributions
Example 7
An instructor is administering a final examination. She tells her class that she will

give an A grade to the 10% of the students who earn the highest marks. Past

experience with the same examination has yielded grades that are normally

distributed with a mean of 70 and SD of 10.

Probability calculate x value

If the same present class runs true to form, what numerical score would a student

need to earn an A grade?

 The applications of normal distributions

Step one
P (One-sided area)=0.1 Z?
X − 80 − 70
Z= = = 1.67
 6

10%
43
 The applications of normal distributions

Step two
Z=1.28 x?

x−x x − 70
Z= = 1.28
s 10

x=82.8
Thanks！

45