Lecture 6


THE NORMAL DISTRIBUTION

Probability Density Function of a Continious Random variable:
Consider a Histogram where the heights of the bars equal the relative frequencies for each class interval. Let the number of observations in the sample grow indefinitely. Then the number of class intervals can be increased indefinitely, and the width of the class intervals can be made increasingly narrower. In the limit the histogram will be approximated by a smooth curve.

This is called the probability density function.

The total area under the curve must equal 1. The probability that a random variable will assume a value between any two points, a and b, equals the area under the random variable’s probability density function over the interval from a and b.

The Normal Distribution:
The most popular and widely use continuous probability density is Normal density given by the formula:

    F(x) = 1/{(2p )1/2 s } exp( -1/2[(x-m )/s ]2,

where m is the expected value (mean), s the standard deviation, e @ 2.718 - the base of the natural logarithms, and p @ 3.1416.

Normal curves vary in shapes because differences in mean and variance, but all have the following characteristics:

1. Symmetrical and bell-shaped. So its

mean=median=mode.


2. Probability that a value will lie within k standard deviations of the mean is always given by the following rule:
    within one s.d. of the mean = 68.3%
    within two s.d. of the mean = 95.4%
    within three s.d. of the mean= 99.7%
    See Diagram 6.4 (page 193).

3. Location and shape determined entirely by
m and s . So the distribution X with mean m and s.d. s 2 is denoted as

N (m , s 2).

Normal distribution has been found to approximate a wide variety of real life distributions, but not always.

STANDARD NORMAL CURVE

For any normal distribution X which is N(
m , s 2),
   

Z = (X-m )/s

has a standard normal density. So if X (weight of an adult male) is distributed as N(170, 25) then (X-170)/5 is a standard normal variable or variate.

How to use the table of the standard normal distribution: See Table A.2 (Page A8 in Text).

The executives at Liz Claiborne want to know the probability that the height of an adult male lies between 65 and 68 inches.

We know that the height distribution is characterized by N(66, 4). So the two points on N(0, 1) are (65-66)/2 = -0.5 and (68-66)/2 = 1.0.

Now let us find the area under the standard normal curve between –.5 and 1.0.
From the table A.2,

The probability from 0 to 1.0 is .3413;
The probability from 0 to .5 is .1915;

But by symmetry,

Prob (0 to -.5) = Prob (0 to .5) = .1915.

Hence, the total probability from -.5 to 1.0 is .3413 + .1915 = .5328.