Economic Statistics :

Sampling Distributions : Central Limit Theorem

I. Normal parent population:

If the population is Normally distributed, then the distribution of the sampling
distribution will also be normal. Therefore, we can standardize the means and use the Z
table to approximate the probabilities of certain ranges of the sampling distribution.

A. Standardizing the sampling distribution of the sample mean: Remember that the
2
variance for the sampling distribution of the sample mean is  2  ,
x n
a. The standard deviation of the sampling distribution is called the standard

error of the mean:  
x n
X 
b. Thus the standardization to the Z distribution is: Z 

n
B. Example 1
Suppose we draw a sample of size 36 from a normally distributed population with mean
30 and variance 900.
1. Find the mean and variance of the sampling distribution of sample means.
E( X )=30
Var( X ) = 900/36 = 25

2. Find the probability that the sample mean is between 20 and 40.

20  30 40  30
ZL   2 and ZU   2 and therefore
5 5
P(20 < X < 40) = P(-2< Z <2)=0.9544.

3. Find P(23< X <35).

23  30 35  30
ZL   1.4 and ZU   1 and therefore
5 5
P(23 < X < 35) = P(-1.4< Z <1)=0.4192+0.3413 = 0.7605

II. Central Limit Theorem:

As the sample size increases and with a large enough a sample size the sampling
distribution will become symmetrical (and bell shaped) regardless of the shape of the
parent population. This remarkable fact is known as the central limit theorem. The
theorem implies that whatever the shape of the parent population, distribution of X is
approximately normal in a large sample. The appropriateness of using the Normal as
an approximation of the sampling distribution depends on how skewed the parent
 distribution is. However, if the n is at least 30, then it can be assumed under the
Central Limit Theorem that the sampling distribution is normally distributed.

A. Example 2
Cola, Inc. maintains records regarding the amount of cola in its jumbo bottle. Their
records indicate that the mean amount per bottle is 31.2 ounces and the population
standard deviation is 0.4 ounces. A quality technician randomly selects 42 bottles from
the filling line. The mean amount of cola, X , contained in the bottles is 31.38 ounces.
1. Is this an unlikely result? Is it likely that the process is putting too much soda in
the bottles?
a. First, since the sample size is large distribution of X is approximately
normal.
b. We are finding the probability that we could get a sample mean of at least
31.38. P( X >31.38);
31.38  31.2 0.18
c. Standardize this sample mean: Z    2.91 ; P(z >
0.4 / 42 0.0617
2.91) = 0.5 – 0.4982 = 0.0018
d. Thus it is unlikely, less than 0.2% chance we could select a sample of 42
and get a mean greater than or equal to 31.38, so we can conclude that the
process is putting to much soda in the each bottle.

B. Example 3
The U.S. Census Bureau announced that the median sales price of new houses sold in
February 2008 was $244,100, whereas the mean sales price was $296,400. Assume that
the standard deviation of the prices is $90,000.
1. If you select samples of size n = 2, describe the shape of the sampling
distribution of X .
The distribution will be positively skewed, so the larger density of values will lie
toward the bottom half of the distribution.

2. If you select samples of size n = 100, describe the shape of the sampling
distribution of X .
The distribution will be approximately symmetrical, the shape of a normal
distribution with mean $296,400 and standard deviation $9000

3. If you select a random sample of n = 100, what is the probability that the
sample mean will be less than $300,000?
P( X < 300,000) = P( z < (300,000-296,400)/(90,000/√100)) = P(z < 0.40)
P(z < 0.40) = 0.5 + P(0 <z < 0.40) = 0.5 + 0.1554 = 0.6554

4. If you select a random sample of n = 100, 95% of the means of home values
will be below what value?
 P( X < X 0 ) = 0.95
P (z < z0) = 0.95; z0 = 1.645
1.645 = ( X 0 - 296,400)/ 9000
X 0 = 311,205

C. Example 4: The amount of time university professors devote to their jobs every
week has a mean of 62 hours and a standard deviation of 8 hours. Suppose you
select a sample of 34 university professors.
a. What is the probability that the mean amount of work for those in your
sample is more than 65 hours?
P( X >65);
65  62 3
Standardize this sample mean: Z    2.19 ;
8 / 34 1.372
P(z > 2.19) = 0.5 – 0.4857 = 0.0143

b. 99% of the means for professor hours will be below what value?
Find the Z value: P(Z < 2.33) = 0.99
De-Standardize this Z to the sample mean:
X  62
2.33   (2.33 *1.372)  62  65.197 ;
8 / 34

99% of the means for professor hours will be below 65.197.

c. 90% of the means for professor hours will be in what range?

Find the Z value: Z upper has .05 above it, and Zlower has .05 below it:
P(0 < ZU < 1.645) = 0.45
P(-1.645 < ZL < 0) = 0.45
De-Standardize this Z to the sample mean:
X  62
1.645  U  (1.645 * 1.372)  62  64.257 ;
8 / 34
X  62
 1.645  L  (1.645 * 1.372)  62  59.743 ;
8 / 34
90% of the means for professor hours will be between 59.743 and 64.257