Presentation on theme: "Presenter – Anil Koparkar Moderator – Bharambhe sir"— Presentation transcript:
1Presenter – Anil Koparkar Moderator – Bharambhe sir Sampling TechniquesPresenter – Anil KoparkarModerator – Bharambhe sir
2Framework Introduction Need and advantages Methods of sampling Probability samplingSimple Random Sampling – With & Without ReplacementStratified Random SamplingSystematic Random SamplingCluster SamplingNon probability samplingConvenient samplingJudgment samplingQuota samplingUses of samplingReferences
3A famous sampling mistake That’s TrumanThey only asked rich, white people with telephones who’d they vote for. Sadly, they published their mistake
4Few Terms Who is the target group for the study? This is called the study populationWho in the target group should be surveyed?This is called the sample.How many people should be surveyed?This is called the sample size.How should the people to be surveyed?This is called the sampling method.
5Need of sampling techniques Census requires enormous time, trained personnel, money etc. and slightest bias can get magnified when no. of observations are increased.Thus most of the time its not feasible to collect data from whole population.Thus with sampling techniques data collected and analyzed in much less time and money and also reduces inspection fatigue
6What is Sampling ?Procedure by which some members of a population are selected as representative of the entire population. ORProcedure of selection of part of an aggregate to represent the whole population.The sub-group thus selected to represent the whole population is known as SAMPLEThus sampling method is the scientific & objective procedure of selecting unit from a population and provides a sample that is expected to be representative of the population as a whole.
8what exactly IS a “sample”? A subset of the population, selected by either “probability” or “non-probability” methods. If you have a “probability sample” you simply know the Probablity of any member of the population being included (not necessarily that it is “random.”)
10Simple Random SampleDef:- The technique of selecting sample in which there is equal probability of individual getting selected at each draw.A sample selected such that each possible sample combination has equal probability of being chosen. Random does not mean haphazard.Two types of Simple Random Sampling 1 ) Simple random sampling without replacement 2 ) Simple random sampling with replacementA sample selected such that each possible sample combination has equal probability of being chosen. Also called Unrestricted random sampling.Two types of Simple Random Sampling 1 ) Simple random sampling without replacement• In this method the population elements can enter the sample only once• The units once selected is not returned to the population before the next draw2 ) Simple random sampling with replacement• The population units may enter the sample more than once .
11Simple Random Sample Get a list or “sampling frame”-coding Let N= no of units in population thenSelection/identification random numbers between 1 to NMethods of selection/identification of a random number:Lottery MethodTable of Random numbersRandom number selections using computer3. Identify sample unit according to random number generated.(Decoding)Sampling frame:- It is identification & listing of all unis comprising the study population & allocating then unique Identity no.
12Simple random sampling without replacement Simplest and basic method of drawing samples.The probability of including a specified unit in a sample of size n at rth draw is 1/N-(r-1). (N=total population size, n =is sample size)
13Simple random sampling with replacement Definition: It is a method of sampling such that every one of the possible sample of size n from N has the same probability.The probability of drawing a specified unit at the rth draw is equal to the probability of drawing it at the 1st draw.The population units may enter the sample more than once This method is used when there is follow up in any survey. E.g School health surveyThe probability of including a specified unit in a sample of size n is n/N. (N=total population size, n =is sample size)probability of including a specified unit in a sample of size n at rth draw = 1/N
14Methods of selection of a simple random sampling: Lottery MethodRandom number proceduresBy computer (Demo)By random number table (Demo)
15Lottery methodEx. Suppose 20 patients from 100 is to be selected by lottery method, then-All the 100 patients can be given serial number 1 to 100.100 pieces of paper with equal size will be marked with 1 t0 100 no they are then folded and shuffled.Draw out 1 & note the no.Replace the piece of paper drawn. Again repeat the process. Reject if same no is selected. Repeat till 20 samples are drawn. Now decode those numbers to select proper subject from our population 100.
16Random number procedures By computer (Demo)By random number table (Demo)Example:- nine blocks in a certain administrative zone contain 793, 170, 970, 657, 1721, 864, 383 and 826 households respectively. If we want to select 6 households using a method of SRS Without Replacement.Answer
17Advantage and disadvantage of SRS Advantages of Simple random samplingIt is simple techniqueGives equal chances of selection for every individual from population.Disadvantages of Simple random samplingIf study is repeated, same samples couldn’t be identifiedIf population is heterogeneous, then SRS is not good technique to draw samples.If population is divided in different strata, then SRS may leave representatives from some strata.
18Stratified Random Sample Also sometimes called proportional or quota random sampling.Stratification means division of population into mutually exclusive and exhaustive groups.This method divide the population into non-overlapping (mutually exclusive)groups (i.e., strata) of size N1, N2, N3, ... Ni, such that N1 + N2 + N3 Ni = N.Identify size of sub-samples (n1, n2, n3,…ni; n=n1+n2+n3….ni)ni=(ni/N) × n.Then using SRS in each strata to get sample of size ni.Combined together to form the required sample from the population.Separate your population into groups or “strata”Do either a simple random sample or systematic random sample from thereNote you must know easily what the “strata” are before attempting thisIf your sampling frame is sorted by, say, school district, then you’re able to use this method
19Stratified Random Sample Advantages• Every unit ina stratum has thesame chance of beingselected.• Using the same sampling fraction for all strata ensuresproportionate representation in thesample of thecharacteristic being stratified.• Adequate representation of minority subgroups of interestcan be ensured by stratification and by varying the samplingfraction between strata as required.Disadvantages• The sampling frame of the entire population has to beprepared separately for each stratum.• Varying thesampling fraction between strata, to ensureselection of sufficient numbers in minority subgroups forstudy, affects the proportiona I representativeness of thesubgroups in the sample as a whole.
20Systematic Random Sample • Divide the population size by the sample size, to get sampling fraction• Select a random number between 1 and sampling fraction, which is the first sampling unit• Systematically select the remaining sample units, by adding sampling fractionSelect a random number, which will be known as kGet a list of people, or observe a flow of people (e.g., pedestrians on a corner)Select every kth personCareful that there is no systematic rhythm to the flow or list of people.If every 4th person on the list is, say, “rich” or “senior” or some other consistent pattern, avoid this method
21Systematic Random Sample Select a random number, which will be known as kGet a list of people, or observe a flow of people (e.g., pedestrians on a corner)Select every kth person.Select a random number, which will be known as kGet a list of people, or observe a flow of people (e.g., pedestrians on a corner)Select every kth personCareful that there is no systematic rhythm to the flow or list of people.If every 4th person on the list is, say, “rich” or “senior” or some other consistent pattern, avoid this method
22Merits of Systematic random sampling 1. Procedure is simple and convenient to use. 2. Relatively less labor and time is needed. 3. If the population is sufficiently large and homogeneous and if numbering of subjects are available, then this method can provide good results. 4. If start number (1st sample) is known, whole samples can be determined. Thus sampling can be repeated or checked.
23Demerits of Systematic random sampling If engulfed in any hidden cyclic trend, then this technique can be dangerous. E.g. studying no of issue of books per day. If sampling interval (k)=7, then it will induce error.
24Cluster Sample• Used in large scale investigations in wide geographical area.• First stage- preparation of large sized sampling units• Randomly selecting a certain number• Second stage- Another list prepared from them• Sub-samples drawn by random sampling
25Cluster SampleIf we have to do a survey of town in India..
26Multi-stage Cluster Sample E.g. 30×7 cluster survey…Arrange clusters in ascending or descending order of population or according to alphabetical order. Calculate cumulative frequency.Identify sampling interval. Population/30.Identify random number. Say K.So 1st cluster will fall in Kth population n later cluster will be calculated by K+1SI, K+2SI,….etc. th cumulative population.Randomly sample people within those clusters.Advantages• Cuts down onthe cost of preparing a sampling frame.• Cuts down onthe cost of travelling between selected units.• Eliminates the problem of 11packing 11 (in health surveys,especially those involving case finding and treatment, it isnot unusual for neighbouring houses not included in thesample to transfer their households temporarily to aselected house).Disadvantages• Sampling error is usually higher than for a simple randomsample of the same size
27Multi- Stage Sampling Stages 1 2 3 4 5 6 Country State District Taluka When sampling procedure is carried out in several stagesThis procedure is used in large scale country wise or region wise surveys. • First stage- preparation of large sized sampling units• Randomly selecting a certain number• Second stage- Another list prepared from them• Sub-samples drawn by random sampling and so on….Stages123456CountryStateDistrictTalukavillagesindividuals
28Multi –Phase Sampling Used to obtain supplementary information Certain items of information collected from all units of sample Other items collected from only some of sampling unitsEx. In health examination survey among school children
29Non probability sampling the techniques which do not provide any basis for estimating the probability of item in the population for getting included in the sampleOther characters-–Representativeness is in question as sampling error cannot be measuredMore suitable for small in-depth enquiries than large surveys.Saves time and money (speed and administration convenience)More flexible
30Convenience/ haphazard Sampling In this technique samples are selected at the convenience of the researcheruseful in formulative or explorative studies, pilot surveys, testing questionnaire, ..etcEx. 1. Choosing fruits from basket, pilot testing of thesis questionnaire, The "person on the street" interviews conducted frequently by television news programs. Sampling those most convenient. Gets a quick reading of public opinion.Sampling those most convenientGets a quick reading of public opinionAlso called Haphazard or Convenience Sampling
31Purposive Sampling Sampling with a purpose in mind Handpicking supposedly typical or interesting casesReaches a targeted sample quicklyTypesJudgment samplingQuota sampling
32Judgment samplingResearcher Purposively or deliberately draws a sample which he/she thinks is representative.Thus personnel biases of investigator have great chances.Ex. selecting talkative children for interviewing to find the cause of tobacco chewing, Selecting mothers who are willing to participate in study of Health Care Seeking for Newborn Danger Signs
33Quota samplingselection is based on some basic parameters like age, sex, income etcfield workers are assigned quotas of number of units satisfying the required characteristics for collecting dataPropertieswhen parameters are large, the number of cells increases.It misleads if relevant parameter is omitted.In this method field worker tend to visit respondents who are more likely to be available and accessible.
34Quota samplingEx. Communication behavior of user to be carry out on a quota sample from a population having the following parameters:%age of grad., post. Grad.& doctorates are 20, 35, 45 resp.Male: female = 60:40If sample size is 200, find the quota sampling as per above 2 parameters.
35Quota samplingAns:- Qualification ratio/ proportion G:P:D = 4:7:9, gender ratio/ proportion M:F = 3:2Therefore graduate male = 4/20 × 60 = 12 or 3/5 × 20 = 12.Likewise all cells can be filled and then samples are collected purposively.Distribution of particulars of population in %
36Distribution of particulars of population in % Quota sample in numbers as per parametersSex\QlnGradPost grad.DoctTotalMale12212760Female8141840203545100Sex\QlnGradPost grad.DoctTotalMale244254120Female162813680407090200
38Uses of sampling techniques In population census – Data on various items e.g. occupation, income, migration, housing, etc…..are collected on sampling basis.To provide counter-checks on data collection.To speed up tabulation & publication of results.To increase operational efficiency in business and industries e.g. estimating size of readership of news magazine – c/a Opinion poll survey.In purely experimental investigations.E.g. Determination of quality of milk, EQA in microscopy quality checks in RNTCP.
39References:Sukhatme PV, Sukhatme S, Sukhatme BV, Asok C. Sampling theory of surveys with applications. Iowa state university press, Ames, Iowa, USA. 3rd edLwanga SK, Cho YT, Ayeni o. Teaching Health Statistics Lesson And Seminar Outlines. WHO Geneva,Sampling Methods in Medical Research. Available online at URL:Rao KV. Biostatistics. A manual of statistical methods for use in health, nutrition = and anthropology. 2nd ed. Jaypee brothers medical publisher (P) ltdSridhar MS. Research methodology; part 5 – sampling and sampling strategy. Available online at URL:Amol R Dongre, Pradeep R Deshmukh and Bishan S Garg. Awareness and Health Care Seeking for Newborn Danger Signs Among Mothers in Peri-Urban Wardha. Indian Journal of Pediatrics. May 9,Dongre AR, Deshmukh PR, Murali N, Garg BS. Tobacco consumption among adolescents in rural Wardha: Where and how tobacco control should focus its attention?. Indian Journal of Cancer ; Sept 2008: 45 (3). 101