SAMPLING Meaning of Basic Terms

Population: Any finite or infinite collection of individuals sharing some common set of characteristics. Population Elements: Individual participant or object on which the measurement is taken. Census: The complete enumeration of the elements of a population. Sample: A subgroup of the elements of the population selected for participation in the study. Sampling: Selecting some of the elements in a population, from which we may draw a conclusion about the entire population. Sample Survey: A survey which is carried out using a sample.

Sampling Design Process

1. Define Target Population: Target population is the collection of elements or objects that posses the information sought by the researcher. It should be defined in terms of elements, sampling units, extent and time. Element is the object that posses the information ex. Housewives. Sampling unit is the basic unit containing the elements of the population. Ex. Household. Defining Sample frame: A representation of the elements of the target population. It could be a directory, list of members of any association etc. Before using the suitability of the frame must be examined. Whether it covers the population adequately, up-to-date, convenient. Select Sampling Technique Determine the sample Size Execute the Sample Process

2.

3. 4.

Types of Sampling Techniques

Non-probability sampling: That does not use chance selection procedure. Rather, rely on the personal judgment of the researcher. Arbitrary or conscious decision of the researcher to select the elements. Convenience, judgment, quota, snowball Probability sampling: Method in which each element of the population has a fixed probabilistic chance of being selected for the sample. Simple random, systematic, stratified, cluster(area)

Non-probality: Convenience
Obtaining a sample of such elements which are convenient to the researcher. Co-operative and easily accessible element are selected. Suitable for exploratory researches where researcher is interested in inexpensive approximation or gross estimation. Not recommended for descriptive or causal researches or marketing researches. Sometimes used even in large surveys.

Least expensive and least time consuming. High chances of biasness.

Not representative of population.

Non-probability: Judgment
A refined or improved form of convenience sampling in which element are purposely selected based on the judgment of the researcher. Based on sound judgment or expertise of researcer. Low cost, convenient and quick Subjective, the value depends on the expertise of the researcher. Does not allow generalisation Can be used when broad population inferences are not required.

Non-probability: Quota
Two-stage judgmental sampling First, the population is segmented into mutually exclusive groups based on certain characteristic such as gender, age, race etc. Second, sample elements are selected based on convenience or judgment. Quota ensures that the composition of the sample is same as the population. Characteristics within a characteristic, if not taken care-of (which is difficult), sample will not be representative. Widely used Convenient, cheaper and quick Under certain conditions obtain results close to probability methods Selection biasness No assurance of representativeness

Non-probability: Snowball
Selection of elements with desired characteristic starts with usually randomly. Subsequent respondents are selected based on the referral or information provided by the initial respondents. It substantially increases the likelihood of locating desired characteristics in the population. Low cost but time consuming Useful when identification of respondents is difficult at the beginning of the study.

Probability: Simple Random

Each element in the population has a known and equal chance of getting selected.

Every element is selected independently of every other element. Sample is drawn by a random procedure from a sampling frame. Lottery method Random table method No personal biasness Representative sample for a homogeneous population Easily understood Results are projectable and can be generalised. Expensive No assurance of representativeness Difficult to administer when population is large and dispersed

Probability: Systematic
Choosing the sample by selecting a random starting point and then picking every kth element in succession. K (skip interval)=Population size/ sample size Procedure: 1. 2. 3. 4. 5. Identify the list and number of elements in population Identify the sample size Identify skip interval Identify random start Draw sample Less costly and easier May not represent the whole population

Probability: Stratified
Two-stage sampling First, sample frame is divided into non-overlapping groups or strata such as gender, age group. Secondly, elements are selected from each stratum by a random procedure. The elements in each stratum should be as homogeneous as possible. Upto six strata. Can be proportionate or disproportionate. In proportionate, size of sample is in proportionate to the relative size of that stratum in population. In disproportionate, same sample size from each stratum is selected.

Includes all possible subpopulation- better coverage of population Difficult to identify appropriate strata. Expensive Not possible to stratify on many variables.

Probability: Cluster
First, the target population is divided into mutually exclusive and collectively exhaustive subpopulations cluster Then, a random sample of clusters is selected based on a probability sampling technique. For each selected cluster, either all the elements are selected or some elements are drawn probabilistically. No unit from the non selected unit is included. Area Sampling: Cluster sampling in which the clusters consists of geographical areas such as blocks, districts etc. Economical, convenient and easy to administer Difficult to compute and interpret result

Stratified vs Cluster
Stratified Cluster

All strata are included

A sample of clusters is chosen

Each stratum should be homogeneous

Each cluster should be heterogeneous

Strata should be heterogeneous

Clusters should be homogeneous

Needed for the entire population

Needed only for selected clusters

Elements selected from each stratum randomly

All elements from each selected cluster are included

Choice of Probability and Non probability Sampling

Probability Non-probability

Suitable for Exploratory Research

Suitable for Conclusive9 Descriptive and Causal) Researches

For homogeneous population

For heterogeneous population

Statistical considerations unfavourable

Favourable

Operational considerations unfavourable Time efficient Economical

Favourable Time consuming Costly

A Good Sample Design

True representative of the population Precision Biasness and accuracy Small sampling error Viable to the fund available Generalization with reasonable level of confidence