Module-3: Statistical Inference 1 Introduction, sampling distribution, standan error, esting of hypothesis, levels of significance lest of signiticances, confidence limits, simple sampling of attributes, test of signi large samples, comparison of large samples, Hours) (RBT Lovels: L112 3.1 Samp Antroduet + Entire group of individuals under study is called population. Quantity associated with population like mean(j.), SD(a) is ealled parameter. Example: Set of all students in a college is a population, Mean weight of students is a parameter. ‘A small part of the population is called samples Quiatitity associated with sample like mean(2), SD(s) is called statist Example: Set of randomly selected 50 students ftom the college is a sample. Mean weight of these students is a statistic. SD [Smoke Population Sampling distribution and standard error: 4 The frequency distribiltion of means of different samples is called sampling distribution of mean, The standard deviation of the sampling distribution of mean is called standard ‘error of mean. The reciprocal of the standard error js called precision. where SE@P) = EE. ifP is known + Find the critical value using the table * Mealculated value < critical value, Accept Ho. Ho is the eohelu Mealeulated value > critical vatue Note: 8.00) = PS itp is known, SE (p) = Wary = [A varcyy Ve = bame= Confidence interval: 24 3ISEG)), if Fis known, PE SISE(PY, it P is known, PES[SEQ)), if P is known, Confidence interval at a level of sign F£ 20(SE(9)), if is known reject Ho. His the conclusion. P £ 2e[SE(P)], if P is known, PE za[SE(p)), if P is known, Or. Narasimhan 6, RNSITA sample of 900 members is found to have a mean of 3.4 em. Can it be reasonably rded as SD 1.61 cm. By data, ¥ uly random sample from a large population with mean 3.25 em and 3.4,n = 900, p= 3.25,0 = This is a large sample. Apply z test. lo? |.er S.E@= [C= |2—=01 @= he 300 = 0.0537 Ho: 61 3:25, Sample istken from the population wih mean 3, 28 3.4 — 3.25} FE Bo 0.0537 | Therefore, calculated value of z = 2.8 2.8 Ata = 0.05, critical value of z = 1.96 Since calculated value > critical value, Reject Ho. Therefore, sample is not taken from the population with mean 3.25 2. Ifa mean breaking strength of copper wire is 575 Ibs with a standard deviation 8.3 tbs. How large a sample must be used in order that there be one chance in 100 that the mean breaking strength of the sample is less than 572 Ibs. (Zq = 2.33) 575.0 = 83. ——..- Brsata, This is a large sample: Apply z test. s1@ eae 575, mean breaking strength of copper wire is 575 Ibs. | To find: n such that p< 572 ‘That is, find n such that Ho is rejected ‘That is, find m such that Calculated value > Critical value. > 233, n>4156 Therefore; n = 42: Or. Narasimhan G, RNSIT3 The mean of a c« normal population is equal (o the standard error of the mean of the samples of 100 from that di . Find the probability that the mean of the sample of 25 from the distribution will be negative, By data, w= SE (z) -E-— To find: P(® < 0) when n = 25, When n = 25, SE(z) = & E =P critical value, Reject Ho. ‘Therefore, the die is biased at 1% level of signifieance. In 324 throws of a die, an odd number turned up 181 times. Is it reasonable to think that ay 1% level of significance the die isan unbiased one? [zz = 2.58) Since n = 324, apply z test, By data, p =2°4 and P= = 3 SEC) Ee ca 0.0278 Ho: P =2,The die is unbiased. ere \al= sec = fo.0278 Therefore, calculated value of z = 2.1084 Ata = 0.01, critical value of z = 2.58 Since calculated value < critical value, accept Ho Therefore, the die is unbiased at 1% level of significance. Note: The die is biased at 5% level of signifi Dr. Narasimhan G, RNSITa sample of 840 families was selected at come of Z ing 18000 Families random, OF these $40 fi 2 i ave a monthly nu hese S40 families, 206 families were found to have a monthly milies have a monthly 28000 or less.It is desired to estimate how many out of 18,000 s would you place your estimate? income of ¥ 25,000 or less. Within wh: 206 iS = 0.2482 FOR = fEO2882)(0.7540) sag 7m B40 Required confidence limits are given by p 3ISE(D)] = 0.2452 + 3(0.0148) 2452 + 0.0444 0.2452 ~ 0.0444 and 0.24524 0.0444 0.2008 and 0.2896 By data, p SEQ) ive frequency of the . An unbiased coi nid Tie between O.49/and 0.51. Find the smallest Appearance of heads sh ‘hat will ensure this result with 90% eonfidence, (2s(0.1) = 1.645) TF ila PSs Ss —— Standard error of proportion: SEQ) = faa 20% confidence intervals siven by 0.5 + 2». SE(p) OS = 25. SE(),05 + 25. SE(p)) ess = (05-498 05 4 204s ( avn By data (05 — 292 605428 ) _ O6.49.953) ‘Therefore, the smallest value of is given by 05 +498 = 9.51 1.645% = 0.01 n= 6765 Or, Narasimhan 6, RINSITA random mple of 500 pineapples was taken from a large consignment and 65 were found to be bad. Show ths the standard error of the proportion of bad ones in a sample of this si is 0.015 that the percents consignment almost certainly lies between 8.5 and 17.5. 5 By data, p= = 0.13 s.£0)= f= FRR = oo1s Required confidence interval= (p — 3[SE(p)],p + 3[SE(p)}) = (0.13 — 3(0.015), 0.13 + 3(0.015)) 0.085, 0.175) = (8.5%, 17.5%) ge of bad ples in, the Dr. Narasimhan G, RNSIT= 3.3 Comparison of large samples Working rule: Write the null hypothesis Ho Find the calculated value using [este lel - {er » if], Xp are known Ise | sf, 73 are not known ® Find the critical value using the table. + Ifcalcutated value < critical value, accept Ho. Ho isthe conclusion. * Ifealculated value > critical value reject Ho. Hi is the coneligion. Standard error: ELS a7, 2 aretnou SEG ~¥) If 4,02 are known If cis known If Py,P2 are known 1 MP: + n2p2 7 Fr, Dr. Narasimhan G, RNSIT1. The means of samples of sizes 1000 and 2000 are 67.5 and 68.0 cms respectively. Can the samples be regarded as drawn from the same population of SD 2.5 cm? 1 zs(0.08) = 1.96] Since samples sizes are n, = 1000, n, = 2000, apply 2 test By data, X7 = 67.5, = 68.0 and o = 2.5 Ho: #2, = #tg, Both the samples are drawn from the same population. SEQ -%) - 2 25 Jro00 * 2000 = [4] = ts ‘06s | ~ Gose8 ‘Therefore, calculated value of 2 = 5.16 At @ = 0.05, critical value of z= 1.96 Since calculated value > criti¢al value, rejéct Ho. | Therefore; Both the samples are not drain from the same population. Dr. Narasimhan G, RNSIT2 A sample of height of 6400 soldiers has a mean of 67.85 inches and a standara ‘leviation of 2.56 inches while a sample of height of 1600 sailors has a mean of 68.55 inches and a SD of 2.52 inches. Does the data indicate that the sailors are on an Serage taller than soldiers? Use 0.05 level of significance. [zy 1.65) Since samples sizes are n, = 6400,np 1600, apply z test. By dats, 37 = 67.85, 8.55, 5, = 256, s, = 252 The sailors are not taller than soldiers. ea SER. ty Se ny * = fee 0.0707 Eseerd Ho: py = pp, l= EES ~ FESS) = 2,499 Therefore, calculated value =9.9 At = 0.05, critical value = igs Since calculated value> critical val Therefores Or. Narasimhan G, RNSIT3, A sumple of 100 electrie bulbs produced by manufacturer A showed a mean lifetime of 1190 hours und a standard deviation of 90 hours. A sample of 75 bulbs produced by manufacturer Wi showed # mean lifetime of 1230 hours with a standard deviation of 120 hours. Is there a difference hetween the mean lifetime of two brands at significant level of 0.05? (Za/2 = 1.96) Since samples sizes are m, = 100,n, = 75, apply z test. By data, ¥] = 1190, xy = 1230, s, = 90, sp = 120 Noi 141 = Hay There is no difference between the mean lifetime of two brands. SEG -%) 16.5227 ‘Therefore, calculated value At a = 0.05, critical value = 1.96 Since calculated value > critical value, reject Ho. ~—“Thefefore, There Fa UiTTercnee between the medn Nifetime of two Brands at significant level of 0.05, Dr. Narasimhan G, RNSITOne type of aircraft i found to develop engine trouble in § Aights out of a total of 100 an. another type in 7 lights out of total of 200 fights Is there {types concerned so, engine defects are concerned? (Zqa eS are ny = 100 and ny = 200, apply z test = 100(0.05) + 200(0.035) © 100 259 = 0.08 = 0.024 ‘Therefore, calculated value = Ata = 0.05, critical value= 1196 Since caleulatedl Value critical value, Reject Ho. ‘Therefore, this difference i unlikely to be hidden in samples of 1200 and 900 respectively from the two populations. Dr. Narasimhan G, RNSIT