‘ Biostatistics of s Hin Leet ty ty ty wes ty be the values of a statistic t (mean or standard Eafe variable distribution) for k possible samples. Thus the statistic t may he regarded as renee which can take any one of the vale ty ty yy +5 ty The set of the values cesta aha is known as the sampling distribution of the statistic t, We can compute tic t. For example, and other statistical constants of the sampling distributions of the statistic t Mean = E (1) = he i (say) 1 iy Variance of ¢ = Var () = i Ep LOP= ZO“ O 12.9 STANDARD ERROR OF A STATISTIC becrvati The statistical measure of standard deviation may be computed both from the ol we kan of the population and also from the observations of the Sampling distribution. Also Patt de that the standard deviation is a measure of the average amount of the variability of al re specrations of variable from their mean, When the average amount of the variability of the observations of a population is computed, it is called the standard deviation. But when the merase amount of the variability of the observations of a,sampling distribution is computed, itis known as Standard Error, Thus the standard deviation computed from the observations of a sampling distribution of a statistic is called the standard error of the statistic. In other Words, the standard deviation used to measure the variability of the values 0 tic from sample to Sample Cae The Han of inetd and stndard error have the same meaning and same connection but are used in different contexts and different circumstances, Both of them are used to measure the variability of observations. Standard error is used to measure the variability of the values of a statistic computed from the samples of the same size drawn from the population, whereas standard deviation is used to measure the variability of the observations of the population itself. Thus the standard deviation of the sampling distribution of a statistic ‘r’s known as standard error of “1”, It is generally, denote by S.E. (#). Thus SB.) = War@ = |x (-7) The following two standard errors are frequently used in statistics, 1. Standard Error of Sample Mean ; It is {he standard deviation of the sampling o dtstribution of sample means. Itis denoted by oF, and is given by o Tq where & is the standard deviation of population and n is the sample size. _ oO HEE) = Tr? Where o is known _ s . or S.E(E) = 7, when ois not known =, § = standard deviation of sample is given Scanned with CamScannerHypothesis Testing and Large Samples Tests as 2. Standard error of difference of two sample means X, and X, drawn from the same 1 population. | L ac) when © = population standard deviation is known; and n, an n, are the sizes of two samples, which are drawn from the same population. 3. Standard Error of difference of two samples means ¥, and Z, drawn from two different populations with standard deviations o, and o, respectively S.E.(¥,-¥,) = ( of my Mz” S.E.(F,-¥,) = where n, and n, are the sizes of the two samples. 12.10 UTILITY OF STANDARD ERROR OF A STATISTIC The standard error of sample mean or the standard error of proportion is used in sampling in order to obtain the following facts: 1. Standard error is used to set up the confidence limits within which the population parameter may lie. 2. Standard error is used to test the hypothesis and to dranw a statistical conclusion from it. 3. Standard error is used to measure the variability of the values of a statistic from its ‘mean. 4. Itis, generally, used for large samples and gives us the idea about the average amount of error which actually occurs in estimating the values of a parameter on the basis of @ statistic. 12.11 ESTIMATION THEORY Estimation of population parameters like mean, variance, proportion, correlation coefficient, etc., from the corresponding sample statistic is one of the very important problems of statistical inference. The theory of estimation was founded by Prof. R.A. Fisher round about 1930 and is divided into two groups. (i) Point Estimation and (ii) Interval estimation. 12.12 POINT ESTIMATION In point estimation a single statistic (numerical value) is used to provide an estimate of the population pardmeter. A particular value of a statistic which is used to estimate a given parameter is known as a point estimate or estimator of the parameter, A good estimator is one which is as close to the true value of the population parameter as possible, The following are some of the criteria which should be satisfied by a good estimator. (i) Unbiasedness; (ii) Consistency ; (iii) Efficiency ; (iv) Sufficiency Scanned with CamScannerBiostatistics ¢ sample observations $ ation parameter 276 ion of thi 1. Unbiasedness : A statistic 1 = 1 (yy Xyy eo Xs 2 Spot ‘ding pop “yey nny Ny iS Said to be an unbiased estimate of the corr 6, if roe aise is equal to parameter i.e, if the mean value of the sampling distribution of the s a 2, Efficiency ; If t, and f, are estimators of a parameter 0 suc Mt, yarn 7 7 “in ‘vords, an estimator with lesser vequently more reliable than the other. then 1, is said to be more efficient than ¢,. variability is said to be more efficient and consequent Efficiency (E) of f, relative to 1, is defined as: Obviously, E <1. 6G yonll HH 3. Efficiency and sufficiency : The concepts of consistency and sufficiency are bey scope of this book. 12.13 INTERVAL ESTIMATION . There are situation where the point estimation is not desirable and we are interested in finding such limits within which with a known probability or to a known degree of reliance, the value of the population parameter is expected to lie. Such a process of estimation is called the interval estimation. In other words, an interval estimation is the range of values used in making estimation of a population paraméter. Thus the interval estimation of the population parameter is the estimation of the population parameter by an interval around. the point, The interval estimation of a population parameter is the estimation of the parameter 0 with the help of the interval |f— s, ¢ +s], where # is the sample statistic, i.c., {-s Z, = 1.96 Confidence interval = ¥ + S.E. = 15 + 0.98, 4. 12.17 TESTING OF HYPOTHESIS imation and testing of hypothesis, Sampling theory deals with two types of problems, vz., estimation Tee iors Op In this section we shall deal with the problem of Testing of Hyp h of statistics which helps Probability plays an important role in decision-making and-the branc} ata esis, hich weg us in arriving atthe criterion for such dectsions is known as testing of hyPOl ss, ft’ is initiated by J. Neyman and E.S. Pearson, It employes statistical techniques 10 AUNT Pes in certain situation where there is an element of uncertainty on the basis of samp! zo Sed in advance, ion parameter. In other \efypothesis : A hypothesis is a statement about the population p: ich is t i ical basis. Words, a hypothesis is a conclusion which is tentatively drawn on logical seo It is tentative conclusion that specifies the properties of a ) x 1.96 0.5) F) x Z, = 15 + 14.02 to 15.98. Statistical Hypothesis = ” © distribution of a random-variable. These properties generally refer to parameters of the population and the hypothetical values with whicl the values of statistic derived from a sample on ae inorder to find the difference between statistic and corresponding parameter. In other words, esis is some assumption or statement, which may or may not be true, in k atstical hypoth ms : 5 about a population or about the probability distribution characterising the given population, which we want to test on the basis of the evidence from a random sample. Test of Hypothesis : Hypothesis testing can be regarded as an example of a decision process, in which data are assembled in a particular way to produce a quantity that leads to a choice between two decisions. Each decision then leads to an action, Because data arise from sampling Process, there is some risk that an incorrect decision will be made with some loss attached to the resulting incorrect action. In this context, hypothesis testing is an example of'a more general study known as decision theory. Many situations in life Tequire a choice between two decisions, whether or not the choice is actually made on the basis of data from a sample. The decisions to Purchase or not to purchase a particular piece of laboratory equipment or, to market or not to market a particular medicine. The testing of hypothesis is a procedure that helps us to ascertain the likelihood of hypothesised population parameter being correct by making use of the from the sample/In testing of hypothesis a statistic is computed from a sample drawn from the parent population and on the basis of this statistic it is observed whether the sample so drawn has come from the population with certain Specified characteristic) The value of sample statistic may differ from the corresponding population parameter due t my in fluctuations. The test of hypothesis discloses the fact whether the difference bets nap ie statistic and the corresponding hypothetical population Parameters significant or n ee n uno ; Thus. the test of hypothesis is also known as the test of significance, az Scanned with CamScannerl p Hypothesis Testing and Large Samples Tests mt 18 PROCEDURE OF TESTING A HYPOTHESIS following are the steps involved in hypothesis testing problems, < Setting up of Hypothesis : There are two types of hypothesis: 1. Null hypothesis. 2. Alternative hypothesis as aa Hypothesis : The statistical hypothesis that is set up for testing a hypothesis is known decid ae ‘The null hypothesis is set up for testing a statistical hypothesis only to Whether to accept or reject the null hypothesis. It asserlé that there is no difference betw: } : eee ete Sample statis populition parameter and whatever difference is there, is uutable to sampling-crror iypothesis i usually denoted by H,- Setting up null hypothesis = The following steps must be taken into consideration while setting up a null hypothesis. z () In order to test the significance of the difference between a sample statistic and the population parameter or between the two different sample statistic, we set up the null \hypothesis H, that the difference is not significant. ‘There may be some difference but that is solely due to Sampling fluctuations. . (ti) To test any statement about the population, we hypothesise that it is true. For example, if we want to find the population mean has a specified value }1,, then the hypothesis H, is set as follows. Hy: B= Hy Prof. R.A Fisher remarked, “Null hypothesis is the hypothesis which is to be tested for possible rejection under the assumption it is true” ‘Alternative Hypothesis ; The negation of null hypothesis is called the Alternative hypothesis. In other words, any hypothesis which is not a null hypothesis is called an alternative hypothesis. It is always denoted by H, or H,, It is set in such a way that the rejection of null hypothesis implies the acceptance of alternative hypothesis. For example, if we want to find the null hypothesis that the average height of the student of a college is 165 cm. i.e., fly = 165 oms (say). Then the Null hypothesis is H, = #= 165 = 1) and the Alternative hypothesis could be (@) Hyrn Hy (ie, u> Hy oF B Hy [One Tailed Test] {One Tailed Test] (ii) Hy th < Hy \_2-Computation of Test Statistic ; After setting up the null hypothesis and alternative hypothesis, we compute the test sta ic, The test statistic is a statistic based on appropriate probability distribution. It is uscd to test ‘whether the null hypothesis sct up should be accepted br rejected. It is the main yard stick which help us to decide whether to accept or reject the null hypothesis. Different probability distribution values are used in appropriate cases while testing the null hypothesis. The following are the most commonly used test statistic. ____—— Scanned with CamScanner282 istribution : We use Z-distribution Biostatistics (Ge., if the sample size n > 30). The Z-statistic is defined a8 E.() EO). y0,1) sn > 2 i one. is a standard normal variable with mean zero and variance Also, SE. (1) = Standard error of the statistic ¢ ttest : t-test is used t-statistic is defined as ie, i size for small sample, i.c., if the sample i iS H___ Difference of sample and population mean’ = Difference of sample and population means i S.E. (=) Standard error of mean with — 1 degrees of freedom, where n is the size of sample. for large samp mal curve for le under the nor n $ 30. The student's — The following table shows the conditions for using Z-test and t-test in testing null hypothesis about the means. : ia TABLE Size of sample | Population s.d. cis known | Population s.d. o is not known n>30 Ztest Z-test 1X30 t-test t-test igs of Errors in Hypo' a null hypothesis may be correct (2) Type I Error : It is the error of rejecting null hypothesis H,, when it is true. When a rence (of mean) is significant and the hypothesis is made. The probability of making a type I error is nificance, In order to control the type I error, the ignificance c, The probability of null hypothesis is true, but the diffe ejected, then a Type I Error is denoted by a, the level of sig Probability of type I error is fixed making a correct decision is tien (1 - a). (b) Type I Error : It is the error and the hypothesis is accepted, type Il error is denoted by p. The following summary table in whi Decisibn Based on Data Accept H, Reject H, at a certain level of si Correct Decision Type I error (a) thesis Testing : There is every chance that a decision regarding OF may not be correct, There are two types of errors. of accepting the null hypothesis H, when it is false. In esis is false, but the difference of means is insignificant Type Il error (B) Correct decision Ml Scanned with CamScannera” - . . 'ypothesis Testing and Large Samples Tests = Ut may be helpful to think of hypothesis testing an analogous to a jury trial of H: the Poteet The type I error corresponds to convicting an eee while responds 10 acquitting a guilty defendant. po eee is probability of rejecting the tested hypothesis when it ts false, t.e., when an © hypothesis is true; denoted by 1 — B, where B is the probability of a type II error. arsipune et Si cance : The next step is the fixation of level of significance. The level P (Rejecting Ht } maximum probability of making a type I error and it is denoted by ti. ace Te 2 H, is true) = o., The probability of making a correct decision is then ae e for fixing the level of significance depends on the seriousness of the aan ‘ypes of error, The commonly used level’ of significance in practice are 5% -05) and 1% (0, = 0.01). If we use 5% level of significance (« = 0.05) we shall mean that the probability of making type I error is 0.05 or 5%, i.e., P [Rejecting H, when H, is true] = 0.05 This mican that there is a probability of making 5 out of 100 (or making 1 out of 20) type Terror. Tis also mean that we are 95% confident that a correct decision has been made. Similarly 1% level of significance (c = 0.01) that there is a probability of making 1 error out of 100. It 1s Important to note that if no level of si icance is given, then we always take a = 0.05. 5. Critical Region or Rejection Region : The rejection region or critical region is the region of the standard normal curve corresponding to predetermined level of significance ct (which is fixed for knowing the probability of making a type I error of rejecting the null hypothesis H, when it is true). The region under the normal curve which is not covered by the rejection region is known as Acceptance Region. Thus the statistic which lcads to rejection of null hypothesis H, gives us a region known as Rejection Region or Critical Region. While those which Icad to the acceptance of H, give us a region called a Acceptance region. The value of the test statistic computed to test the null hypothesis H, is known as the critical value. The critical value separates the rejection region from the acceptance region. 6. Two Tailed Test and One Tailed Test : The probability curve of the sampling distribution of the test statistic is a normal curve. In any test, the critical region is represented by a portion of the area under this normal curve. This curve has two sides (or ends) known as two tails. The rejection region may be represented by a portion of arca on cach of the two sides or by only one side of the normal curve and correspondingly the test is known as Two Tailed Test (or Two sided Test) or one Tailed Test (or one sided test) Two Tailed Test or Two sided Test : When the test of hypothesis is made. on the basis of rejection region represented by both sides of the standard normal curve, it is called a vo tailed test or two sided test (sce Fig. 12.1). In other words, a test of statistical hypothesis where is led or two tailed such has the alternative hypothesis Null Hypothesis : H, : # = Hy Alternative Hypothesis: H,:##H) (H> p, or p< p,) It is called as two tailed test or two sided test. j Scanned with CamScannerBiostatistics 284 One Tailed or One sided Test : A test of statistical hypothes!s where ether oon hypothesis is one sided is called as one tailed test or one sided test. There tailed or one sided tests. 1, Right Tailed Test : In the right tailed test the reje cntirely on the right tail of the normal curve. (sce Fig. 12.2). 2. Left d Test : In the left tailed test the critical region oF reject on the left tail of the normal curve. (sce Fig. 12.3). a For example, in the test for testing the mean (1!) of the population. Null hypothesis Hy} B= Hy Alternative hypothesis H,: p> hy and Hy: h Z,) = a. cted region or critical region lies tion region lies entirely (Right Tailed Test) (Left Tailed Test) , when a two tailed test will be Selection Rejection c region (/2) | Acceptance \ region (o/2) Z=0 Fig. 12.1 For Left Tailed Test P(Z<-Z)=a Also for a two tailed test: P (|J)>Z,) = = P(Z>Z)+P(Z<-Z) =a. mm P(Z>Z,)+P(Z< Z) =a, (By symmetry) a * PZ> 2.) = > = the area of cach tail is > Scanned with CamScanner‘ Hypothesis Testing and Large Samples Tests ‘he , Rejection Rejection region (c/2) tegion («v/2) Acceptar nee region (c/2) Z-0 2, Z-0 Fig. 12.2 : Right Tailed Test at significance level Fig 12.3 : Left Tailed Test at significunce level pee table will give the critical value of Z viz., Z at 1%, 5%, 10% level of Table for Critical Values of Z, of Z Critical value | Level of Significance & (Z,) | 1% 5% 10% Two tailed Test Zl = 2.58 1, = 1.96 IZ = 1.645 Right Tailed Test Z, = 233 Z, = 1.645 Z, = 128 Left Tailed Test Z, = 233 Z, = 108 | 2, = 128 8. Decision : The last step is the decision about the null hypothesis, i.c., whether to accept it or to reject it. In this regard we compare the computed valuc of Z (obtained in step 2) with the critical value or significant value or tabled value Z, (given by table for critical value in step 7) at a given level of significance a and decide as under: (i) IZ) <|Z,, then we accept the null hypothesis H, => It the calculated valus of Z ts less than the tabled Z, of Z at a level of significance a, then the difference between [1 — £ (0)] is not significant (and this difference may be due to fluctuation of sampling), so we accept the null hypothesis. In this case the test statistic falls in the region of acceptance. (u) Mf |Z) > |ZJp then we reject the null hypothesis Hy and accept the alternative hypothesis H,. In this case the computed value of Z is numerically greater than the critical value Z, at a level of significance «, and therefore, the computed. value of test statistic falls in the rejection region, So we reject the null hypothesis and accept the alternative hypothesis at a level of significance or confidence level (1 ~ @), 42.19 THE RELATIONSHIP BETWEEN HYPOTHESIS TESTING AND CONFIDENCE INTERVAL ESTIMATION There is ordinarily a close rclationship between a test of hypothesis concerning a parameter or parameters and the corresponding confidence interval, To illustrate this relationship we me examine the (J - @) level confidence interval for jt, the mean of a normal distribution wil known variance 6? and sample size 1. 1 (wn (G/Vit) and ¥ + 24a (ov): The interval limits are: ¥ — Scanned with CamScannerSa a 286 Biostatistics Now consider a particular numerical value, say jl, between the limit that is o a o wan) Te Sie < F + 2) tara Va ‘eversing the order By the opposite sequence algebraic steps to derive the interval, that is eee ns in this ing ¥, multiplying by of terms in this inequality and using > signs instead of < signs, subtracting ¥, multiplying by ~1, and dividing by O/Jn, we have equivalent chain inequality X= Wy ~F-(a2) < Ofdn <%~(w2) or recalling that 25g =— Zp = ty Zan < critical value |Z we reject the null hypothesis H, and accep of significance a, then | at a level pothesis H,. ¢ the alternative by! mean was found to be 99. Test Example : A sample of size 400 was drawn and the sample 'n mean 100 and standard jon wil wheye6t4HS sample could have come from a normal populati deviation 8 at 5% level of significance. Solution : Here ¥ = 99, n= 400, p= 100, o=8. 1, Null hypothesis : The sample have come from a normal popw! sd. 8 ie., H, : w= 100. Alternative Hypothesis, © H, : ##100 = 2. Test Statistic. Jation with mean 100 and It is a case of two tailed test. c 8 _8 2 Heres SE. (3) = 5 = Jago ~ 20° 5 99-100__5 __o5 ‘Test Statistic Z = Q) 2 i || = 2.5 3. Level of Significance : a = 0.05. 4, Critical Value : The value of Z,, at 5% level of significance = 1.96. (from the table). 5. Decision : Since the calculated value of Z is greater than the critical value as 25>1.96 = |Z|>|Z,| => Null hypothesis H, is rejected => the sample has not been drawn from a normal population with mean 100 and s.d. 8. Example 5 : A stenographer claims that she can type at the rate of 120 words per minute. Can we reject her claim on the basis of 100 trials in which she demonstrates a mean of 116 words with a standard deviation of 15 words. Use 5% level of significance Solution : 1. Null Hypothesis : H, : Stenographer’s claim is true, ie, Hy: p= 120 Alternative hypothesis. H, + ##120 = Itis a case of two tailed test. 2. Calculation of Test Statistic : Here n = 100, ¥ = 116, w= 120, 5 = 15 Standard Error of Mean = S.E. (=) Test Statistic = Z=———— = 3. Level of Significance : a = 0,05. 4, Critical value : Z, = 1.96 from the Normal table, Scanned with CamScannerHypothe pothesis Testing and Large Samples Tests amy 5. Decision : Since |Z)= \ ull ition: Since 21 = 2.67 is peatr than [2,1 = 1.96 at 5% level of significance, so the ‘ : ejected. ty alah i Sere eecceret a i eo stenographer’s claim of typing at the rate of 120 normally di: baby pe ena with pt = 13.0 in. (33 cm) and o = 0.7 in. (1.8 cm). A group of 49 newborn pip as te @ population group living ina remote region and thought perhaps to constitute this evidence ae are studied and found to have an average chest circumference of 12.6 in. In tat the group of 49 come. jon wit 7 een evra pao from a population with parameter values different from Solution : ir 5 . -_s fad Recasting the problem in our terminology, we shall assume that the 25 observation cae: sample from a normally distributed population. If the population has the same pe jeviation as the population of presumably normal infants, then o = 0.7. We may test pothesis H : 1 = 11, = 13.0. Suppose that we agree to use a level of significance a = 0.05. Then the critical values of the test statistic z = (¥-[o)/(o/Vn) are 55 = -1-96 and Example 6 : yl ple 6 : Suppose that chest circumference of presumably normal newborn baby girls is 2o575 = 1.96. But for the sample we are given that ¥ = 12.6 and thus 126-130 __ 04 _ 4 0.7/449 (10m Because this value is less than ~1.96, it falls into the critical region and we reject the hypothesis. Our conclusion is that the infants cannot (at the 0.05 level of significance) be as the population which is normal. Notice that a 95 per cent confidence interval for the mean of the population from which the sample of 49is drawn would be ¥ £ 1.96 (a/-in) = 12.6 + (1.96) (0.10) = 12.6 + 0.196 = 12.404 and 12.796. The hypothetical value 13.0 does not fall wichin this interval. «ZExample 7 : The mean life time of sample of 400 fluorescent ight tube produced by a company is found 10 be 1570 hours with a standard deviation of 150 hours, Test the hypothesis that the mean life time of the bulbs produced by the company is 1600 hours against the alternative ‘at it is greater than 1600 hours at 1% level of significance. it is a large sample. Z= hypothesis th Solution : Here n = 400 = = 1570 standard deviation of sample mean s = 150, 0 = 1600. + The mean life time of the bulbs is 1600 hours, ie., Hy : = 1600. :p> 1600 => Itisa case of right tailed test. Also ¥ 1. Null hypothesi: Alternative Hypothesis H, 2. Computation of Test Statistic Z : _,_ 8 _ 150 _ 150 Standard error = S.E. (=) = 7 = Yaoo = 30 =15. 5 F-p_ _ 1570-1600 __ 30 Test Statistic Z = S.E.(¥) > “as. aa Scanned with CamScannerBiostatistics 290 ficance : Here 0 = 0.01 F i 1 the Right tailed 4. Critical value : The critical value of Z at 1% level of significance fo test (from the table)) is Z, = 2.33. 5. Decision : Since the calculated value of [2] is 4 which i 12Z,] (at a = 0.01) = 2.33, so the null hypothesis is rejected ant 3. Level of is greater than the critical value J the alternative hypothesis is accepted. Hence the mean life time of bulbs is greater than 1600 hours. G1 cms. Can the sample Example 8 : A sample of 900 members has a mean 3.4 ems and sd. 2.6) cms rant rr be regarded as one drawn from a population with mean 3.25 cms. Using as 0.05, is the claim acceptable? Solution : Here n = 900 => it is a large sample. 3.4 cm, s.d. of sample s = 2.61 cm, p = 3.25. Also, = 1. Null hypothesis : The sample has been drawn from a population with mean 3.25 cms, ie, Hy: w= 3.25 Alternative hypothesis. Hy: # 3.25. Tt is a case of two tailed test. 2. Calculation of test statistic: Qo error of mean = S.E. (¥) 2.61 * 4900 30° = 0,087, ¥-p oe statistics 2 Z = 3. Level of Significance : Here 0. = 0,05, 4. Critical value : From the table [2] the value of Z at & = 0.05 is |z l= 1.96. 5. Decision : Here calculated value of |Z| = 1.72 < Critical value |Z,| = 1.96 al = 1S => Null hypothesis is accepted, Hence, the sample has been drawn from a population with mean 3.25 cms. Scanned with CamScannerBiostatistics NN TWO MEANS 292 erence BETWEE! 4221 TEST OF SIGNIFICANCE OF oF cameos 6,2 antl Gr? FeSpectively " MPLES and vances oy? a : anaes ptaions with mea 1 tom these two populations Let ¥, and mie and 1) Let A and B be two p Cm , size Let us take two independent samples of size 20 ig sample means. Then! UF PT | xo popuiation means HE?’ the difference between em is _ bed test whether Hy = Ha: Or independent sample means viz, 3, be the correspondiny (i) To test the equality of To test the significance of ance difference between two (i sting the signific! 5, 2 The following steps should be taken for (¢ 5 is set in any © m from 4 an ng of following forms: 1, Null Hypothesis : The null hypothesis ee ae Hy: a = re the two samples have been draw sifferent pop’ me same means and equal standard deviation oF, es have been drawn from the same parent population H, : Hy = Hy i, the two samp! ; Alternative hypothesis : H, : }, #H, (TW° tailed test) ; or Hy:ty>H, — (One tailed test) he following two cases: ws Hy: Hy It is a case of Two tailed test. 2. Calculation of Test Statistic : Si [usar 407) _ 195 nm On, 100 200 . z S.E. (% ES |Z] = 0.769 3. Level of significance : Here 0 = 0.10, 4. Critical value : At c= 0.10 is Z, = 1.645. 5. Decision : Since calculated value of Z= 0.769 < critical value Z, (at «= 0.10) = 1.645 = H, is accepted. Hence we conclude that the two means are statistically equal. 3.9, 0, = 14.8: ' 2 S.E. of difference means :S.E. (F,- Test Statistic : Z = Example 10 : A random sample of 1000 workers from south India show that their mean wages are Rs. 47 per week with a standard deviation of Rs. 28. A random sample of 1500 workers from north India gives a mean wage of Rs. 49 per week with a standard deviation of Rs. 40. Is there any significant difference between their mean level of wages? Solution : We are given the following information. South India North India Size of Sample : nm, = 1000 Mean of Sample : =47 S.D. of Sample : 8 Population Mean = Hi = Population Mean Hy = Population mean 1, Null hypothesis : There is no significant difference between two mean level of wages, ie., Hy: By = Hy. Alternative hypothesis : H, : 1, #1, => It represents a two tailed test y Standard Error of : S.E. (¥ ,- ¥,) Test Statisti 2. Calculation of Test Statistic : x,-%, 1.36 “136 7 (28), (40y SOY = 1.36, 1000 ~ 1500 147. Scanned with CamScanner294 ‘ given so we take o = 0) - : is not give .05, A f significance : Since the level of signin ble) is 1.96 pines = 005 (from NE Ot valuc of Zyl = 1.96 = 4. Critical value : Value of Z, 3 y= 1.47 < Critical he the culated value of | nificant difference between the two means cis no sig 5. Decision : Since the cal null hypothesis is accepted => ther level of wages. ” ganl ‘xample 11 : The research unit in an org ae rent for male and scholastic aptitude test are diffe male applicant’ file are taken and summarised below: srmine whether scores on ry ishes 10 determine the isation ea applicants. Random samples i Applicants Female Male ¥ 502.1 510.5 a 86.2 90.4 7 399 204 | Using the above sample data, test the null hypothesis that the average score ae Sor the Population male and female applicants. Use 5% significance level and assume that the scores are normally distributed in each case. Solution : Let x, and x, be the random variables of female and male applicants. Then we have ¥, = 502.1 F,= 5105 5, = 86.2 5, = 904 n, = 399 04 H, = population mean. H, = population mean Population of male and female, 1. Null hypothesis : The average score is same for the he, Hy: oy = py. Alternative hypothesis : H. : Wéb > It PH, H, #p, Tepresents a Two taile: 2. Calculation of Test Statistic : ui S.E. of difference of Means = S.b, (= _ = ys iW F)= ny 86.2)? ? = (862) | (0.4) = [730 8i76 399 204 = 399 * 504 204 = 18. | VBL AOS V5868 = 7.66 Test Statistic : z = __¥,~¥, 5 14 = 1.097 *6 Tas Scanned with CamScannerHypothesis Testing and Large Samples Tests 295 3. Level of Significance : Here a = 0.05. 4. Critical Value : The value of Z,, (at a = 0.05 from the table) = 1.96. 5. Decision : Calculated value of |Z] = 1.097 < Critical value |Z,| = 1.96 => Null hypothesis is accepted => Average score is same for the population of male and female. Example 12 : The mean yield of wheat from a district A was 210 kgs, with standard deviation 10 kgs per acre from a sample of 100 plots. In another district B, the mean yield was 220 kgs, with standard deviation 12 kgs from a sample 150 plots. Assuming that the standard deviation of the yield in the entire state was 11 kgs., test whether there is any significant difference between the mean yield of crops in the two districts. Take d = 0.05. Solution : We have the following data. District’ A District B n, = 150 ¥,= 220 2 o,=1 Population Mean [1 1. Null Hypothesis : There is no significant difference between the mean yield of crops in the two districts i.e., Hy = Hy =-Hy- Alternative hypothesis : H, : 1, # #,- It represents a Two Tailed Test. 2. Calculation of Test Statistic = Oo; Oo; a e Standard Error: SE. (,- ¥,) = +5 = ie +o 1M BL _ are 0807 = VE0T7 = 1.42 150 5,-i, 210-220 _ -10 Test Statistic : Z = —te = =-705 est Statistic: Z= SE Gey ra Tap 777.05. fe [2] = 7.05 3. Level of significance : We take o = 0,05 as level of significance is mentioned in the question. 4. Critical value at a = 0,05, The critical value of Z, i.c., Z,, (at 0, = 0,05) = 1,96, 5. Decision : Since calculated |2| = 7.05 > Critical value [Z,| = 1.96 => H, is rejected => There is a significance difference between the mean yield of crops in the two districts. Scanned with CamScanner