MopuLeé 5 ts (animate or inanimate) under study is called population or ‘collection of individuals or of their attributes (qualities) or of results of etee, ct, specified. a finite number of individuals or members is 2 finite the universe of the weights of students in a particular c number of members is known as an infinite universe. For at various points in the atmosphere. can happen is called a hypothetical universe. The universe a coin an infinite number of times (provided that it does h the problem of discussing universe of which he complete enumeration is impracticable. For average per capita income of the people of India, country is a very difficult task. Naturally, the Lis of which we can examine only a limited in of the Theory of Sampling. A sample is thus a small portion of the called the sample size. The process ofef —_—a TECHNIQUES - III find if they are properly cooked and ‘ % ¢2Spoonful of it. contain the proper quantity of salt or sugar, by 41, Probability sampling and 2. Non-probability sampling Inthe ee the reseee knows the exact possibility of selecting each member of pppalation while in the latter, the chance of being included in the sample is not known. A sasbility sample tends to be more difficult and costly to conduct. However, probability samples ‘ue only type of samples where the results can be generalized from the sample to the gation. In addition, probability samples allow the researcher to calculate the precision of estimates obtained from the sample and to specify the sampling error. Non-probability samples, incontrast, donot allow the study’s findings to be generalized jmthe sample to the population. When discussing the results of a non-probability sample, researchers must limit his/her findings to the persons or elements sampled. ‘This procedure also does not allow the researcher to calculate sampling statistics that mide information about the precision of the results. The advantage of non-probability sam- fagis the case in which it can be administered. probability samples tend to be less complicated and less time consuming than prob-4is unable to subdivide the population from which tn, jive income categories prior to drawing the sample. °° umpling: In judgemental or purposive sampling. the judgement about who to include in the sample gr" lare used in selecting the respondents or clemente't” ¢ of sample would be a i d to those persons who li ee past experience indicates that most of the use of ny lis living within two miles. However, by limiting the sammn 0% may not be reliable if the usage ch ics ofthenew ho lexperienced. As with all non-probability sampling metnags "introduced by the researcher cannot be measured and « st commonly used for conducting probability s, ling: Simple random sampling provides the b: methodologies are derived sampling, the researcher must first prepare an exhausti of the population of interest. From this list, the sample is has an equal chance of being drawn during eacl is done without replacement ; that is, a person or item rom the population for all subsequent selections. At any the population not already drawn. To draw a simple researcher bias, computerized sampling programs and ly select the members of the population to be sample would be a survey of state employees. An 3 as of a certain date could be obtained from the 00 names were selected from this list using a random program, then a simple random sample would be has the advantage of reducing bias and “ha nd the precision of the estimates derived throug! d random sampling involves categorizing sive and collectively exhaustive fore from each group. Stratified samp’ a notemiques - "I! ee 397 and rural groupings. Ifthe differences ini fj SZ aur an the income differences within the regions of Gata peal upin, isi Jeet coved. in addition, if the research organization has trench eee te pis pa eng : located in Pos. the administration of the survey can be decentralized and perhaps conducted in f° gpofficient manner ie — ae Ad fy cluster Sampling: C luster sampling is similar to stratified sampling because the inet be sampled is subdivided into mutually exclusive groups. However in cluster yy, Pigg, the groups are defined so as to maintain the heterogeneity of the population. It is if a garcher'S goal to establish clusters that are representative of the population as a whole, fly Mh in practice this may be difficult to achieve. After the clusters are established, a g te random sample of the clusters is drawn and the members of the chosen clusters are fig poi fall of the elements (members) of the clusters selected are sampled, then the sampling ine is defined as one-stage cluster sampling. If a random sample of the elements of gaselected cluster is drawn, then the sampling procedure is defined as two-stage cluster sapling. Cluster sampling is frequently employe d when the researcher is unable to compile a ‘eprchensive list of all the elements in the population of interest. A cluster sample might be cibyaresearcher attempting to measure the age distribution of persons residing in Mumbai. frould be much more difficult for the researcher to compile a list of every person. residing in finbai than to compile a list of residential addresses. In this example, each address would Jements contained in the clus- ‘present a cluster of elements (persons) to be sampled. If the el r pulation, then estimates derived from cluster sampling are ‘sare as heterogeneous as the po] Be as thove soon Smple random sampling. However, ifthe heterogeneity of the clus- unis less than that of the population, the ease ee =e pee mee Par (iv) Systematic Sampling: Systematic sampling, 2 07 ee : rien used in place of ciple aaaeaieaamnnne: 1 systematic es ee Bes every n neniben atte selecting the first through ” element as the starting : : a orresides to sample every 20°" member ofthe population, a er example, if the rese? point for the sample is randomly selected from the first yer cent sample, the starting Pol! spe of cluster sample because each of the first members. A systematic luster that contains 5 percent of the population. “members of the sampling gystematic sample instead of a simple random tas tend to be easier to draw and execute. The forward through the sampling frame to draw i may spread the members selected for mea- pulation than simple random sampling. ‘Therefore, me more representative of the population and more mn systematic sampling is that this method can allow without complete prior knowledge of the sam- wrs to the state’s publications desk could be con- randomly selecting the first through 10% visitor as in this manner, it would not be necessary for the ye list of visitors prior to drawing the sample. mpling, systematic sampling is as precise as simple contained in the clusters are as heterogeneous as theA TEXTBOOK OF ENGINEER INg w an Ifthis assumption is not valid, then systematic sam indom sampling, In conducting systematie sampling, itis agg = 8 Pg doesnt introduce bias into the sample by selecting an inappaye sl an dee, when conducting a sample of financial records, or oto 8 schedule, the researcher would not want to select "7" ae the et sample would then be comprised of observations that were all. thee ’ of-the-week influences may cause contamination of the on : e day sau i mple, giving the Testa ling Will be leg, ultiStage Sampling: Mult-stage sampling slike cluster sampling sample within each chosen cluster, rather than including all uni he Iti-stage sampling involves selecting a sample in at least two stages i Groups or clusters are selected. These clusters are designed tocotan ot than are required for the final sample. oy second stage, population units are chosen from selee imore than two stages are used, the process of choos s until the final sample is achieved. of multi-stage sampling is where, firstly, electoral sub-division dum a city or state, Secondly, blocks of houses are selected from wile and, thirdly, individual houses are selected from within the sleaa ted clusters to derive af ‘ing population units witis of multi-stage sampling are convenience, economy and efficiency. Muli Not require a complete list of members in the target population, wich Preparation cost. The list of members is required only for thosedus final stage. The main disadvantage of multi-stage sampling is the same ss! accuracy due to higher sampling error. |" iE ts of the population such as mean, the variance ete. Seed | concepts of the sample from the members of the aol pulation from which the sample has been drawn's samples | a ce are denoted by |i and o2, while those of the a < Cain ling distribution of a statistic is nt role in the theory of large samples tis Wy statistic, for large sample. 7 = g.0Statistic Standard error olin o*/2n 3 | Difference of two sample means ¥ ~ z ot , oe | | mo Ny 4. | Difference of two sample standard deviation s, ~s, amas we Qn,” Ong 5. | Difference of two sample proportions p, ~p, a 7 = | 6 | Observed sample proportion p PY 58 TEST OF SIGNIFICANCE An important aspect of the sampling theory is to study the test of significance which will enable us to decide, on the basis of the results of the sample, whether (i) the deviation between the observed sample statistic and the hypothetical parameter value or Xs (ii) the deviation between two sample statistics is significant or might be attributed due tochance or the fluctuations of the sampling. is which is tested for possible rejection under the Hi, in clear terms.ATEXTBOOK OF ENGINEERING May Dilty of the value of the variate falling in the critical region i A region corresponding to a statistic ¢ in the sample space § i null hypothesis Hy is called as critical region or ro to acceptance of Hy is called acceptance re, of the statistic ¢ belongs to the critical region is mn 8 ly e ich aroun gion, The rt icon yh Krom: The probability qt own as the level otsiga Pite w/H,) =a ‘ignit i size of the type I error (refe q significance is the size o1 LPs Teler art. 5.7) or the Maxima the appropriate level of significance in advance depending on the reliability og statistic (or test criterion): We compute the test statistic 2 under the larger samples corresponding to the statistic t, the variable 2 = = a d with mean 0 and variance 1. The value of z given above under the nul) as test statistic. n: We compare the computed value of z with the critical value 2, at 1). The critical value ofz,, of the test statistic at level of significance for by, icAL TECHNIQUES = II ee Level of significance 1% (0.01) 5% (0.05) en Two tailed test 12,122.58 Right tailed test 2, = 2.33 lz l= ee Left tailed test 2,=-2.33 y= aS 2,=-1: If |2| > 2. We reject Hy and conclude that there is significant ai zept Hy and conclude that there is no significant differece on erence: If [2] <2, we 10 ERRORS IN SAMPLING The main aim of the sampling theory is to draw a valid conclusion about the population a on the basis of the sample results. In doing this we may commit the following two pes of errors: Type I Error When Hy is true, we may reject it. P(Reject Hy when itis true) = P(Reject H,/H,) = oa sealed the size of the type I error, also referred to as producer's risk. Type II Error: When H, is wrong, we may accept it. P(Accept H, when it is wrong) = ‘cept H,/H,) = B. B is called the size of the type II error, also referred to as consumer’s fk, Note. The values of the test statistic which separates the critical region and acceptance region ‘rcalled the critical values or significant values. This value is dependent on (i) the level of signifi- ‘ce used and (ii) the alternative hypothesis, whether it is one tailed or two tailed, j1_TEST OF SIGNIFICANCE FOR LARGE SAMPLES _ If the sample size n > 30, the sample ‘ely distributions, as Binomial, Poisson,rv Ps 3. in ed proportion of successes are P+ 3 Gy the obser he probable limits for the population arep +, for the proportion in le limits the probabl known, eS ee NP ne a ‘Significant va an estim: pand 2, is the signifi nes id chi taken as . Proportion, tae ofAL TECHNIQUES i Null hypothesis H,: The manufact i: Sol. urer claim is accepted i.e., 5 ote Q=1-P=1 0.05 =0.95 100 ~"” alternative hypothesis: P > 0.05 Hence we use Right tailed test, 30 _ observed proportion of sample = —— = P , 'p! 400 0.075 Test statistic Under H,, the test statistic. z= =r aiummnenE Deena /PQin (005% 095 ~ 2.2941. 400 Conclusion: The tabulated value of z at 5% level of significance for right tailed test is 2,= 1.645. Since | z | = 2.2941 > 1.645, Hyis rejected at 5% level of significance, 2. the snqartion of defective is larger than the manufacturer claim, To find 95% confidence limits of the proportion. Itis given by P+2,./PQin ie, 0.05 + 1.96 pos 086 = 0.05 + 0.02135 = 0.07136, 0.02865, of defective bolts are (0.07136, 0.02865). t Hence 95% confidence limits for the proportion [rerommow ‘wero born in a week. Do these figures confirm the 1. Ina hospital 475 female ‘porn in equal number? hypothesis that males ane . ‘and out of them 540 ‘are vegetarian and the rest are ‘of eating (vegetarian or non-vegetarian) are significance (ii) 5% level of significance? found to be smokers. Does this information in the city are smokers? 2 d at 5% level, accepted at 1% level rence of Proportions os n, and n respectively taken from two different “fference between the sample proportions p, and Hi, that there is no significant difference between , where P= ™P*7P2 ond Q=1—P. mya hlATEXTBOOK OF ENGINEERINg Ma ILLUSTRATIVE EXAMPLES | . Before an increase in excise duty on tea, 800 people out of q to be tea drinkers. After an increase in the duty, 800 persone ue 10a) ‘a sample of 1200 people. Do you think that there has been « "to umption of tea after the increase in the excise duty? 4 Sigifican, n, = 800, n. = 1200 X,_ 800 4 _X,_ 800 2 tee, 1000 6°72 n. 1200 3 Pim + Pog _ Xi +Xy _ _800+800 _ 8 3 1 + Mg my +ny ~ 100041200 11° °* esis Hy: p, =p, i-., there is no significant difference in the eonsumptiog “increase of excise duty. ot othesis H,:p, >?» 0.8 — 0.6666 ee i=, | aes ra( 2+) 11“ 111000 * 1200 e calculated value of | z | > 1.645 and also | z | > 2.33, boththe and 1% level of significance, hence H, is rejected ice, thereiss iption of tea due to increase in excise duty. luced 16 defective articles in a batch of 500. After overhai- tch of 100. Has the machine improved? = 6.842,TEST YOUR KNOWLEDGE of 400 men and 600 women were asked whether they would like to have a school j, Random “efenidence. 200 men and 325 women were in favour of proposal. Tet the hypothesis Meat the proportion of men and women in favour of the proposal are same at 5% level of significance. were 956 births of which 52.5% were males while in towns A and B combined, jon in total of 1406 births was 0.496. Is there any significant difference in the propor- is proport ea male pirths in the two towns ? Answers 1, H,: Accepted 2. H, : Rejected. ee {113 Testing of Significance for Single Meat e i : To test whether the ference betioen ample mean cd ppusain met i seni” wnot. ’ Let X,, X, x, be random sample asin rom leer ay Xgy ooeeen r of size n from a population with dsize N with mean and variance 0. the standard error of mean of a random sample variance 6? is of ‘To test whether the fie,, to test whether the d tant or not under the null iy o . o Foe pei ae it are ¥-1.96 Fr 1.96, the significant value, sted i.e., the sample is not drawn from the popu on) given by = + 2.58 o/yn = 64+ 2.58/10sample of 200 measurements from Af. ‘Determine 95% confidence ital or jeqiance 0,2. The test statistic isence tests were given to two groups of boys and girl, irls, Mean S.D. Size 5 8 60 73 10 100 nce between mean scores is significant. H,: There is no significant difference betw ‘een mean scores a ne calculated value of | z | < 1.96, the significant value of at 5% nted i.e,, there is no significant difference between mean sere 00 and 2000 farms gave an average yield of 2000 kg and 2050kg farms in the country may be taken as 100 kg. Examine ficantly in yield. sf 1000 and 2000 members are 168.75 ems and 170 ems esp sd as drawn from the same population of standardple. Random samples drawn we jeights of adult males: (i) Is the difference between the n (i Is the difference between the Sol. Given: n, = 1000, n, = Since the samples size are () Null hypothesisATEXTBOOK OF ENGINEERING. SANCE OF SMALL SAMPLES vofthe sample is less than 30, then the sample is called he possible for us to assume that the random sampling y*etple pmormal and the values given by the sample far ttbutn’ are Ps jalues and canbe used in their place for theealeulaton yy RIBUTION (t-Test) (G.B.T.U. (MBA) 2011 ; G.B.7U, cuca, > MCA) 2019, on of the sample ‘s’ is given then t-statistic is defined as ad probability integral of t-distribution. These of freedom and when the degrees of lent to normal distribution and the probTECHNIQUES - Ill : we use the statistic sample mean % and the population eur a ne ma where g= [2-3 SiVn at ipeeeree of freedom n — 1. Atgiven level of significance o: and degrees of freedom (n ~ 1 jalor one tailed). If calculated ¢ value Jit] > tay Hy is rejected. si Fiducial Limits of Population Mean sani Ift,is the value of at level of significance a at (n ~ 1) degrees of freedom then, i"TEST YOUR KNOWLEDGE duals are chosen at random from a normal population of students ang y.. tobe 63, 63, 66, 67, 68, 69, 70, 70, 71, 71. In the light of these data, dice’ Mi mar, i f students is 66. » CASCUSS the », ate ‘mark of the population of s cere values gives the lengths of 12 samples of Egyptian cotton taken from 52, 45, 43, 47, 47, 46, 45, 50. Test if the mean length of th Consign © consignment ean of 18 items has a mean 24 units and standard deviation 3 units. Test the h ‘arandom sample from a normal population with mean 27 uni YPothesis sample of 10 boys had the I.Q.’s 70, 120, 110, 101, 88, 83, 95, 98, 107 and 10, py rt the assumption of a population mean I.Q. of 160? "Doth Xn Ids Yn, of sizes two normal populations with mean j1, and , respectively under ation variance are equal (o, = 6, = 9). re been drawn from the normal population with means j, andj, ns of the two samples. ic t is given by t = 222) + deviations s,, s, are given then we have S'= 7 5, = Beary ~ 5)? + Ely — ny +My -2 1otal hypothesis: Hy Hy = Hye, hypothesis: ‘alternativehypothesis Hy: py > Hy (one tailed test) ‘of two sample means : Eh 7 n 2.34 | 3.34 | 4.34 7.0756 | 2.7556 | 1.7956 | 5.4756 | 11.1556 | 18.8356 jficance for 13 d.f. is 1.77. @fenen-?yup of 5 patients t ot patients from the i G2 kg. Do you agree with Gat the value of at 10% ‘rest whether the dif 5, The average number ard deviation 20 and regard both the machit» independent samples have been drawn from the normay Poulain, ILLUSTRATIVE EXAMPLES Two random samples drawn from 2 normal populations are as fo independent samples have been drawn from the same population we f the means by applying the t-test and (ii) equality of population nes that the sample variances are equal, we shall first applythe jis Hy: 0,2 =o i.e., the population variance do not difirelusion: The table vah calculated value of F wees 3 fend that the variability in two p ‘test: Null hypothesis Hy, = pi iaiternative hypothesis Ht, Test of statistic 2K, -X,ATEXTBOOK OF ENGINEERING ~1 dF for lated 89, Since the tabulated value of F is less than the caleule is no significant difference between the variance, the normal population with same variance. (ii) the equality of variance. that the sample variance are equal, we first apply Fest of = 03 not differ significantly.- oat TECHNIQUES =I aires 9.490 for (nj + ny ~2) df. mom * “V16 Conclusion: The caleulated value of | t | ve! of significance is 1.96, Since the calculated value is greater than the tabulated value, Hy is rejected, there is significant difference between means, Le, Hy # Hy. "Since there is significant difference between means, and no significant difference srance, we conclude that the samples do not come from the same normal population. TEST YOUR KNOWLEDGE 1. From the following two sample values, find out whether they have come from the same population: Semplet | 7 | a7 | 18 | 25 | 27 | 2 | a7 | 2s 2 Sample? | 16 | _167]] 1260) 18 | Noo) [alia [oa 2 The daily wages in Rupees of skilled workers in two cities are as follows: is 9.490. The tabulated value oft at 39 d.f. for Size of sample of workers S.D. of wages in the sample aes ch City A ; 16 City B 13 3. The standard deviation respectively. Can the| A TEXTBOOK OF ENGINEERING yn. Mar EE, =N (total frequency) and degrees of freedom (d.£) = (n= 1), (@ Ifz2 = 0, the observed and theoretical frequencies agree exactly. (ii) If 2 > 0, they do not agree exactly. ES OF FREEDOM of freedom. e to choose any four numbers whose sum is 50, we can exercise our i ‘any three numbers only, the fourth being 50 minus the total of the three rag sd. Thus, though we were to choose any four numbers, our choice was reduced ty ‘one condition imposed. There was only one restraint on our freedom and our a were 4—1= 3. If two restrictions are imposed, our freedom to choose will ‘and degrees of freedom will be 4 — 2 = 2. ‘the number of degrees of freedom is the total number of observations less lependent constraints imposed on the observations. Degrees of freedom (df) y table, the expected frequency of any class in which it oocurs the most general test known. Itis applicable ©2°°7 h can be summed up under theTECHNIQUES — III y 423 non | number of fre i ON, the total ‘equencies should be reasonabl; is dil titutes largeness, but as an arbitrary f y Neree, Its dine at cons! igure, we may say that N should be at least #% owever, few the cells. (d) No theoretical cell-frequency should be small. Here again, it is difficult to say what pstitutes smallness, but 5 should be regarded as the very minimum and 10 is better. If au theoretical frequencies occur (i.e., < 10), the difficulty is overcome by grouping two or sore classes together before caleulating (O — B). It is important to remember that the jumber of degrees of freedom is determined with the number of classes after regrouping. Note 1. If any one of the theoretical frequency is less than 5, then we apply a correction given by Yates, which is usually known as ‘Yates correction for continuity’, we add 0.5 to the cell frequency sbich is less than 5 and adjust the remaining cell frequency suitably so that the marginal total is not hanged. Note 2. It may be noted that the y? test depends only on the set of observed and expected fre- quencies and on degrees of freedom (4.f.). It does not make any assumption regarding the parent popu- lation from which the observations are taken. Since x2 does not involve any population parameters, it is termed as a statistic and the test is known as Non-parametric test or Distribution-free test. 621 THE x? DISTRIBUTION For large sample sizes, the sampling distribution of y? can be closely approximated by a continuous curve known as the chi-square distribution. The probability function of x distri- bution is given by fe) = (2) er where e = 2.71828, v = number of degrees of freedom; c = a constant depending only on v. Symbolically, the degrees of freedom are denoted by the symbol v or by dif. and are obtained by the rule v = n —R, where k refers to the number of independent constraints. In general, when we fit a binomial distribution the number of degrees of freedom is one ls than the number of classes; when we fit a Poisson distribution the degrees of ecdom ave the arithmetic mean 2less than the number of classes, because we ‘ to get the parameter of the Poisson di Blsearo tee aera degrees of freedom are 3 less than the PUSGET ee total frequency, mean and standard ; h : * If the data is given in a . B ae In the case of Binomial Z In the case SE In the cas 522 TESI— ATEXTBOOK OF ENGINEERIN IG Mary EWAN Ss | ILLUSTRATIVE EXAMPLES 1. In experiments on pea breeding, the following frequencies of seu 'S werg Wrinkled and Round and Wrinkled and Tot yellow green green a 108 32 iz s that the frequencies should be in proportions 9 Examine te ental result support the theory i.e., there is no significant difference d and theoretical frequency. theoretical (expected) frequencies can be calculated as follows: _ 556x9 116 _ 556 x3 16 E, = 312.75 = 104.25nical TECHNIQUES ~ Il oa, Null hypothesis H, : The accidents are uniformly distributed over Under this Ho, the expected frequencies ofthe aesident on ech of the Conclusion: Since the calculated val repted i., the accidents are unifoi | ATEXTBOOK OF ENGINEERING wary MATiog ‘Null hypothesis H,: The data are consistent with the hypothesis of “male and female births. i.e, p = q = 1/2. Ye use Binomial distribution to calculate theoretical frequency given by. p Ni) = Nx P(X =1r)=Nx"C, p’g"-" Rk is the total frequency, N(r) is the number of families with r male children dq are probabilities of male and female births respectively, n is the nuke, : a qual prop, 236 a | 200 5 | 1296 ws | 6.48 3.92 5% level of significance for 5 - 1= 4 df. is 9.49. the calculated value of x? is greater than the tabulated value, His re not consistent with the hypothesis that the Binomial law holds snd birth is not equal to that of a female birth. / licts the proportion of beans in the four groups, Gy G+ 23:3: 1. Inan experiment with 1600 beans the numbersx thejoor Does the experimental result support the theory? z experimental result support the theory. i. there ved and theoretical frequency. iTECHNIQUES ~ II qublular value of 2 at 5% level of Significance for 3 at * onclusion. Since the calculated value of 72 is i accepted. i.e., the experimental rat Example 6. The following table shows the dist m a telephone directory: Sol. Null hypothesis Hy : The there is no significant difference Under H,, the expected f Calculation of x2ae ATEXTBOOK OF ENGINEERING Man 4 ay ney of x mistakes per page is given by the poisson law as follow. ~F N(@) = N- P(x) 242.05(0.4821)* e =e ; = 30 |- ay 0 F, ‘abulatea» the null hypothesis is rejected and the differ- tate is significant and! wel Cometic satis ee ot ene eee machines.show variations more than could be covered of F for 0, = 2 and v, = 12 at 5% level ofjnstiCAL TECHNIQUES ~ I 449 gl ‘aatstical ae Control” when statistical techniques are employed to control, improve god maintain quality or to solve quality problems. Building an information system to satisfy jpe concept of Prevention and control and improving upon product quality requires statistical hiking . Statistical quality control (S.@.G.)s systematic as compared to guess-work of haphazard cess inspection and the mathematical Statistical approach neutralizes personal bias and Treovers poor judgement. S.Q.C. consists of three activities: (1) Systematic collection and graphic recording of accurate data (2) Analyzing the data (3) Practical engineering or management action if the information obtained indicates significant deviations from the specified limi Modern techniques of statistical quality control and acceptance sampling have an important part to play in the improvement of quality, enhancement of productivity, ereation ofconsumer confidence and development of industrial economy of the country. Following statistical tools are generally used for the above purposes : ( Frequency distribution: Frequency distribution is a tabulation of the number of times a given quality characteristic occurs within the samples. Graphic representation of frequency distribution will show : (a) Average quality (6) Spread of quality (c) Comparison with specific requirements _(d) Process capability. (ii) Control chart: Control chart is a graphical representation of quality characteris- tics, which indicates whether the process is under control or not. (iii) Acceptance sampling: Acceptance sampling is the process of evaluating a por- tion of the product/material in a lot for the purpose of accepting or rejecting the lot on the basis of conforming to a quality specification. It reduces the time and cost of inspection and exerts more effective pressure on quality improvement than it is possible by 100% inspection. Itis used when assurance is desired for the quality of materials/produets either produced or received, ‘ed (iv) Analysis ofA TEXTBOOK OF ENGINEg; F detection of faults: In statistical quality control after plot (&, R, P, C, U) etc. when the points fall above the upper control limits font, ‘control limit, it gives an indication of deterioration in quality. Negece, action may be then taken immediately. 4. Adherence to specifications: So long as a statistical control continues, sre pe accurately predicted for future by which it is possible to assess whether thems are capable of producing the products with the given set of specificatig tia "5: Increases output and reduces wasted machine and man hours. 6. Efficient utilization of personnel, machines and materials results in high ity ler ty. cessary om re Creates quality awareness in employees. However, it shoul is not a panacea for assuring product quality. s a common language that may be used by all three groups design juction personnel and inspectors in arriving at a rational solution of mutual problems.” "9. Points out when and where 100% inspection, sorting or screening is required, jlimination of bottlenecks in the process of manufacturing. id be noted that staisg ‘action can be based. Without such action, the method is ineffective, ’ re ‘the application of standard procedures is also very dangerous without adequae similar things are rarely obtained in nature. This fact holds good fe well. No production process is good enough to produce all items The variations are due to two main reasons: Variations due to chance causes are inevitable ina difficult t6 trace and to control also even under ‘esti ona be due to some inherent characteristic of the process chance factors alone, the observations will r of chance variation is the foundation 0” s which produce these variations are ace e causes, then in this case, ‘as it may be concluded thaoa ~ Te ~ — aL TECHNIQUES =I ns 451 j) Differences among materials i (jv) Differences in each of these factors over time (v) Differences in their relationship to one another. e variations may also be caused di i ; These variations may al ised due to ch in worki it fi pepaptof operator ee mee. ‘king conditions, mistake on t sat TECHNIQUES OF STATISTICAL QUALITY CONTROL, 1, Process Control: It is a process of monitoring and measuring variability in the serformance of a process or a machine through the interpretation of statistical techniques and iisemployed to manage in-process quality. This technique ensures the production of requi- ste standard product and makes use of control charts, 2, Product control: This technique is concerned with inspection of already produced aaods to ascertain whether they are fit to be despatched or not. To achieve the objectives, it makes use of sampling inspection plans, To control the quality characteristics of the product, there are two main techniques : 532 CONTROL CHART A control chart is a graphical representation of the collected information. It detects the ‘ariation in processing and warns if there is any departure from the specified tolerance limits. Inother words, control charts is a device which specifies the state of statistical control or is a device for attaining quality control or is a device to. whether the statistical control has been attained. s 7 The control limits on the chart are so placed lose the presence or absence of the assignable causes of quality variati nosis possible and brings sub- stantial improvements in product r 0ATEXTBOOK OF ENG! wn process can meet the existing specificati line or to tell whether the process is in ee me Tol andi, y determine whether a givel change in the production ‘To secure informatio 4. To secure information _5.'To provide a basis for c product. i jon to be used in establishing or hat i i 6. To secure information to be i g or changing inspection procedure op procedure or both. RUCTION OF CONTROL CHARTS FOR VARIABLES - to be used in establishing or changing production pr ve it is necessary to widen the tolerances, surrent decisions or acceptance or rejection of manufact tured or taken during a manufacturing process overg of all, a random sample of size n is x, are noted and quality measurements *,, 2, ae le mean ga n n taal range R= pax, ~*min ‘process is found stable, # consecutive samples are selected and for each sample lated ‘Then we find ¥ and R as = when tolerance limits are not given| |, when tolerance limits are given is upper control limit Band UCL (for X-chart) = 2+ Aak are be found from the following table