UNIT I DATA COLLECTIONPopulation • All the items that fall under the purview of research are called “Universe” or “Population”. • Algebraically represented by “N”.Sample: • It is the small unit of the population, that represents all the characteristics of the population. • The sample is selected by using several techniques; called Sampling Techniques. • Algebraically represented by “n”.Census and SurveyCensus: It is the Enquiry done on the population. i.e. it is the enquiry that covers all theitems in the population. More complicated Involves great deal of time, money and energy. Can be done only by government.Example: population census.Survey: it is the study conducted on the selected few items called sample. TOPIC II SAMPLINGConcept of Sampling: • Sampling Method means selection of a limited number of items representing the population or universe for studying the characteristics of the whole population or universe. • Example: to know the IQ of the Students of age between 15-16, suppose in the class of 70 students. We conduct the study on 20 students who represent the class.Population = 70Sample = 20Essentials of Sampling • Sample should posses same characteristics as the population. • Absolute accuracy is not essential. • Regulating conditions should be same for every individual item in the sample.Advantages of Sampling • the result obtained is generally more reliable than that obtained from a complete count. • Total financial burden of a sample survey is generally less than that of complete census. • Possible to collect more detailed information in a sample survey. • Causes less damage and wastage.
Disadvantages: • Shortage of experts in the sampling field is a serious hurdle in the way of reliable statistics. • Sampling plan may be complicated that it requires more time, labour and money than a complete count. • Must be carefully planned and executed otherwise the results obtained may be inaccurate and misleading.Reasons for Sampling: 1. Universe Size 2. Financial Constraints 3. Sufficiency of an approximation 4. Time Constraints 5. Destructive Nature of smapling TOPIC III SAMPLING DESIGNSample DesignIt comprises of: 1. Sampling Frame 2. Selection of sampling items 3. Sample SizeTypes of Sampling Design • Probability Sampling: • Non-Probability Sampling: Simple Random Convenience Stratified sampling Judgment Area or Cluster Quota Multi-stage random Panel Systematic PurposiveSimple Random Sampling • Merits: Represents universe in the better way Desired level of precision can be achieved by increasing or decreasing the sample size. More scientific method. • Demerits: May not be true representative if its size is small. Units of population should be dependent. Needs a complete list of finite population, without.
Stratified Sampling • Merits: Assures representativeness Decreases chances of excluding units of universe. representative character can be achieved with fewer items. Replacement of units is easily possible. Saves time & money • Demerits: Requires accurate knowledge of the universe. If stratified list is not available, it will be costly to prepare the same. Bias or error may be made in the sample through improper stratificationCluster or Area Sampling • Merits: If clusters are geographically defined, yield lowest field costs. Requires listing only individuals in selected clusters. Characteristics of clusters as well as those of population can be estimated. • Demerits: Larger Errors for comparable size than other probability samples, and Requires ability to assign each member of population uniquely to a cluster; inability to do so may result in duplication or omission of individuals.Multi-stage Random Sampling: • Merits: Complete listing of universe is not required. If sampling units are geographically defined, then, it cuts down field costs • Demerits: Errors are likely to be larger than the other types. Errors will increase as number of selected sampling units decrease.Sampling with probability proportional to size: • Merits: Equivalent to simple random sampling. It is less cumbersome. It is less expensive. • Demerits:
Sequential Sampling: • It is repetitive and can give more accurate results.Judgment Sampling • Merits: Reduces the cost of preparing sample and field work, since ultimate units can be selected so that they are close together • Demerits: Variability and bias of estimates cannot be measured or controlled, and Requires storing assumptions or considerable knowledge of population and sub-group selected.Quota Sampling: • Merits: Same as in judgement sampling. Introduces some stratification effect. • Demerits: Introduces bias of observer’s classification of subjects and non-random selection within classes.