——— m™ LARGE SAMPLES § Test of significance of Large Samples If the size of the sample > 30, then that sample is called large | sample. There are four important test to test the significance of large samples, 1. Test of significance for single proportion 2. Test of significance for difference of proportions 3. Test of significance for single mean 4. Test of significance for difference of means & Test of significance for single proportion j Suppose a large sample of size n is taken from a normal population. To test the significant difference between the sample proportion p and the population proportion P, we use the statistic =P = = where 7 - sample size Note 1 : Limits for population proportion P are given by ps3. where q=1-p. = EXAMPLE 1 & { A manufacturer claimed that,atleast 95% of the equipment which * he supplied to a factory conformed to specifications. An examination of a sample of-200 pieces of equipment revealed that 18 were faulty. Test his claim at Ss) level of significance. Given sample size n = 200 Number of pieces confirming to specification = 200 —18 P= is “- a¢ pROBABILITY. & STATISTICS p = proportion of pieces confirming to specifications 162 © = 200 =0.91 = Population proportion , 95 = 700 = 0.95 Q=1-P = 1-0.95=0.05 Null hypothesis Hy : The proportion of pieces confirming to specifications ie., P= 95%. ‘ Alternative hypothesis Hy : P < 0.95 (Left tail test) The test statistic _ p-P __0.91-0.95 z= "PQ. [0950.05 \ n 200 0.04, yy - = Q01s4 77259 Since alternative hypothesis | is left tailed, the tabulated value of z at 5% level of significance is 1.645. Since calculated value of | z | = 2.6 is greater than 1 645 we riject the null hypothesis Hy at 5% level of significance. Hence the manufacturer claim is rejected. EXAMPLE 20 In a sample of 1000 people in Karnataka re rice eaters and the rest are wheat easers. Can we assume that both rice and wheat are equally popular in this state at 1% level of significance ? Given n = 1000 ~s P = sample proportion of rice eaters = 240. Tooo = 954 P = population proportion of rice eaters 1 = 7 =05 t SS _—_— UNIT 4 ~~ao rere er—rwww Q = 05 Null Hypothesis Hy : Both rice and wheat are equally popular in . state. Alternative Hypothesis Hy : P # 0.5 (two tailed alternative) Test statistic is ook 7, [Po n = 0.54-0.5 =2.532 0.5 x 0.5 1000 The calculated value of z =2.532. The tabulated value of z at 1% level of significance for two tailed test is 2.58. Since calculated z < tabulated z we accept Hg. ie. Both rice and wheat are equally popular in the state. = EXAMPLE 3 & In a big city 325 men out of 600 men were found to be smokers. Does this information support the conclusion that the majority of men in this city are smokers? ae Given n = 600 Number of smokers = 325 P = sample proportion of smokers ~ 25255 : = 600 = 0.5417 P = population proportion of smokers in the city oes = 7 =05 Q=1-P =1-05 =05 Null hypothesis Hy : The number of smokers and non-smokers a, « equal in the city. Alternative Hypothesis ; P > 0.5 (Right - tailed) Test statistic — = UNIT 4= 0.5417 -0.5 10.5 x 0.5 600 = 2.04 Calculated value of z= 2.04 ight tail test is Tabulated value of z at 5% level of significance for ri 1,645. Since calculated value of z > tabulated value of z. We reject the null hypothesis. The majority of men i smokers. in the city are EXAMPLE 4 ee t In a sample of 400 parts manufactured by a factory, the number of defective parts ‘was found to be 30, The company, however, (aimed that only 5% of their product is defective. Is the claim tenable ? 4} Given n = 400 4) Number of defectives in the sample = 30 = Proportion of defectives in the sample p= 30 _ = 500 = 0.075 P = the population proportion ane = Too = 0.05 .Q = 0.95 table. The company’s claim p= 0.05 is accept Null hypothesis Ho : + P > 0.05 [Right-tailed alternative] Alternative hypothesis Hy Test statistic is eee ge "ps _, 0.075 - 0.050. = 9.29 0.05 x 0.95 400 8 neROBABILITY 4,72 TESTING OF HYPOTHESIS p & STATISTIog =2.29 +. Calculated value of 2 2.2! / w Calcul at 5% level of significance for right-tailed tesp is ue of z Wer > tabulated 2. 1.645. Here calculated z We reject the null hypothesis. / ie, The company’s claim that only 5% of their product is defective jg not acceptable. m EXAMPLE 5 —_—_—_— OO O_O OO Fl ‘A die was thrown 9000 times and of these 3220 yielded a 3 or 4. Is this consistent with the hypothesis that the die was unbiased ? Given n = 9000 P = Proportion of successes of getting 3 or 4 in 9000 throws 3220 . : . = 9000 ~ 0.3578 “ps Population proportion of successes = P (getting a3 or 4) 2 4h i2 6 6 = 0.3333 = we ae = 2_ 1 3 = 0.6667 -P 1 3 Null hypothesis Hy : The die is unbiased, Alternative hypothesis H, : P ¢ 4 (Two-tailed) The test statistic is z= PoP _ _ 03578-03333 ‘ = —03578=0.3333 Ey [0.3333 x 0.6667 ~ 44 Calculated 2 = 4.94 DD Since z > 3, the null hypothesis Hp is rejected. ie., The die is biased, : S UNIT 4‘ TESTING OF HYPOTHESIS _4.73 jaoBABILITY & STATISTICS EXAMPLE 6 & : A random sample of 500 apples was taken from a large consignment and 60 were found to be bad. Obtain the 98% confidence limits for the percentage number of bad apples in the consignment. Given n = 500 @) p = Proportion of bad apples in the sample " W Ss 8 2g = 0.88 ; / We know that 98% confidence limits for population proportion are saat? 98% confidence coefficient is 2.33] ¢ 98% confidence coefficient is 233] \ = 0.12 £2.33 x 0.01453 / = 0.12 + 0.03385 = (0.08615, 0.15385) 98% confidence limits for percentage of bad apples in the consignment are (8.61, 15.38) . MPLE 7 @. A random sample of 500 pineapples was taken from a large consignment and 65 were found to be bad. Find the percentage of bad pineapples in the consignment. Given n = 500 P = Proportion of bad pineapples in the sample os 500 = 0.13 q = |-p=0.87 We know that the limits for population proportion Pare given by Pt [AL s n [0.13 «0.87 = 0. 0.13 x 0.87 _ 1343 300 0.13 + 0.045 = (0.175, 0.085) ; in; The percentage of bad pineapples in the consignment lis % and 8.5%), 2 eee Tn jes between ~~m= DIFFERENCE OF PROPORTIONS - Suppose 2 large samples of sizes 7 and My are taken respectively frc 2 different populations. To test the significant difference between the sample proportions p; and pz find and g=1 -p. a Pia=P2 h ny Pi tha P2 Zz vf) yee ane o aa wy 2 ONT 4 A ee efFfFye popaBuntTy & STATISTICS TESTING OF HYPOTHESIS 4.75. EXAMPLE 1 Random samples of 400 men and 600 women were asked whether they would like to have a flyover near their residence. 200 men and 335 women were in favour of the proposal. Test the hypothesis that proportions of men and women in favour of the proposal are same, at 5% level. Given sample sizes 7 = 400, n = 600. ; 200 -,..— Proportion of men py = 499 = 0.5 + F Proportion of women py = 38 =0.54h Null hypothesis Hg : Assume that there is no significant difference between the option of men and women as far as proposal of flyover is concerned. fo a He - Hoipt = P2-P |] Alternative hypothesis Hy : py #72 (two tailed) \ The test statistic is 2. = cam Pq Gin) + mpi tne where p= Tomy +My — 200, 325 __ 400° 99 * 609 * 600 aan = 400+ 600 “ ~ SL gai = To00 7,955 > q =" 1-0.525=0.475° 0.5 — 0.541 1 0.525 x 0.425 ea 0.041 ° = 9032 = ~!34 jz] = 134 Since | z | < 1.96, we accept the null hypothesis Hy at 5% level of significance. ie., There is no difference of opinion between men and wom as proposal of flyover is concerned. a UNIT 40 jen as far4.76 TESTING OF HYPOTHESIS PROBABILITY & STATISTICS = EXAMPLE 2 &8—@—€£@£§£@€24A———————- increase in excise duty on tea, 800 persons out of a samele "ot 1000 persons were found to be tea drinkers. After an increase in duty, 800 people were tea drinkers in a sample of 1200 people. Using standard error of proportion, state whether there is a, significant decrease in the consumption of tea after the increase in excise duty. Given m = 1000 - ny = 1200 800 “+ 800 Pi ~ 7900 -- P2 = 7200 = 08 = 0.667 _ Mm Pit ng pr Now, eo 300 800 - 1000 (Tyan) + 1200 (#4 = 1000 + 1200 — 1600 _ _ = 3909 = 0.727 “Aq = 1-0.727=0.273 —~, Null hypothesis Hg : Assume that there is no significant difference in the consumption of tea before and after the increase in excise duty. ie, Ho: 7, = py Alternative hypothesis Hy : p; > py (right tailed test) — The test statistic is z = —-PI=P2 : ( f rt) - . nm ng 0,8 0.667 \ Td 0.727 x 0.273 Gaba a a) | | 4 = 0.8~0.667 “0019. = 7 Since z > 1.64 (one tail) we reject the null hypothesis at 3% level of significance. ie. there is a difference in th Paes ie consumption of tea before and after the increase in excise duty, mw UN 4 ———???____vme TESTING OF HYPOTHESIS 4.77 gop aBsusry & STATISTICS. Note 1: If we want to test the significance of the difference between n and p where mi pit ma P2 p= nyt istic i PZ? ten the statistic is Z = [mpd ny (y+ Nn) v- EXAMPLE 3 © In a random sample of 400 students of the university teaching department, it was found that In another sample number of failures together. Given 300 students failed in the examination. of 500 students of the affiliated colleges the in the same examination was found to be 300. i Find out whether the proportion of failures in the university teaching departments significantly greater the university teaching departments and affiliated colleges taken than the proportion of failures in ny = 400, ny = 500 300 - 300 Pi = 400 =0.75 P2 = 500 =0.6 _ m Prt m2 P2 ony tn : = 300 300 - a 400 + 500 mee oq = 0333 . Null hypothesis Hg : Assume that there is no signifidugt difference between p; and p. The test statistic Pi-P is,z = fm Pd ny (ny + 1) 0.75 — 0.667 ~~ [500 x 0.667 x 0.333 400 (400 + 500) ~ _—_—_—_ UNIT 48 oO4.78 TESTING OF HYPOTHESIS $B STATISNics 0.083 | = ors 7474 Calculated z > 1.96 (at 5% level of significance). Hence H is rejected, <. The proportion of failures in the affiliated colleges is greater than the proportion of failures in university departments and affiliated colleges taken together. Note 2 : Suppose the population proportions P; and P are given and P, # Pa. If we want to test the hypothesis that the difference (P; — Pz) in population proportions is likely to be hidden in simple samples of sizes 7, and ny from the two Populations respectively then (Py =P2) (Py) ~ Pa) PiQ Po z= = x Ss If the sample proportions are not known then we use IPi=Pal Bae EEE PrQi | PQ na ra ™ EXAMPLE 4 © A cigarette manufacturing firm claims that its brand A-line of cigarettes outsells its brand B by 8%. If it is found that 42 out of a i ~ sample of 200 smokers prefer brand A and 18 out of another sample | ; of 100 smokers prefer brand B, test whether the 8% difference is a | \ valid claii eB 2Q Z é s se \ R 8 S ny = 100 ° = _ 18. Pi * 300 P2 * 100 0.18 3 Wl e I 4 } 8 P}-Py = 8% =~ = [7 1~P2 = 8% 700 0.08 Ny Null hypothesis Hy : Assume that 8% difference in the sale of t¥° j | Hy: A | | 5 _\ brands of cigarettes is a valid claim. i.e., Hp : py —P2 = 0.08 | / ‘() Alternative hypothesis Hy : p, — py #0.08 (Two tailed test) | = UNIT4 ee . iopABILITY & STATISTICS TESTING OF HYPOTHESIS 4.79 i The test statistic is,z (1 —P2)~ (Pi - Po) mati) _ Pit Pz nytny 200 (0.21) + 100 (0.18 200 + 100 = 42418 _ s " o 300 1-0.2 = 0.8 0.03 — 0.08 02x08(s55+ +h 0.05 = 0.0489 = ~1.02 lz, = 1.02 Since | z | < 1.96 we accept ae null hypothesis Hy at 5% level of ‘ignificance, lam, ie, 8% difference i in the sale of two brands of cigarettes is a valid - Excerahora TEST OF SIGNIFICANCE FOR SINGLE MEAN Suppose we want to test whether the given sample of size 7 has been ‘ drawn from a population with mean p. We set up null hypothesis. that there is no difference between x and p where x is the sample mean. rae x- The test statistic is,z = oa ln in |where o is the S.D. of the population. | If the population S.D. is not known, then use the statistic | - \where s is the sample S.D. — | jt \ . woa w SABILITY & STATISTI 4.82 TESTING OF HYPOTHESIS. » f. PROB: ics -yf Note : The values ¥ + 1.96 -& are called 95% fiducial limits or a re . . : 0 confidence limits for the mean of the populations corresponding to the 4 given sample. Similarly ¥ + 2.58 i are called 99% confidence limits, iss m EXAMPLE 1 &- A sample of 900 members has a mean of 3.4 cms and S.D. 2.61 . ems. Is the sample from a large population of mean 3.25 cm and $.D. jis 2.61 cms. If the population is normal and its mean is unknown find 458 the 95% fiducial limits of true mean. : E Given "n= 900 2p = 3.25 : ‘ ¥ = 34cm s = 2.61 Null hypothesis Hp : Assume that the sample has been drawn from the population with mean p= 3.25. m Alternative hypothesis H, : 1 #3.25 The est statistic iz = SHH 234-325 | PN ° s 2.61 . .. vn p00“ Y L724 ; os " So 1.724 < 1.96 “GY We accept the null hypothesis Ho. 0 z j i-e., The sample has been drawn from the population with mean = 3.25. | 95% confidence limits are-=" + 1.96 a ' in 2.61 ' a = 3.441.96x ¥900 = 3.440.1705 = 3.57 and 3.2295 UNIT 4pROBABILITY. ESTAS TICS TESTING OF HYPOTHESIS 4.83 s EXAMPLE 2 @ ‘ An insurance agent has claimed that the average age of policy’ : 4 polders who issue through him is less than the average for all agents which is 30.5 years. A random sample of 100 policy holders who had jssued through him gave the following age distribution. Age 16-20 | 21-25 | 26-30 | 31-35 | 36-40 No. of persons 12 22 20 30 16 Calculate the Arithmetic mean and standard deviation of this distribution and use these values to test his claim at 5% level of significance, — AY fidi ¥ = At ww 5x16 oD 28+ = 28.8 100 ; ; S72 ap ¢ 4-C : ; 164 (16) rea-(ren) th=5] x } = 635 no Null hypothesis Hy : The sample is drawn from a population with rg Mw w w W mean pie., ¥ and p do not differ significantly. ie., 1 =30.5 years Alternative hypothesis H, : 1 < 30.5 years (left tail test) Now, x = 288, 5 = 635 B= 305 yrs n = 100 ‘ The test statistic is,z j = -2.677 i 4 . } c. [2] = 2.68 : Tabulated value of z at 5% level of significance is 1.645 (Left tail test) / So UNIT 40 /4.84 TESTING OF HYPOTHESIS PROBABILITY & STANSHeg f Here calculated z > tabulated z. . The null hypothesis Hg is rejected. “¥ and p differ significantly. ie., the sample is not drawn from a population with mean p, = TEST OF SIGNIFICANCE FOR DIFFERENCE of MEAN Let x] be the mean of a sample of size m1, from a population with mean ty and variance of. Let Xz be the mean of a sample of size n from a population with 2 mean jig and variance 03. ° To test whether there is any significant difference between % and ey we have to use the statistic X1-% a ot oS . m "ny Note : If the samples have been drawn from the same population then met as’ . of = 03 = 02. , 3-H 7 Tee Gt mm : If is not known we can use a estimate of o2 given by > ht 5 . 2 = MMS ' _A mtn @ EXAMPLE 1 & The means of 2 large samples 1000 and 2000 members are 67-5 inches and 68.0 inches respectively. Can the samples the regarded 8 ( drawn from the same population of S.D. 2.5 inches. = UNIT4omm 1000 ny = 2000 x = 67.5 inches Xq = 68 inches i = 2.5 inches Population SD, o Null hypothesis Hp : The samples have been drawn from the same opulation of S.D. 2.5 inches. iF ie, 1 = Hg and o = 2.5 inches. Alternative hypothesis Hy : py # p- The test statistic is,z = hy Ee 0.0968 " ie. The samples are not drawn from-the same population of $.D. 2.5 = -5.16 |z| = 5.16> 1.96 . The null hypothesis Hg is rejected at 5% level of significance. ees EXAMPLE 2 @———— eee mean yield of wheat from a district A was 210 pounds with SD. 10 Pounds per acre from a sample of 100 plots. In another istrict the mean yield was 220 pounds with $.D. 12 pounds from a 4 ample of 150 plots. Assuming that the S.D. of yield in the entire state ts 11 pounds, test whether there is any significant difference [tween the te yield of crops in the two districts. aa a _ ~ % + 210 % = 200 m = 100 ny = 150 : ttPROBABILITY & STATISTicg f 4.86 TESTING OF HYPOTHESIS: Population $.D.o = 11 ¢ Null hypothesis Hg : There is no difference between x; and x. Ho: =%2 ‘Alternative hypothesis Hy : ¥ # cy oH _ __210-200__ The test statistic is,z = 75-4 2? (tetas m mm 100 © 150, = 7.04178 ie, z = 7.041> 1.96 ie. ‘:. The null hypothesis Hg is rejected at 5% level of significance. sho ie, There is a difference between X; and x7 . ie, There is a significant difference between the mean yield of crops in the two distinct. 1 Note : If the two samples are drawn from two populations with unknown standard deviations then SS. 4,2 m Mm = EXAMPLE 3 & In a survey of buying habits, 400 women shoppers are chosen #1 random in super market ‘A’ located in a certain section of the city Their average weekly food expenditure is Rs.250 with a S.D. of Rs.40- For 400 women shoppers chosen at random in super market ‘B’ in another section of the city, the average weekly food expenditure is Rs.220 with a S.D. of Rs.55, Test at 1% level of significance whether the average weekly food expenditure of the two populations of shopper are equal. 4 Given +m = 400, ¥ =Rs.250, 5, = Rs.40 m, = 400, X =Rs.220, 52 = Rs.55 Null hypothesis Hy : Assume that the average ‘weekly f0% ON 4 —prOBABILITY & STATISTICS TESTING OF HYPOTHESIS 4.87 expenditure of the two populations of shoppers are equal. ie, Ho: %) => Alternative hypothesis Hy : x; #2, ’ Test statistic is, z = —41= ss, 21,2 m m _ 250-220 _ 30 _ gg (40)? , (55) 3.4 400 ” 400 ie, | Zz 8.82 > 2.58. , . The null hypothesis Ho is rejected. ie., The average weekly food expenditure of the two populations of shoppers are not equal. tl = SVEDrICE m