STATISTICS FOR MANAGEMENT-II LEARNING OBJECTIVES: To learn how to estimate certain characteristics of a population from sample. To learn the strengths and shortcomings of point estimates and interval estimates. To calculate how accurate our estimates really are. To learn how to use the t- distribution to make interval estimates. INSTRUCTOR’S NAME 1 CHAPTER-7 Statistical Inference: The process of drawing inferences about population on the basis of information contained in sample (taken from population) is called statistical inference. Statistical inference is divided in to two major areas, 1- Estimation of parameters. 2- Testing of hypothesis In this chapter we deal only with estimation of parameters. It is a process by which we obtained an estimate of the unknown population parameter by using samples (taken from population). For example, we may estimate the population mean (µ) by calculating the mean of sample (taken from population). Sample statistic used to estimate population parameter, is called estimator. For example x is estimator of µ , S is estimator of 𝜎 etc. 2 1- Unbiasedness 2- Efficiency 3- Consistency 4- Sufficiency 1-Unbiasedness: An estimator is unbiased if mean of its sampling distribution is equal to the p population parameter. For example sample mean x is unbiased estimator of population mean µ. because mean of sampling distribution of sample mean is equal to population mean i.e. E ( x ) = µ 2- Efficiency: An estimator is efficient if it has smaller standard error than any other estimator of that population parameter. For example sample mean 𝑥 is efficient estimator of population median than sample median because in large samples, the sample mean has small standard error than sample median. 3 3-Consistency: An estimator is consistence if as sample size is increases, the value of that estimator becomes very close to the value of population parameter. For example sample mean 𝑥 is consistent estimator of the population median Because as sample size increases the value of sample mean becomes very close to the population median. 4-Sufficiency : An estimator is sufficient if it uses all information that is contained in sample about parameter. Any estimator that is not computed from all values in samples is not sufficient. An estimate is the numerical value of an estimator .It is used to estimate unknown parameter. There are two types of estimate, 1-Point Estimate 2-Interval Estimate 1-Point estimate: A single numerical value of an estimator is called point estimate. Here are some estimators used to get point estimate. x = i) ii) 𝑆2 = S ∑𝑥 𝑛 ∑( 𝑥− x =√ )2 𝑛 ∑( 𝑥− x 𝑛 )2 or or ∑ 𝑥2 𝑛 ∑ 𝑥2 √ 𝑛 - ( ∑𝑥 2 ) 𝑛 − ( ∑𝑥 𝑛 {𝑏𝑖𝑎𝑠𝑒𝑑 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑜𝑟 } )2 4 iii) S2 = S iv) ∑( 𝑥− x )2 𝑛−1 =√ ∑ ( 𝑥− x or )2 𝑛−1 ∑ 𝑥2 𝑛−1 - 𝑛( x )2 𝑛−1 ∑ 𝑥2 or √ − 𝑛−1 {𝑢𝑛𝑏𝑖𝑎𝑠𝑒𝑑 𝑒𝑠𝑡𝑖𝑚𝑎𝑡𝑜𝑟} 𝑛( x )2 𝑛−1 𝑋 P = 𝑛 Ex: Find point estimates of the mean, variance and standard deviation of population from which the given sample was drawn. 45 32 37 46 39 36 41 48 36 Solution: 𝑥 𝑥− x 45 32 37 46 39 36 41 48 36 5 -8 -3 6 -1 -4 1 8 -4 360 a) n ( 𝑥 − x )2 25 64 9 36 1 16 1 64 16 232 = 9 ∑ 𝑥 = 360 x = ∑𝑥 𝑛 5 360 = S2 = b) = c) 9 = 40 ∑( 𝑥− x )2 𝑛−1 232 9−1 232 = ∑ ( 𝑥− x = √ S = 29 8 )2 𝑛−1 = √ 232 = √ 9−1 232 8 =√29 = 5.39 Ex : (EX Sc 7-1 pg 352) Find point estimates of the mean and variance and standard deviation of population from which the given sample was drawn. Solution: 𝑥 𝑥− x 8.8 14.0 21.3 7.9 12.5 20.6 16.3 14.1 13.0 128.5 a) -5.5 -0.3 7.0 -6.4 -1.8 6.3 2.0 -0.2 -1.3 ( 𝑥 − x )2 30.25 0.09 49.00 40.96 3.24 39.69 4.00 0.04 1.69 168.96 n = 9 ∑𝑥 = 128.5 6 x = = b) S2 = = c) S 𝑛 128.5 = 14.3 9 ∑( 𝑥− x )2 𝑛−1 168.96 = 9−1 ∑ ( 𝑥− x = √ 168.96 8 = 21.12 )2 𝑛−1 = √ H.W: ∑𝑥 168.96 9−1 = √ 168.96 8 =√21.12 = 4.60 D0 EX 7-11 (pg 353) 2-Interval Estimate: A range of values to estimate an unknown parameter is called interval estimate. Confidence Interval: Confidence interval is the range of the estimate. We express the confidence interval in terms of standard error rather than in numerical values. i.e. x + 𝑧 σx̄ and x − 𝑧 σx̄ 7 Confidence Limit: The lower limit ( x − 𝑧 σx̄ ) and upper limit ( x + 𝑧 σx̄ ) of confidence interval are called confidence limits. Confidence Level: The probability associated with interval estimate is called confidence level. EX : (EX Sc 7-5 pg 360) Express the lower and upper limits of the confidence interval for the given levels in terms of x and σx̄ . a) 54% b) 75% c) 94% d) 98% Solution: a) P(z) = 54% = = 54 100 0.54 2 = 0.54 = 0.2700 = 0.74 Upper limit of confidence interval = x + 𝑧 σx̄ = x + 0.74 σx̄ Lower limit of confidence interval = x − 𝑧 σx̄ = x − 0.74 σx̄ Or Confidence limits = x ± 0.74σx̄ b) P(z) = 75% = = 75 100 0.75 2 = 0.75 = 0.3750 = 1.15 8 Upper limit of confidence interval = x + 𝑧 σx̄ = x + 1.15 σx̄ Lower limit of confidence interval = x − 𝑧 σx̄ = x − 1.15 σx̄ Or Confidence limits = x ± 1.15σx̄ c) P(z) = 94% = = 94 100 = 0.94 0.94 = 0.4700 = 1.88 2 Upper limit of confidence interval = x + 𝑧 σx̄ = x + 1.88 σx̄ Lower limit of confidence interval = x − 𝑧 σx̄ = x − 1.88 σx̄ or =x Confidence limits ± 1.88σx̄ d) P(z) = 98% = = 98 100 = 0.98 0.98 2 = 0.4900 = 2.33 Upper limit of confidence interval = x + 𝑧 σx̄ = x + 2.33 σx̄ Lower limit of confidence interval = x − 𝑧 σx̄ = x − 2.33 σx̄ Or Confidence limits H.W: = x ± 2.33σx̄ D0 EX 7-25 (a, b, c) (pg 361) 9 𝜎 Known , n>30 , population infinite EX : (EX 7-28 PG 365) From sample of 250 from a population with known 𝜎 of 13.7, the sample mean is found to be 112.4, Find a 95% confidence interval for the mean. Solution: a) n = 250 𝜎 = 13.7 x = 112.4 σx̄ = 𝜎 √𝑛 13.7 = P(z) = 0.87 √250 = 95% = = 95 100 0.95 = 0.95 = 2 x 0.4750 = 1.96 ± 𝑧σx̄ x − 𝑧 σx̄ x + 𝑧 σx̄ 112.4 – 1.96(0.87) 112.4-1.71 110.69 So 95% confidence interval 𝜎 Known , 112.4+1.96(0.87) 112.4+1.71 114.11 = (110.69 , 114.11) n> 30 , population finite EX: From a sample of 200 from population of 1000 with known 𝜎 of 1.08, the sample mean is found to be 69.2, Find 95% confidence interval for the mean. 10 Solution: n N = 200 = 1000 x = 69.2 𝜎 = 1.08 σx̄ x √ √𝑛 1.08 = P (z) 𝜎 = √200 x √ 𝑁−𝑛 𝑁−1 1000−200 1000−1 X √ 800 = 0.08 = = 0.08 x 0.89 0.07 999 = 95% = 95 100 = 0.95 2 = 0.95 = 0.4750 = 1.96 x x − 𝑧 σx̄ ± 𝑧σx̄ 69.2 – 1.96(0.07) 69.2- 0.14 69.06 x + 𝑧 σx̄ 69.2+ 1.96(0.07) 69.2 + 0.14 69.34 So 95% confidence interval = (69.06 , 69.34) H.W: Do EX 7-28 (b) pg 365. 11 𝜎 n > 30 Unknown , , population infinite EX : A sample of 100 is taken from population with sample mean 182 and sample standard deviation 17.29.Calculate 99% confidence interval for the mean. Solution: n = 100 x = 182 𝜎̂ = s =17.29 𝜎 ̂ ̂ = σx̄ √𝑛 = P(z) 17.29 = 1.73 √100 = 99% = = x 99 100 = 0.99 0.99 = 0.4950 = 2.58 2 ̂ ± 𝑧 σx̄ ̂ x − 𝑧 σx̄ ̂ x + 𝑧 σx̄ 182 – 2.58(1.73) 182- 4.46 177.54 182+ 2.58(1.73) 182+4.46 186.46 So 99% confidence interval = (177.54 , 186.46) 𝜎 Unknown , n> 30 , population finite EX : (EX Sc 7-6 pg 365) a) Find estimated standard error of the mean i.e.𝜎 b) Find 96% confidence interval for the mean 12 Solution: a) n = 60 N = 540 x = 6.2 = s 𝜎̂ ̂ σx̄ ̂ 𝜎 = √𝑛 = 1.368 √60 = 1.368 x √ 𝑁−𝑛 𝑁−1 x√ = 0.18 x √ 540−60 540−1 480 539 = 0.18 x 0.94 = 0.17 b) P(z) = 96% = = x 95 100 0.96 2 = 0.96 = 0.4800 = 2.05 ̂ ± 𝑧 σx̄ ̂ x̄ x − 𝑧 σx̄ 6.2 – 2.05(0.17) 6.2 – 0.35 5.85 ̂ x + 𝑧 σx̄ 6.2+2.05(0.17) 6.2 + 0.35 6.55 So 96% confidence interval = (5.85, 6.55) H.W: a) Do EX 7-30 (Pg 365) . b) The height of random sample of 50 students has mean 174.5 and sample standard deviation is 6.9.Find 98% confidence interval for the mean. 13 t- Distribution: We use t-distribution when sample size i.e. n is less than equal to 30( n≤30) and 𝜎 is unknown and we assume that population is normal or approximately normal. t – distribution is symmetrical-distribution. It is lower at the mean and higher at the tails than the normal distribution so we can say that t-distribution is flatter than the normal distribution. As the sample size increases then it loses its flatness and becomes approximately equal to the normal distribution. There is different t-distribution for every sample size. Degree of Freedom: Number of values in a sample that we can choose freely is called degree of freedom. EX : (EX 7-44 PG 377) For the following sample sizes and confidence levels, find the appropriate t values for constructing confidence interval. Solution: a) n=15 P(t) =90% 90 = = 0.9 100 = 1 - 0.9 = 0.1 df = n-1 = 15-1 = 14 t0.1(14) = 1.761 b) n= 6 p(t) =95% 95 = = 0.95 100 14 = 1 - 0.95 = 0.05 df = n-1 = 6-1 = 5 t0.05(5) = 2.571 c) n = 19 p(t) = 99% 99 = = 0.99 100 = 1 - 0.99 = 0.01 df = n-1 = 19-1 = 18 t0.01(18) =2.0878 d) n= 25 p(t) = 98% = 98 100 = 0.98 = 1 - 0.98 = 0.02 df = n-1 = 25-1 = 24 t0.02(24) = 2.492 e) n = 10 p(t) = 99% 99 = = 0.99 100 = 1 - 0.99 = 0.01 df = n-1 = 10-1 =9 t0.01(9) = 3.250 15 f) n = 41 p(t) = 90% 90 = = 0.90 100 = 1 - 0.90 = 0.1 df = n-1 = 41- 1 = 40 t0.1 (40) = 1.684 H.W: Do EX Sc 7-10 (pg 377) 𝜎 n ≤ 30 Unknown, , population infinite EX : (EX Sc 7-11 pg 377) Solution: n = 7 x = 39.2 𝜎̂ = s = 3.2 ̂ σx̄ = = ̂ 𝜎 √𝑛 3.2 √7 = 1.21 P(t) = 95% 95 = = 0.95 100 = 1 - 0.95 = 0.05 df = 7-1 =6 t 0.05(6) = 2.447 ̂ x ± 𝑡σx̄ ̂ x − tσx̄ ̂ x + tσx̄ 39.2 – 2.447(1.21) 39.2 – 2.96 36.24 39.2 + 2.447(1.21) 39.2 + 2.96 42.16 So 95% confidence interval = (36.24 , 42.16 ) H.W: Do EX 7-46( Pg 377) 16 OBJECTIVE SECTION Q-1 WRITE SHORT ANSWERS FOR THE FOLLOWING 1- Write the suitable name for the process stated below, It is the process by which we obtained an estimate of unknown Population parameter by using sample. Answer: Estimation. 2-Write criteria of good estimator. Answer: i) ii) iii) iv) Unbiasedness Efficiency Consistency Sufficiency 6-Define estimator. Answer: A sample statistic used to estimate population parameter is called estimator. 8-Define degree of freedom. Answer: Number of values in the sample that we can choose freely is called degree of freedom. Q-2 CHOOSE THE CORRECT ONE. 1- Statistical inference is divided in to ------- areas. a) Two b) Five c) Ten 17 2- Formula for S is, ∑𝑥 a) 𝑛 ∑𝑥 b) 𝑛−1 c) None of these 3- Degree of freedom can be calculated as, a) n -1 b) N-9 c) 3n-1 4- When 𝜎 unknown ,n ≤ 30 and population assume to be normal or approximately normal then we use, a) t- distribution b) F-distribution c) Binomial distribution Q-3 Write true or false for the following. 1- x Is the point estimate of µ. 2- There are two types of estimates. 3- When an estimator has larger standard error than any other estimator it means it is efficient estimator of that estimator. 4- S = ∑(𝑥− x) 𝑛−1 18