Design of Sample Surveys Chapter 4. INTRODU! y [All items in any field of inquiry constitute a ‘Universe’ or Population.” A complete enumeration ofall items in the population’ is known as a census inquiry or census survey. It is obvious that for any Studs, or investigation census survey is rather infeasible, For example, (0 hive a idea of average per | t capita monthly income of the people in India, we will have to enumerst 1 the earning individuals in the country, which is a very difficult task. Census survey is impossible in the situations w' fe but the units are destroyed while Eg. inspection of crackers. Further, many a time iti population, and sometimes it is pos Ie to obtain suffic petotal population. In such cases there is no utility of censu However, it needs to be emphasised that when the w toa sample survey. When field studies are undertaken in practical life, con: cost almost invariably lead to a selection of respondents i.e. selection of only & few items. The respondents selected should be as represental of the total population as possible in order to produce a miniature cross-section, The selected respondents constitute what is technically Hed a tsample” Election process is called ‘sampling technique.’ The survey ves when hen population is infinite. In's spected, census survey is not at all desirable, not possible to examine every item in the scurate results by studying only a part surveys. population is fin mall one, it is no use resort iderations of time and erse is a and the “sample survey’. Algebrai sally, let the population size be N d if'a part of size 7 (whieh iss Not this populatio is selected according to some rule for studying some characteristic of the population: o units is known as “sample?. Researcher must prepare a sample J would be the group consisting of these Pee eatady ic, he ust plan how a sample should be selected anu of what size such as pipe FEE. e at ‘A sample design is a definite plan for obtaining a sample from a given population. Ut rete technique or the procedure the researcher would adopt in selecting items for the ample. design may as well lay down the number of items to he included in the sample ie. the sz @ sample of the& oe Design of Sample Surveys ple designs sample. Sample design is determined before data are collected. There are many from which a researcher can choose. Some designs are relatively more precise and easier to apply than others. Researcher must select prepare a sample design which should be reliable and appropriate for his research study. The main steps of sampling design are as follow (i) Objective: The first step of sampling design is to define the objectives of survey in clear and concrete terms. The sponsors or the researchers of the survey should confirm that the objectives are commensurate with the money. manpower and time limit available for the survey. Population: In order to meet the objectives of the survey. what should be the population? This question should be answered in the second step. The population should be clearly defined. Sampling units and fram selecting sample. Sampling unit may be a geographical one such as state, district, ete., ora construction unit such as house, flat, ete.. or it may be a social unit such as family, club, school, ete... or it may be an individual. The researcher will have to decide one or more of such units that he has to select for his study. The list of sampling units is called as “frame” ‘or sampling frame. Sampling frame contains the names of all items of a universe (in case of finite universe only), Such a list should be comprehensive, correct, reliable and appropriate. It is extremely important for the source list to be as representative of the population as possible. (iv) Size of sample: This refers to the number of items to be selected from the universe to constitute a sample. This is a major problem before a researcher. The size of sample should neither be excessively large, nor too small. It should be optimum. An optimum sample is one which fulfills the requirements of efficiency, representativeness, reliability and flexibility. While deciding the size of sample. researcher must determine the desired precision as also timate. The size of population variance needs to be considered as in case of larger Variance usually a bigger sample is needed. The size of population must be kept in view for this also limits the sample size. The parameters of interest in a research study must be kept in view, while deciding the size of the sample. Costs too dictate the size of sample that we can draw, As such, budgetary constraint must : A decision has to be taken concerning a sampling unit before lage, an acceptable confidence level for the invariably be taken into consideration when we decide the sample size. The formulation of sample size determination is provided in Chapter 10. (¥) Parameters of interest: Statistical constants of the population are called as parameters, we get the €4., population mean, population proportion ete. When we do census survey actual value of parameters. On the other hand, when we do sample surv« estimates of unknown population parameters in place of their actual values, In determining the sample design, one must consider the question of the specific population Parameters which are of interest. For instance, we may be interested in estimating the Proportion of persons with some characteristic in the population, or we may be interested in knowing some average or the other measure concerning the population. There may also be important sub-groups in the population about whom we would like to make estimates. All this has a strong impact upon the sample design we would accept. CHAPTER 4Research Methodology and no essential information ould be colle should be very much clear in the mind of iscussed in Chapter 6. al difficultic be collected for all the w the results, The reasons for non-response handled with caution (vi) Data collection: No irrelevant information sh ed should be discarded. The objectives of the surv surveyor. The methods of data collection are d (vii) Non-respondents: Becaus of practic jonse tends to chai sampled units. This non-resp should be recorded by the investigator. Such ¢ Selection of proper sam he will use i.e., he must decide sample. In fact, this technique or procedure s ands for the sarap al sample designs (explained in the pages that follow) out must choose one for his stud: “he must select that design 8 ven cost, has a smaller sampling ctor 5 of a survey depends on the frand trained personnels for the 1 design on a sniall seale before the better idea of os, data may not ases should be cher must decide the type of sampte to be used in selecting the items for the sample design itself There are fof which the researcher hich, for a given ing design: The resea ‘about the technique sample size and for a (ix) Organizing field work: ‘There should be efficient supervisory s (x) Pilot survey: It is always helpful to try out the resea going tothe field, This is called as “pilot survey” oF spretest’. Iemight gt practical problems and troubles. Cost considerations, from pra only the size of the s of a non-probability reliable field work. field work. he ctical point of view, have a major mple but also to the type of ample. (xi) Budgetary constraint: impact upon decisions relating to not sample. This fact ean even lead to the us BVM oo statacodasd FRM ecru?) NON: The errors involved in the collection of data are classified into sampling and non-sampling errors. 4.3.1 Sampling Errors Sampling errors arise due to the fact that only’ a part of the population has been used to estimate population parameters and to draw inferences about the population, Sampling errors are absent in a census survey. Sampling error can be measured fora given sample design and Be Th : Mi . wrement of sampling rors tisually called the “precision of the sampling plan’. Ifwe eee recisi a e oreasli . . ™ 7 * the pression can be improved. But increasing the size ofthe sample has its vn Jimitations viz ra d sample increases the cost of collecting data and also enhances the cr bias. Thus the effective way to increase precision is usually to select a bet ann bie Se cease eee avealeiple eels aiven cost in ter sampling desizn whieh To jess precise design because itis casier to adopt the alven coat In practice, however, eve” prefer’ bias can be controlled in a better way in anche ue and also because ofthe at that stents In brief, while i (while selecting ie Faia fs oe ae procedure, researcher must ensure that the procedure sampling error sIps way. ipling error and helps to comrol the systematic bias in a bellDesign of Sample Surveys 4.3.2. Non-sampling Errors Non-sampling errors ee Of collection and preparation of data and thus are presen both the sample surve: as well as the censtis survey, Thus the data obtained in census survey is fre from sampling emors, however subjected to non-sampling errors. Non-sampling errors ean he eles by defining the sampling units, frame and the population corrvetly in the investigations. J and by employing efficient people 44 SAMPLE SURVEY. VS. CENSUS SURVEY. 2 Sample survey, since we study only a'subpart of the whole population, requires less money and less time. Most of the times, non-sampling errors are so much large that the recalte sample survey, Rue much more accurate than those of census survey, Non-sampling errors arive due toa number or factors such as inefficieney of field workers, non-response, bias due to interviewers ere, These errors are likely to grow when the number of units inspected increases. However. if the objective of study is very serious in nature and informat SA5h and every sampling unit, there is no way out but to resort to census survey. Morewecn iftime and money are not important factors or if population is not so large. 4 cen survey may provide better results than any sample survey, provided efficient and trained statt ic employed. n is required about mati R n= ON ae ene Zhe method of selecting a sample is of fundamental importance and depends upon the and investigation, The techniques of selecting a sample are classified as. “nom = probabi and “probability sampling,”. We take up these two designs separately. ature of data sampling” 4.8.1 Non-probability Sampling Non-probabi ty sampling is that sampling procedure which does not afford any bas! for the probability that each item in the population has of being included in the sample. Non-probability sampling is also known by different names such as delibe judgement samplin, researcher; his choice ate sampling, purposive sampling and In this type of sampling, items for the sample are selected deliberately by the Concerning the items remain supreme. In other words, under non-probability sampling the organisers of the inquiry purposively choose the particular units of the universe for tepeaiuting a sample on the basis that the small mass that they so select out ofa huge one will be Epical or representative of the whole. For instance, if economic conditions of people living it a state pring BE studied, a few towns and villages may be purposively selected for intensive study on the The Ple that they can be representative of the entire state. ‘Thus, the judgement of the organisers o the study plays an important part in this sampling design. In such @ design, personal e of enter ion of the ment has a great chan fing inte the se cots and Lhe investigator may select a sample which shall yield results favou abe ois point of Sie end if that happens, the entire inquiry muy pet vitiated. Thus, there is Stes ere eet “nlcring into this type of sampling technique. ut in the investigators are impartial, CHAPTER 4and have the necessary experience so as to take sound judgement, the results obtained from an analysis of deliberately selected sample may be tolerably reliable. However. in such 2 sampling, ies is no assurance that every element has some specifiable chance of being included. Sampling error in this type of sampling cannot be estimated and the element of bias. great or small. is always there. As such this sampling design is gets rarely adopted in large inquiries of importance. However. in smal] inquiries and researches by individuals, this design may be adopted because of the relative advantage of time and money inherent in this method of sampling. Quota sampling is also an example of non probability sampling. Under quota sampling the interviewers are simply given quotas to be filled from the different strata, with some restrictions on how they are to be filled. In other words, the actual selection of the items for the sample is left to the interviewer's discretion. This type of sampling is Very convenient and is relatively inexpensive. But the samples so selected certainly do not possess the characteristic of random samples. Quota samples are essentially judgement samples and inferences drawn on their basis are not amenable to statistical treatment in a formal way. 4.5.2 Probability Sampling Probability sampling is also known as ‘random sampling’ or “chance sampling". Under this sampling design, every item of the universe has an equal chance of inclusion in the sample. It is. so to say. 8 lottery method in which individual units are picked up from the whole group not deliberately but by some mechanical process. Here it is blind chance alone that determines whether one item or the other is selected. The results obtained from probability or random sampling can be assured in terms of probability ie., we ean measure the errors of estimation or the significance of results obtained from a random sample, and this fact brings out the superiority of random sampling design over the deliberate sampling design. Random sampling ensures the law of Statistical Regularity which states that ifon an average the sample chosen is a random one, the sample will have the same composition and characteristics as the universe. This is the reason why random sampling is considered as the best technique of selecting a representative sample. 4.5.2.1 Simple Random Sampling Simple random sampling from a finite population refers to that method of sample selection which gives each possible sample combination an equal probability of being picked up and each item inthe entire population to have an equal chance of being included in the sample. This applies to sampling «without replacement ie., once an item is selected for the sample, it cannot appear in the sample as3i8 (sampling with replacement is used frequently less in which procedure the element selected for the sample is returned to the population before the next element is selected. In such a situation the same element could appear twice in the same sample before the second element is chosen). In brief. the implications of random sampling (or simple random sampling) are: (a) It gives each element in the population an equal probability of geting into the sample: all choices are independent of one another, (b) It gives each possible sample combination an equal probability of being chosen. Keeping this in view we can define a simple random sample (or simply’ random sample) fF a finite population as a sample which is chosen in such a way that each of the *C, possible same has the same probability, 1/*C,, of being selected. ‘To make it more clear we take a certain finilDesign of Sample Surveys population consisting of six elements (say 4, b, €,dse,f)ie..N = 6, Suppose that we want to take a reple of size 1 = 3 from it. Then there are ‘C, = 20 possible distinct samples ofthe required size, and they consist of the elements abc, abd, abe, abf, acd, ace, acf, ade, alf, aef. bed, bee, bef, bde, hf bof, ede, cdf cof. and def. If we choose one of these samples in such a way that each has the 20 of being chosen, we will then call this a random sample or a simple random sample. probability 1/ With regard to the question of how to take a random sample in actual practice, we could, in simple cases like the one above, write each of the possible samples on a slip of paper, mix these slips thoroughly in a container and then draw as a lottery either blindfolded or by rotating a drum or by any other similar device. Such a procedure is obviously impractical, if not altogether impossible in complex problems of sampling. In fact, the practical utility of such a method is very much limited. Fortunately, we can take a random sample in a relatively easier way without taking the trouble of enlisting all possible samples on paper-slips as explained above. Instead of this, we can write the name of each element of a finite population on a slip of paper, put the slips of paper so prepared into abox ora bag and mix them thoroughly and then draw (without looking) the required number of slips for the sample one afier the other without replacement. In doing so we must make sure that in successive drawings each of the remaining elements of the population has the same chance of being selected. This procedure will also result in the same probability for each possible sample. We can verify this by taking the above example. Since we have a finite population of 6 elements and we want to select a sample of size 3, the probability of drawing any one element for our sample in the first draw is 3/6, the probability of drawing one more element in the second draw is 2/5 (the first element drawn is not replaced) and similarly the probability of drawing one more element in the third draw is 1/4, Since these draws are independent, the joint probability of the three elements which constitute our sample is the product of their individual probabilities and this works out to 3/6 x 2/5 « 1/4= 1/20. This verifies our earlier calculation. Even this relatively easy method of obtaining a random sample can be simplified in actual practice by the use of random number tables. Various statisticians like Tippett, Yates, Fisher have prepared tables of random numbers which can be used for selecting a random sample. Generally, Tippett’s random number tables are used for the purpose. Tippett gave10,400 four figure numbers. He selected 41,600 digits from the census reports and combined them into fours to give his random numbers which may be used to obtain a random sample. __ We can illustrate the procedure by an example. First of all we reproduce the first thirty sets of Tippett’s numbers 2952 6641. «3992, 9792-7979 SOIL 3170 ©5624. «4167-9525 «ISHS 1396 7203 5356 = 1300» 2693.) 23707483 3408 2769-3563. «610769137691 0560 = 5246.=—sNI2.—=s«9025 GOB B126 ee we are interested in taking a sample of 10 units from a population of $000 units, bearing aan from 3001 10 8000, We shall select 10 stich figures from the above random numbers which ng than 3001 and not greater than 8000. we randomly decide to read the table numbers SOI 3170 et santing from the first row itself, we obtain the following numbers: 6641, 3992, 7979, »3170, $624, 4167, 7203, 5356, and 7483, CHAPTER 4@ Research Methodology The units bearing the above serial numbers would then constitute our required random sample, One may note that it is easy to draw random samples from finite populations with the aid of random number tables only when lists are available and items are readily numbered. But in some situations it is often impossible to proceed in the way we have narrated above. For example, ifwe want fo estimate the mean height of trees ina forest. it would not be possible to number the tees, ang ‘choose random numbers to select a random sample. In such situations what we should do isto select sere Gees Or ee cept Inge es ee ee random sample for study purposes. Other alternative may be to divid consider these blocks as sampling units. Height of each and eve should be measured. ple random sample, we discussed above Here. a unit which is selected once is not replaced back to the population. ie. be selected only once, Other alternative to this design is ‘simple random sample with replacement’ Because of the possibility of over-representation of some sampling units, this design is not of much practical usage. So far we have talked about random sampling. keeping in view only the finite populations. Bt what about random sampling in context of infinite populations? Its relatively difficultto explain the concept of random sampling from an infinite population. However, a few examples will show the basic characteristic of such a sample. Suppose we consider the 20 throws ofa fair dice a5 9 sample from the hypothetically infinite population which consists of the results ofall possible throws of the dice. If the probability of getting a particular number, say 1, is the same for each throw and the 20 throws are all independent, then we say that the sample is random. Similarly it would be said to be sampling from an infinite population if we sample with replacement froma finite population and our sample would be considered as a random sample ifin each draw all elements of the population have the same probability of being. selected and successive draws happen to be independent. In brief, one can say that the selection ofeach item ina random sample from an infinite population is controlled by the same probabilities and that successive selections are independent of one another. the whole forest into a finite number of equal-sized blocks ang cy tree falling into selected blocks in “simple random sample without replacement’ one sampling unit ean 45.3 Complex Random Sampling Designs Some complex random sampling designs, which are the mixture of probability and non-probabili sampling methods, are given below. 4.5.3.1 Systematic Sampling In some instances, the most practical way of sampling is to select every ith item on of this type is known as systematic sampling. An element of randomness is introduced into this ki of sampling by using random numbers to pick up the unit with which to start. For instance, ifa cent sample is desired, the first item would be selected randomly from the first twenty thereafter every 25th item would automatically be included in the sample. Thus, in systematic only the first unt is selected randomly and the remaining units of the sample are selected at FO intervals. Although a systematic sample is not a random sample in the strict sense of the ter™ bul is often considered reasonable to treat systematic sample as if it were a random sample. alist. SamplingDesign of Sample Surveys systematic sampling has certain plus points, 1 ean be taken as an improvement over a simple random sample in as much as the systematic sample is spread more evenly over the entire population Itis an easier and less costlier method of sampling and ean be conveniently used even in case of large populations. But there are certain dangers too in using this type of sampling. If there isa hidden periodicity in the population, systematic sampling will prove to be an inefficient method of sampling. For instance, every 25th item produced by a certain production process is defective. If we are to select a 4% sample of the items of this process in a systematic manner, we would either get all defective items or all good items in our sample depending upon the random starting position. [fall elements of the universe are ordered in a manner representative of the total population, ie., the population list isin random order, systematic sampling is considered equivalent to random sampli But if this is not so, then the results of such sampling may, at times, not be very reliable. In practice, systematic sampling is used when lists of population are available and they are of considerable length. 4.5.3.2 Stratified Sampling Ifa population from which a sample is to be drawn does not constitute a homogeneous group. stratified sampling technique is generally applied in order to obtain a representative sample. Under stratified sampling the population is divided into several sub-populations that are individually more homogeneous than the total population (the different sub-populations are called ‘strata’ and then we select items from each stratum to constitute a sample. Since each stratum is more homogeneous than the total population, we are able to get more precise estimates for each stratum and by estimating more accurately each of the component parts, we get a better estimate of the whole. In brief, stratified sampling results in more reliable and detailed information. The following three questions are highly relevant in the context of stratified sampling: (a) How to form strata? (b) How should items be selected from each stratum? (c) How many items be selected from each stratum or how to allocate the sample size of each stratum? Regarding the first question, we can say that the strata be formed on the bi of common characteristic(s) of the items to be put in each stratum. This means that various strata be formed in Such a way as to ensure elements being most homogeneous within each stratum and most heterogencous between the different strata. Thus, strata are purposively’ formed and are usually based on past experience and personal judgement of the researcher, One should always meres dal endertion of he rea an rsh ma Condu ae fee esloated ote nomi by ff efficient stratification plan. We can do so by taking sre actczmining & more approprinns We osed stata and then examining the variances Within a StmPles of equal size from each of he Fran appropriate stratification plan for our and among the possible stratifications, we can dect¢ PPLOP inquiry, Inrespect of the second question, we can say that the usual method, for selection of items for the Sample fr a i Je random sampling, Systematic sampling can tom each stratum, resorted to is that of simp! at “Sed if it is considered more appropriate in certain situations. CHAPTER 4@ Research Methodology Regarding the third question, we usually follow the method of proportional allocation under wh the sizes ofthe samples from the different strata are kept proportional to the sizes of the tata, Te is if? represents the proportion of population included in stratum and n represent the tal sp vie. the number of elements selected from stratum iis m . P To illustrate it, let us suppose that we ‘vant a sample of size n ~ 30 to be drawn from a population of size N = 8000 which is divided int three strata of size N, ~ 4000, N_ = 2400 and N, = 1600. Adopting proportional allocation, we shal gx the sample sizes as under for the different strata: For strata with N, = 4000, we have P, = 4000/8000 and hence n, =n . P, = 30 (4000/8000 Similarly, for strata with V,=2400, we have P= 30 (2400/8000) = 9, and for strata with N, = 1600, we have P= 30 (1600/8000) = 6. n, ‘Thus, using proportional allocation, the sample sizes for different strata are 15,9 and 6 respectively proportion to the sizes of the strata viz., 4000 : 2400 : 1600. Proportional allocation: Considered most efficient and an optimal design when the cost of selecting an item is equal for ea stratum, there is no difference in within-stratum variances, and the purpose of sampling happens i tstimate the population value of some characteristic. But in case the purpose happens to compar the differences among the strata, then equal sample selection from each stratum would be mor efficient even if the strata differ in sizes. In cases where strata differ not only in size but also! | variability and it is considered reasonable to take larger samples from the more variable strata a Sinaller samples from the less variable strata, we can then account for both (differences in satu size and differences in stratum variability) by using disproportionate sampling design by requiring: which is n/N Sx ny/NyOy = 12/302 | and 6, denote the standard deviations of the k strata, N,, N....y N, denote denote the sample sizes of k strata. This is called ‘optim ing. The allocation in such a situation result where 6), 62+ sizes of the k strata and 11,5 Ryy---+ My ‘allocation’ in the context of disproportionate samp] the following formula for determining the sample sizes different strata: n-N,G i=1,2 = Wig, + Nz G2 tt NO, for i= 1,2, ... and k. n, We may illustrate the use of this by an example, Example 4.1: 4 population is divided into thre 1, = 3000, Respective standard deviations ae: 6, = 15, 6, = 18 and 63 =5, -e strata so that N, How should a sample of size 1 = 84 be allocated to the three strata, if we want optimum alloct using disproportionate sampling design’? “Solution: Using the disproportionate sampling design for optimum allocation, the sample size different strata will be determined as under:Design of. ‘Sample Surveys Sample size for strata with V, = $000 / 84(5000) (15) (5000) (15) + (2000) (18) + (3000) (5) = 6300000/126000 = 50 n Sample size for strata with N, = 2000 me 84(2000) (18) * (8000) (15) + (2000) (18) + (3000) (5) = 3024000/126000 = 24 Sample size for strata with V, = 3000 84(3000) (5) (5000) (15) + (2000) (18) + (3000) (5) = 1260000/126000 = 10 In addition to differences in stratum size and differences in stratum variability, we may have differences in stratum sampling cost, then we can have cost optimal disproportionate sampling design byrequiring ny ny NiO} ne Oe VO where C, = Cost of sampling in stratum 1 Cost of sampling in stratum 2 > C,= Cost of sampling in stratum k and all other terms remain the same as explained earlier. The allocation in such a situation results in the following formula for determining the sample sizes for different strata: a n-Njo/{G, "Wa, + N,0IC +4 MOC, f is not necessary that stratification be done keeping in view a single characteristic. Populations, ae often stratified according to several characteristics. For example, a system-wide survey designed ‘0 determine the attitude of students towards a new teaching plan, a state college system with 20 Colleges might stratify the students with respect to class, section and college. Stratification of this Me is known as cross-stratification, and up to a point such stratification increases the reliability of "mates and is much used in opinion surveys. From what has been stated above in respect of stratifi oe isthe result of successive application of purposive (involved instr first hayenPling methods. As such itis an example of mixed sampling. T Stratification and then simple random sampling is known as strat for i= 1,2... ve can say that the sample so fication of items) and ed sampling, con: rocedure wherein We ied random sampli@® Research Methodology 4.5.3.3 Cluster Sampling Ifthe total area of interest happens to be a big one, a convenient way in which a sample can be taken is to divide the area into a number of smaller non-overlapping areas and then to randomly select a weber ofthese stmafler areas (usually called clusters), with the ultimate sample consisting ofall (or samples of) units in these small areas or clusters. “Thus, in cluster sampling the total population is divided into a number of relatively small sub- divisions which are themselves clusters of still smaller units and then some of these clusters are randomly selected for inclusion inthe overall sample, Suppose, we want to estimate the proportion of achine-parts in an inventory which are defective. Also assume that there ae 20,000 machine parts in the inventory at a given point of time, stored in 400 cases of 50 each. Now, using a cluster Sampling, we would consider the 400 cases as clusters and randomly seleet‘7" eases and examine all the machine-parts in each randomly selected case. Cluster sampling, no doubt, reduces cost by concentrating surveys in selected clusters. But certainly its less precise than random sampling There is also not as much information in '7t observations sethin cluster as there happens to be in‘n’ randomly drawn observations. Cluster sampling is used only because ofthe economic advantage it possesses; estimates based on cluster samples are usually more reliable per unit cost. Ifclusters happen to be some geographic sub-divisions, in that case cluster sampling is better known asarea sampling, In other words, cluster designs, where the primary sampling unit represents ‘cluster of units based on geographic area, are distinguished as area sampling. The plus and minus points of cluster sampling are also applicable to area sampling. 45.2 4 Multi-stage Sampling Multi-stage sampling isa further development of the principle of cluster sampling. Suppose, we was to investigate the working efficiency of nationalised banks in India and we want to take a sample of few banks for this purpose. The first stage is to select large primary sampling unit such as states ia a country. Then we may select certain districts and interview all banks in the chosen districts. This Ho represent a two-stage sampling design with the ultimate sampling units being clusters of jistricts. . ave of taking a census of all banks within the selected districts, we select certain towns ay int Eewall banks in the chosen towns, This would represent a three-stage sampling design. If tia faking s sansa of all banks within the selected towns, we randomly sample banks tte?! cach selected own, then itis a case of using fourstage sampling plan. If we select randomly atall ewe wile what ow ay nls nom sampling desi ; 5 oesearty sau sage sampling is applied in big inquires extending to a considerable La geographical ac, say, thee There are two advantages of this sampling design Vie+ (2) esr o administer than not singe stage designs nal because of the fact th et -stape evelope: fi ae ze sampling is developed in partial units. (b) A large number of U sampled for a given cost J ‘en cost under multi-stage si 8 because of: i : sample rains on stage sampling because of sequential clustering, whereas this sampling its eanDesign of Sample Surveys 4.5.3.5 Sampling with Probability Proportional to Size In case the cluster sampling units do not have th luster sam J me number or approximately the same number of elements, it is considered appropriate to use a random selection process where the probability of cach cluster being included in the sample is proportion 0 we have to list the number of elements in e: cluster. Then we must sample sys to the size of the cluster. For this purpose, h cluster irrespective of the method of ordering the matically the appropriate number of elements from the cumulative totals. The actual numbers selected in this way do not refer to individual elements, but indicate which clusters and how many from the cluster are to be selected by simple random sampling or by systematic sampling. The results of this type of sampling are equivalent to those of a simple random sample and the method is less cumbersome and is also relatively less expensive. We can illustrate this with the help of an example. Example 4. a The following are the number of departmental stores in 15 cities: 35, 17, 10, 32, 70, 6, 19, 26, 66, 37, 44, 33, 29 and 28. If we want to select a sample of 10 stores, using cities as clusters and selecting within clusters proportional to size, how many stores from each city should be chosen? (Use a starting point of 10). Solution: Let us put the information as under (Table 4.1): Since in the given problem, we have 500 departmental stores from which we have to select a sample of 10 stores, the appropriate sampling interval is 50. As we have to use the starting point of 10°, so we add successively increments of 50 till 10 numbers have been selected. The numbers, thus. obtained are: 10, 60, 110, 160, 210, 260, 310, 360, 410 and 460 which have been shown in the last column of the table (Table 4.1) against the concerning cumulative totals. From this we can say that two stores should be selected randomly from city number five and one each from city number 1, 3, 7, 9.10.11, 12, and 14, This sample of 10 stores is the sample with probability proportional to size. Table 4.1 City mumber enc 1 35 2 52 3 62 4 32 5 710 6 28 a : 26 * 19 9 26 10 — _ u 37 - 12 rr a —_ a 3) a 0 1s 38 500 “Ithe starting point is not mentioned, then the same can randomly’ be selected CHAPTER 4® Research Methodology 4.5.3.6 Sequential Sampling This sampling design is some what complex sample design. The ultimate size of the sample under ths technique is not fixed in advance, but is determined according to mathematical decision rules on the Inasis af information yielded as survey progresses. This is usually adopted in case o! anypling plan in contest of statistical quality control, When a particular lot is to be rejected on the basis of a single sample, itis Known 4s single sampling; when the dec taken on the basis of two samples, it is known as double sampling and in case the decision rests on the basis of more than two samples but the number of samples is certain and decided in advance. the sanypling is knovwn as multiple sampling But when the number of samples is more than two but itis Neither certain nor decided in advanee, this type of system is often referred to as sequential sampling “Thus, in brief, we ean say’that in sequential sampling, one can go on taking samples one after another as long as one desires to do so. rom abrief description of the various sample designs presented above, we ean) that normally one should resort to simple random sampling because under it bias is generally eliminated and the sampling errorean be estimated, But purposive sampling is considered more appropriate when the ‘universe happens to be small and a known characteristic of itis to be studied intensively. There fare situations in real life under which sample designs other than simple random samples may be considered better (say easier to obtain, cheaper or more informative) and as such the same may te used. In a situation when random sampling is not possible, then we have to use necessarily 8 sannpling design other than random sampling. At times, several methods of sampling may well be used in the same study. 1. Whatdo you mean by “Sample Design’? What points should be taken into consideration by ar wrcher in developing a sample design for this research project. 2. How would you ditfereniate between simple random sampling and complex random sampling desis Explain clearly giving examples. - 3. Why probability sampling is generally preferred in compari ili 7 4 t parison to non-probability sampling? EXP the procedure of selecting a simple random sample, eee 4. Under what circumstances stratified random sampling design is consi ratified ra pling design is considered iat you select such sample? Explain by means of an example <2 5. Distinguish between: 2 How wou! (a) Restricted and unrestricted sampling; (b) Convenience and purposive sampliny (©) Systematic and stratified sampling; (8) Cluster and area sampling.Design of Sample Surveys 6. Under what circumstances would you recommend: (@) A probability sample? (b) Anon-probability sample? (©) A stratified sample? (@ Acluster sample? Explain and illustrate the procedure of selecting a random sample. How will you optimize sampling and non-sampling errors? ee systematic bias results from errors in the sampling procedures”. What do you mean by such a systematic bias? Describe the important causes responsible for such a bias. 10. (2) The following are the number of departmental stores in 10 cities: 35,27, 24, 32, 42, 30, 34, 40, 29 and 38. If we want to select a sample of 15 stores using cities as clusters and selecting within clusters proportional to size, how many stores from each city should be chosen? (Use a starting point of 4).. (b) What sampling design might be used to estimate the weight of a group of men and women? A certain population is divided into five strata so that N, = 2000, N, = 2000, N, = 1800, N= 1700, and 1, = 2500. Respective standard deviations are: 6 = 16, 0) = 2.0,03= 4.4, 64= 48, 05= 60 and further the expected sampling cost in the first two strata is Rs 4 per interview and in the remaining three strata the sampling cost is Rs 6 per interview. How should a sample of size n=226 be allocated to five strata if we adopt proportionate sampling design; if we adopt disproportionate sampling design Considering (i only the differences in stratum variability (i) differences in stratum variability as well as the differences in stratum sampling costs. i. CHAPTER 4