Scribd
Upload
0 views

Sampling Distributuion

Uploaded by

ella.chavez
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
0 views

Sampling Distributuion

Uploaded by

ella.chavez
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

46 Y.

LIU

6. Sampling distribution

6.1. Introduction to Sampling Distributions.


Example 6.1. Toss a coin n times. Take X1 to be 1 if 1st flip is head, and 0 if tail. Take
Xi to be 1 if ith flip is head, and 0 if tail. What is the average value? In other words, what
is
X1 + · · · + Xn
X̄ = ?
n
X1 + X2 X1 + X2 + X3 X1 + X2 + · · · + X10
e.g.
2 3 10
(These are called the sample mean, or average of the sample.) We know that the number of
head will be roughly n/2 (heads and tails are equally likely to happen). So the numerator is
roughly n/2, and the average is 1/2.

6.2. Sampling Distribution of the Mean. Toss a coin two times, and assign value 1 to
a head and assign 0 to a tail. Then the average/sample mean could be 0, 1/2, 1.

toss a coin two times again. We will have an average with the same possible values and
same distribution.
..
.

Repeat this experiment many times. what is the average of the average value? In other
word, what is the mean of the sample mean? Answer: µX̄ = µ = 1/2.

What is the variance of the sample mean? (We would expect that the fluctuation/variance
is getting smaller as the size of the sample getting larger.) the variance of the sample mean
X1 +···+Xn 2
2
n
can be calculated: X̄ = n . Here is the standard deviation of the population.

6.3. Central Limit Theorem. Now we know the mean and variance of the sample mean.
What about the distribution of the sample mean? (Recall the bell shape distribution.)

We already encountered this problem while approximating a binomial distribution. In


fact, the bell shape distribution is observed for many other distributions.

The central limit theorem states that: Given a variable with a mean µ and a variance 2 ,
the distribution of the sampling mean approaches a normal distribution with a mean of µ
and a variance of 2 /n as n, the sample size, increases.

Note: A bell shape distribution comes up when repeating an experiment many times. Not
only binomial. This is the so-called CLT.
P
Flip a coin. Each outcome is a variable Xi . The number of head is X.
LECTURE NOTES FOR MATH 21 - TUFTS UNIVERSITY 47

6.4. Sampling Distribution of p. Tossing a biased coin. head shows up with probability
p and tail with probability 1 p. p = 0.4 for instance. Let Xi be the outcome value of ith
toss, such that Xi = 1 if the outcome is a head and Xi = 0 if a tail.

According to previous section, the mean of the sample


p mean is p, variance of sample mean:
pq/n. Standard deviation of sample mean: p := pq/n.
q
(0.4)·(0.6)
If sample size is 10, then p = 10
.

The sampling distribution of X̄ is approximately normally distributed if n is fairly large


and p is not close to 0 or 1. A rule of thumb is that the approximation is good if both np
and nq are greater than 10.

You might also like