Sampling Distribution of a Sample Proportion
Lecture 26 Section 8.4 Robb T. Koether
Hampden-Sydney College
Mon, Mar 5, 2012
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
1 / 35
Outline
Computing the Sampling Distribution of p
The Central Limit Theorem for Proportions
Applications
Assignment
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
2 / 35
Outline
Computing the Sampling Distribution of p
The Central Limit Theorem for Proportions
Applications
Assignment
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
3 / 35
Sampling Distributions
Denition (Sampling Distribution of a Statistic)
The sampling distribution of a statistic is the distribution of values of that statistic over all possible samples of a given size n from the population. We may sample with or without replacement. For our purposes, it will be simpler to sample with replacement.
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
4 / 35
The Sample Proportion
for sample sizes of We will work out the sampling distribution for p 1, 2, and 3. for sample Then I will show you the sampling distribution for p sizes of 4, 5, and 10.
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
5 / 35
Example
Suppose that 45% of all people approve of President Obamas performance. Suppose that we select one person at random. We may diagram the 2 possibilities.
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
6 / 35
Example
Y 0.45
0.55 N
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
7 / 35
Example
Now we take a sample of 2 people, sampling with replacement. . Find the sampling distribution of p
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
8 / 35
Example
Y YY 2 (0.45)2 = 0.2025
0.45 Y 0.45 0.55
YN
1 (0.45)(0.55) = 0.2475
0.45 0.55 N 0.55
NY
1 (0.55)(0.45) = 0.2475
NN
0 (0.55)2 = 0.3025
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
9 / 35
Example
Let x be the number of people (out of 2) who strong disapprove of President Obamas performance. is The probability distribution of p ) p P (p 0 1/2 1 0.2025 0.4950 0.3025
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
10 / 35
Example
Now we take a sample of 3 people, sampling with replacement. . Find the sampling distribution of p
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
11 / 35
Example
0.45 0.45 Y 0.45 0.55 N 0.55 0.45 0.45 0.55 N 0.45 0.55 N 0.55 Y 0.55 0.45 Y 0.55
YYY
3 (0.45)3 = 0.0911
N Y
YYN YNY
2 (0.45)2(0.55) = 0.1114 2 (0.45)(0.55)(0.45) = 0.1114
N Y
YNN NYY
1 (0.45)(0.55)2 = 0.1361 2 (0.55)(0.45)2 = 0.1114
N Y
NYN NNY
1 (0.55)(0.45)(0.55) = 0.1361 1 (0.55)2(0.45) = 0.1114
NNN
0 (0.55)3 = 0.1664
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
12 / 35
Example
be the sample proportion of people who strong disapprove Let p of President Obamas performance. is The sampling distribution of p ) p P (p 0 1/3 2/3 1 0.1664 0.4084 0.3341 0.0911
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
13 / 35
Samples of Size n = 4
If we sample 4 people, then the sampling distribution of the sample proportion is ) P (p p 0 0.0915 1/4 0.2995 2/4 0.3675 3/4 0.2005 1 0.0410
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
14 / 35
Samples of Size n = 5
If we sample 5 people, then the sampling distribution of the sample proportion is ) p P (p 0 0.0503 1/5 0.2059 2/5 0.3369 3/5 0.2757 4/5 0.1128 1 0.0185
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
15 / 35
Samples of Size n = 6
If we sample 6 people, then the sampling distribution of the sample proportion is ) p P (p 0 0.0277 1/6 0.1359 2/6 0.2780 3/6 0.3032 4/6 0.1861 5/6 0.0609 1 0.0083
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
16 / 35
Samples of Size n = 10
If we sample 10 people, then the sampling distribution of the sample proportion is
p 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 ) P (p 0.0025 0.0207 0.0763 0.1665 0.2384 0.2340 0.1596 0.0746 0.0229 0.0042 0.0003
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
17 / 35
Outline
Computing the Sampling Distribution of p
The Central Limit Theorem for Proportions
Applications
Assignment
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
18 / 35
The Central Limit Theorem for Proportions
Theorem (The Central Limit Theorem for Proportions)
For any population and any sample size, the sampling distribution has the following mean and standard deviation: of p p = p p = p (1 p ) . n
is approximately Furthermore, the sampling distribution of p normal, provided n is large enough.
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
19 / 35
The Central Limit Theorem for Proportions
The Sample Size
n is large enough if np 5 and n(1 p) 5. If n is small, then we have to work out the distribution by hand.
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
20 / 35
Outline
Computing the Sampling Distribution of p
The Central Limit Theorem for Proportions
Applications
Assignment
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
21 / 35
Applications
Suppose that 60% of all high-school students own a cell phone. If we survey 3 high-school students, how likely is it that we will nd that at least 2 of them own a cell phone?
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
22 / 35
Applications
Suppose that 60% of all high-school students own a cell phone. If we survey 150 high-school students, how likely is it that we will nd that at least 65% of them own a cell phone?
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
23 / 35
Applications
is normal with If p = 0.60 and our sample size is n = 150, then p = 0 . 60 and mean p p = (0.60)(0.40) = 0.0016 = 0.04. 150
0.65. We want to know the probability that p
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
24 / 35
Applications
10
0.55
0.60
0.65
0.70
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
25 / 35
Applications
10
0.55
0.60
0.65
0.70
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
26 / 35
Applications
is greater than 0.65 is The probability that p normalcdf(.65,E99,.60,.04) = 0.1056.
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
27 / 35
Applications
What if our sample size were 600 instead of 150? is normal with mean p Then p = 0.60 and p = (0.60)(0.40) = 0.0004 = 0.02. 600
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
28 / 35
Applications
20
15
10
0.55
0.60
0.65
0.70
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
29 / 35
Applications
20
15
10
0.55
0.60
0.65
0.70
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
30 / 35
Applications
is greater than 0.65 is The probability that p normalcdf(.65,E99,.60,.02) = 0.0062.
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
31 / 35
Guessing on a Test
Example (Guessing on a Test)
A student takes a math placement test with 25 multiple-choice questions.
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
32 / 35
Guessing on a Test
Example (Guessing on a Test)
A student takes a math placement test with 25 multiple-choice questions. Each question has 5 choices.
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
32 / 35
Guessing on a Test
Example (Guessing on a Test)
A student takes a math placement test with 25 multiple-choice questions. Each question has 5 choices. If a student guesses at each answer, what will his score most likely be?
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
32 / 35
Guessing on a Test
Example (Guessing on a Test)
A student takes a math placement test with 25 multiple-choice questions. Each question has 5 choices. If a student guesses at each answer, what will his score most likely be? If a student scores 10 out of 25, can we be sure that he did not guess at all 25 answers?
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
32 / 35
Guessing on a Test
Example (Guessing on a Test)
If he guesses at all 25 answers, then the probability of getting any one answer correct is p = 0.20. . The proportion that he actually gets correct is p
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
33 / 35
Guessing on a Test
Example (Guessing on a Test)
If he guesses at all 25 answers, then the probability of getting any one answer correct is p = 0.20. . The proportion that he actually gets correct is p is normal with mean 0.20 and standard The distribution of p (0.20)(0.80) deviation = 0.04. 25
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
33 / 35
Guessing on a Test
Example (Guessing on a Test)
If he guesses at all 25 answers, then the probability of getting any one answer correct is p = 0.20. . The proportion that he actually gets correct is p is normal with mean 0.20 and standard The distribution of p (0.20)(0.80) deviation = 0.04. 25 is almost If we allow three standard deviations each way, then p certainly at least 0.08 and at most 0.32.
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
33 / 35
Guessing on a Test
Example (Guessing on a Test)
If he guesses at all 25 answers, then the probability of getting any one answer correct is p = 0.20. . The proportion that he actually gets correct is p is normal with mean 0.20 and standard The distribution of p (0.20)(0.80) deviation = 0.04. 25 is almost If we allow three standard deviations each way, then p certainly at least 0.08 and at most 0.32. That represents from 2 to 8 correct answers.
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
33 / 35
Outline
Computing the Sampling Distribution of p
The Central Limit Theorem for Proportions
Applications
Assignment
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
34 / 35
Assignment
Homework
Read Sections 8.1 - 8.2, pages 499 - 508. Exercises 7 - 14, page 526.
Robb T. Koether (Hampden-Sydney College) Sampling Distribution of a Sample Proportion
Mon, Mar 5, 2012
35 / 35
