Chapter 8: Statistical Inference: Estimation for Single Populations 1 Chapter 8 Statistical Inference: Estimation for Single Populations LEARNING OBJECTIVES The overall learning objective of Chapter 8 is to help you understand estimating parameters of single populations, thereby enabling you to: 1. Know the difference between point and interval estimation. 2. Estimate a population mean from a sample mean when σ is known. 3. Estimate a population mean from a sample mean when σ is unknown. 4. Estimate a population proportion from a sample proportion. 5. Estimate the population variance from a sample variance. 6. Estimate the minimum sample size necessary to achieve given statistical goals.
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Chapter 8: Statistical Inference: Estimation for Single Populations 1
Chapter 8 Statistical Inference: Estimation for Single Populations
LEARNING OBJECTIVES The overall learning objective of Chapter 8 is to help you understand estimating parameters of single populations, thereby enabling you to: 1. Know the difference between point and interval estimation. 2. Estimate a population mean from a sample mean when σ is known. 3. Estimate a population mean from a sample mean when σ is unknown. 4. Estimate a population proportion from a sample proportion. 5. Estimate the population variance from a sample variance. 6. Estimate the minimum sample size necessary to achieve given statistical
goals.
Chapter 8: Statistical Inference: Estimation for Single Populations 2
CHAPTER TEACHING STRATEGY
Chapter 8 is the student's introduction to interval estimation and estimation of sample size. In this chapter, the concept of point estimate is discussed along with the notion that as each sample changes in all likelihood so will the point estimate. From this, the student can see that an interval estimate may be more usable as a one-time proposition than the point estimate. The confidence interval formulas for large sample means and proportions can be presented as mere algebraic manipulations of formulas developed in chapter 7 from the Central Limit Theorem. It is very important that students begin to understand the difference between mean and proportions. Means can be generated by averaging some sort of measurable item such as age, sales, volume, test score, etc. Proportions are computed by counting the number of items containing a characteristic of interest out of the total number of items. Examples might be proportion of people carrying a VISA card, proportion of items that are defective, proportion of market purchasing brand A. In addition, students can begin to see that sometimes single samples are taken and analyzed; but that other times, two samples are taken in order to compare two brands, two techniques, two conditions, male/female, etc. In an effort to understand the impact of variables on confidence intervals, it may be useful to ask the students what would happen to a confidence interval if the sample size is varied or the confidence is increased or decreased. Such consideration helps the student see in a different light the items that make up a confidence interval. The student can see that increasing the sample size, reduces the width of the confidence interval all other things being constant or that it increases confidence if other things are held constant. Business students probably understand that increasing sample size costs more and thus there are trade-offs in the research set-up. In addition, it is probably worthwhile to have some discussion with students regarding the meaning of confidence, say 95%. The idea is presented in the chapter that if 100 samples are randomly taken from a population and 95% confidence intervals are computed on each sample, that 95%(100) or 95 intervals should contain the parameter of estimation and approximately 5 will not. In most cases, only one confidence interval is computed, not 100, so the 95% confidence puts the odds in the researcher's favor. It should be pointed out, however, that the confidence interval computed may not contain the parameter of interest.
This chapter introduces the student to the t distribution to estimate population means when σ is unknown. Emphasize that this applies only when the population is normally distributed. The student will observe that the t formula is essentially the same as the z formula and that it is the table that is different. When the population is normally distributed and σ is
known, the z formula can be used even for small samples.
Chapter 8: Statistical Inference: Estimation for Single Populations 3
A formula is given in chapter 8 for estimating the population variance. Here the student is introduced to the chi-square distribution. An assumption underlying the use of this technique is that the population is normally distributed. The use of the chi-square statistic to estimate the population variance is extremely sensitive to violations of this assumption. For this reason, exercise extreme caution is using this technique. Some statisticians omit this technique from consideration. Lastly, this chapter contains a section on the estimation of sample size. One of the more common questions asked of statisticians is: "How large of a sample size should I take?" In this section, it should be emphasized that sample size estimation gives the researcher a "ball park" figure as to how many to sample. The “error of estimation “ is a measure of the sampling error. It is also equal to the + error of the interval shown earlier in the chapter.
CHAPTER OUTLINE
8.1 Estimating the Population Mean Using the z Statistic (σ known). Finite Correction Factor Estimating the Population Mean Using the z Statistic when the Sample Size is Small Using the Computer to Construct z Confidence Intervals for the Mean 8.2 Estimating the Population Mean Using the t Statistic (σ unknown). The t Distribution Robustness Characteristics of the t Distribution. Reading the t Distribution Table Confidence Intervals to Estimate the Population Mean Using the t Statistic
Chapter 8: Statistical Inference: Estimation for Single Populations 4
Using the Computer to Construct t Confidence Intervals for the Mean
8.3 Estimating the Population Proportion Using the Computer to Construct Confidence Intervals for the Population Proportion 8.4 Estimating the Population Variance 8.5 Estimating Sample Size Sample Size When Estimating µ Determining Sample Size When Estimating p
KEY WORDS Bounds Point Estimate Chi-square Distribution Robust Degrees of Freedom(df) Sample-Size Estimation Error of Estimation t Distribution Interval Estimate t Value
Chapter 8: Statistical Inference: Estimation for Single Populations 5
SOLUTIONS TO PROBLEMS IN CHAPTER 8
8.1 a) x = 25 σ = 3.5 n = 60 95% Confidence z.025 = 1.96
x + zn
σ = 25 + 1.96
60
5.3 = 25 + 0.89 = 24.11 < µ < 25.89
b) x = 119.6 σ = 23.89 n = 75 98% Confidence z.01 = 2.33
x + zn
σ = 119.6 + 2.33
75
89.2 = 119.6 ± 6.43 = 113.17 < µ < 126.03
c) x = 3.419 σ = 0.974 n = 32 90% C.I. z.05 = 1.645
x + zn
σ = 3.419 + 1.645
32
974.0 = 3.419 ± .283 = 3.136 < µ < 3.702
d) x = 56.7 σ = 12.1 N = 500 n = 47 80% C.I. z.10 = 1.28
x ± z1−
−N
nN
n
σ = 56.7 + 1.28
1500
47500
47
1.12
−−
=
56.7 ± 2.15 = 54.55 < µ < 58.85
Chapter 8: Statistical Inference: Estimation for Single Populations 6
8.2 n = 36 x = 211 σ = 23 95% C.I. z.025 = 1.96
x ± zn
σ = 211 ± 1.96
36
2 = 211 ± 7.51 = 203.49 < µ < 218.51
8.3 n = 81 x = 47 σ = 5.89 90% C.I. z.05=1.645
x ± zn
σ = 47 ± 1.645
81
89.5 = 47 ± 1.08 = 45.92 < µ < 48.08
8.4 n = 70 σ2 = 49 x = 90.4 x = 90.4 Point Estimate 94% C.I. z.03 = 1.88
x + zn
σ = 90.4 ± 1.88
70
49 = 90.4 ± 1.57 = 88.83 < µ < 91.97
8.5 n = 39 N = 200 x = 66 σ = 11 96% C.I. z.02 = 2.05
x ± z1−
−N
nN
n
σ = 66 ± 2.05
1200
39200
9
11
−−
=
66 ± 3.25 = 62.75 < µ < 69.25 x = 66 Point Estimate
Chapter 8: Statistical Inference: Estimation for Single Populations 7
8.6 n = 120 x = 18.72 σ = 0.8735 99% C.I. z.005 = 2.575 x = 18.72 Point Estimate
x + zn
σ = 18.72 ± 2.575
120
8735.0 = 8.72 ± .21 = 18.51 < µ < 18.93
8.7 N = 1500 n = 187 x = 5.3 years σ = 1.28 years 95% C.I. z.025 = 1.96 x = 5.3 years Point Estimate
8.9 n = 36 x = 3.306 σ = 1.17 98% C.I. z.01 = 2.33
x ± zn
σ = 3.306 ± 2.33
36
17.1 = 3.306 ± .454 = 2.852 < µ < 3.760
Chapter 8: Statistical Inference: Estimation for Single Populations 8
8.10 n = 36 x = 2.139 σ = .113 x = 2.139 Point Estimate 90% C.I. z.05 = 1.645
x ± zn
σ = 2.139 ± 1.645
36
)113(. = 2.139 ± .03 = 2.109 < µ < 2.169
8.11 µ = 27.4 95% confidence interval n = 45 x = 24.533 σ = 5.124 z = + 1.96
Confidence interval: x + zn
σ = 24.533 + 1.96
45
124.5 =
24.533 + 1.497 = 23.036 < µµµµ < 26.030 8.12 The point estimate is 0.5765. n = 41 The assumed standard deviation is 0.1394 99% level of confidence: z = + 1.96 Confidence interval: 0.5336 < µ < 0.6193 Error of the estimate: 0.6193 - 0.5765 = 0.0428
Chapter 8: Statistical Inference: Estimation for Single Populations 9
8.13 n = 13 x = 45.62 s = 5.694 df = 13 – 1 = 12 95% Confidence Interval α/2=.025 t.025,12 = 2.179
n
stx ± = 45.62 ± 2.179
13
694.5 = 45.62 ± 3.44 = 42.18 < µ < 49.06
8.14 n = 12 x = 319.17 s = 9.104 df = 12 - 1 = 11 90% confidence interval α/2 = .05 t.05,11 = 1.796
n
stx ± = 319.17 ± (1.796)
12
104.9 = 319.17 ± 4.72 = 314.45 < µ < 323.89
8.15 n = 41 x = 128.4 s = 20.64 df = 41 – 1 = 40 98% Confidence Interval α/2=.01 t.01,40 = 2.423
n
stx ± = 128.4 ± 2.423
27
6.20 = 128.4 ± 9.61 = 118.79 < µ < 138.01
x = 128.4 Point Estimate
Chapter 8: Statistical Inference: Estimation for Single Populations 10
Chapter 8: Statistical Inference: Estimation for Single Populations 30
8.63 n = 27 x = 4.82 s = 0.37 df = 26 95% CI: t.025,26 = 2.056
27
37.0056.282.4 ±=±
n
stx = 4.82 + .1464
4.6736 < µ < 4.9664 We are 95% confident that µ does not equal 4.50. 8.64 n = 77 x = 2.48 σ = 12 95% Confidence z.025 = 1.96
77
1296.148.2 ±=±
nzx
σ = 2.48 ± 2.68
-0.20 < µ < 5.16
The point estimate is 2.48
The interval is inconclusive. It says that we are 95% confident that the average arrival time is somewhere between .20 of a minute (12 seconds) early and 5.16 minutes late. Since zero is in the interval, there is a possibility that on average the flights are on time.
8.65 n = 560 p̂ =.33 99% Confidence z.005= 2.575
560
)67)(.33(.575.233.
ˆˆˆ ±=⋅±
n
qpzp = .33 ± (2.575) = .33 ± .05
.28 < p < .38
Chapter 8: Statistical Inference: Estimation for Single Populations 31
8.66 p = .50 E = .05 98% Confidence z.01 = 2.33
2
2
2
2
)05(.
)50)(.50(.)33.2(=⋅E
qpz = 542.89
Sample 543 8.67 n = 27 x = 2.10 s = 0.86 df = 27 - 1 = 26 98% confidence α/2 = .01 t.01,26 = 2.479
.0026 < σ2 < .0071 8.69 n = 39 x = 1.294 σ = 0.205 99% Confidence z.005 = 2.575
39
205.0575.2294.1 ±=±
nzx
σ = 1.294 ± (2.575) = 1.294 ± .085
1.209 < µ < 1.379
Chapter 8: Statistical Inference: Estimation for Single Populations 32
8.70 The sample mean fill for the 58 cans is 11.9788 oz. with a standard deviation of .0556 oz. The 99% confidence interval for the population fill is 11.9607 oz. to 11.9970 oz. which does not include 12 oz. We are 99% confident that the population mean is not 12 oz. indicating an underfill from the machine. 8.71 The point estimate for the average length of burn of the new bulb is 2198.217 hours. Eighty-four bulbs were included in this study. A 90% confidence interval can be constructed from the information given. The error of the confidence interval is + 27.76691. Combining this with the point estimate yields the 90% confidence interval of 2198.217 + 27.76691 = 2170.450 < µ < 2225.984. 8.72 The point estimate for the average age of a first time buyer is 27.63 years. The sample of 21 buyers produces a standard deviation of 6.54 years. We are 98% confident that the actual population mean age of a first-time home buyer is between 24.02 years and 31.24 years. 8.73 A poll of 781 American workers was taken. Of these, 506 drive their cars to
work. Thus, the point estimate for the population proportion is 506/781 = .648. A 95% confidence interval to estimate the population proportion shows that we are 95% confident that the actual value lies between .613 and .681. The error of this interval is + .034.