RESEARCH NOTE PART 6SELECTION OF SUBJECTS1. Population2. Sample3. Sampling Frame4. Sample Size5. Sampling Procedures:-Probability:Simple Random SampleStratified SampleCluster Sample-Non ProbabilityConvenience JudgmentQuota

POPULATIONThe aggregate of the elements defined prior to selection of the sample. All members of any well-defined class of people, events, or objects.Target PopulationAccessible Population

SAMPLEA small group drawn from the population and contains the characteristics of the population.PURPOSE OF SAMPLE To obtain information concerning the population. Concept of inductive reasoning is part of the scientific approach. Method involves making observation and then drawing conclusions from those observations. Sample must be representative if one is to be able to generalize with confidence from the sample to the population

SAMPLING Which sample/organization to study out of the total population of organizations. Samples are often chosen by precise procedure that aim to represent the population. Select a sample that will facilitate the analysis to be made of the data. - Provides a means of identifying and locating the population elements. Organization of the frame often exerts a strong influence on the sampling design. Ideal sampling frame would list each population element once and once only.

SAMPLING FRAMEA sampling frame is a list of all the sampling units available for selection at a stage of the sampling process. At the final stage, the actual sample is drawn from such a list. Some of the more creative thinking in a research project may be related to the specification of a sampling frame. A frame may be a class list, list of registered voters, a telephone book, an employee lists, or even a map. In the case of a map we would be sampling pieces of geography. A city block would be an example. The frame list may be printed or stored in a computer file, on tape or disk. Once a population is specified, you then search for a good sampling frame. Often the availability of a sampling frame defines the population, as no perfect fit is available between population and frame.

SAMPLE SIZETo determine an appropriate sample size: First, specify the degree of precision required. Eg. 95% Confidence Interval - should be a sample percentage plus or minus 2.5%. This specification, thus requires that 1.96 SE (p) =2.5%, where p= population percentage. Non response rate has to be taken into consideration.

FACTORS DETERMINING SAMPLE SIZE1. Number of groups to be analyzed2. Value of information and accuracy required of the results.3. Cost of accessing the sample4. Variability of the population

SMALL OR LARGE SAMPLES?Between the economy and convenience of small samples lies a trade off point balancing practical considerations against statistical power and generalizability. Small samples are more appropriate for in depth case studies or where complex techniques of eliciting or evaluating behaviour are involved, such as psychodrama, role playing, intensive counseling, or interviewing procedures, or projective instrumentsNevertheless, it remains true that the larger the sample, the smaller the sampling error and other things being equal, it is preferable to increase the sample size wherever practical.

STEP IN SELECTING A SAMPLESTEP 1: Define the population. This would include: the elements, the sampling units, the extent, and the time.STEP 2: Identify the sampling frame from which the sample will be selected.STEP 3: Decide on a sample size. Here we determine how many elements to include in the sample. Deciding when a sample is too big or too small is a difficult problem.STEP 4: Select a specific procedure by which the sample will be determined. Exactly how will the decision be made on which population elements to include in the sample?STEP 5: Physically select the sample based upon the procedure previously described.

SAMPLING PROCEDURESThere are many different procedures by which researchers may select their samples, but one fundamental concept must be established at the outset - the distinction between:(1) a probability sample, and (2) a non probability sample. PROBABILITY SAMPLINGIn probability sampling, each element of the population has a known chance of being selected for the sample. The sampling is done by mathematical decision rules that leave no discretion to the researcher.A known chance and not an equal chance of being selected. An equal-chance probability sampling is only a very special case of probability sampling called sample random sampling. The probability sample allows the calculation of the likely extent to which the sample value differs from the population value of interest. This difference is called sampling error.

TYPES OF PROBABILITY SAMPLING1. RANDOM SAMPLING2. STRATIFIED SAMPLING3. SYSTEMATIC SAMPLING4. CLUSTER SAMPLING

RANDOM SAMPLINGThe best known of the sampling procedures is random sampling. The basic characteristic of random sampling is that all members of the population have an equal and independent chance of being included in the sample. That is, for every pair of elements X and Y, Xs chance of being selected equals Ys chance, and the selection of X in no way affects Ys probability of selection.

The steps in Random Sampling are:1. Definition of the population2. Listing of all members of the population.3. Selection of the sample by employing a procedure where sheer chance determines which members on the list are drawn for the sample.With a small population one might put the name or identification number of each member on a slip of paper and place these slips in a box, mix them well, than blindly draw the needed number of slips from the box. Those slips drawn make up the sample.

A more systematic procedure for drawing a random sample is to refer to a Table of Random Numbers. This is a table containing columns of digits that have been mechanically generated, usually by a computer, to assure a random order. Each member in the population is identified with a distinct number and then n numbers are selected from the Table of Random Numbers. The n numbers chosen from the table represent the n elements selected for the study.

STRATIFIED SAMPLINGWhen the population consists of a number of subgroups or strata that may differ in the characteristics being studied, it is often desirable to use a form of sampling called stratified sampling. For example, if one were conducting a poll designed to assess opinions on a certain political issue, it would be advisable to subdivide the population into groups on the basis of age or occupation. This is because one would expect opinions to differ systematically among various age or occupational groups. IN stratified sampling, one first identifies the strata of interest and then draws a specified number of subjects from each stratum. The basis for stratification may be geographical or it may involve characteristics of the population, such as income, occupation, gender. In the study of youths, for example, one may be interested not merely in surveying the attitudes of youths who reside in small towns with those who live in medium-size and large cities. In such a case, one would divide the youth population into three groups, based on the size of the towns in which they reside, and then randomly select independent samples from each stratum.

SYSTEMATIC SAMPLINGStill another form of sampling is called systematic sampling. This procedure involves drawing a sample by taking every kth case from a list of the population.One first decides how many subjects he wants in the sample (n). Since he knows the total number of members in the population (N), he simply divides N by n and then determines the sampling interval (k) to apply to the list.

The first member is randomly selected from the first k members of the list, and then every kth member of the population is selected for the sample. For example, let us assume a total population of 500 subjects and a desired sample size of 50 (n); thus, k = N/n = 500/50 = 10.

One would start near the top of the list so that the first case could be randomly selected from the first ten cases, and then every tenth case thereafter would be selected. Say the third name or number on the list was the first to be selected. Then the sampling interval k or 10 is added to 3 and the thirteenth person falls in the sample, so does the twenty-third and so on, adding the constant sampling interval until the end of the list is reached.

Systematic sampling differs from simple random sampling in that the various choices are not independent. Once the first case is chosen, all subsequent cases to be included in the sample are automatically determined.

CLUSTER SAMPLINGIt is very difficult, if not impossible, to list all members of a target population and select the sample from among them. The population of Malaysia secondary school students, for example, is so large that one cannot list all of its members for the purpose of drawing a sample.

In addition, it would be a very expensive undertaking to study a sample that is scattered all over the country. In this case, it would be more convenient to study subjects in naturally occurring groups or clusters. That is, the researcher would choose a number of schools randomly from a list of schools and then include all the students in those schools in his sample.

This kind of sampling is referred to as cluster sampling since the unit chosen is not an individual but a group of individuals who are naturally together. These individuals constitute a cluster insofar as they are alike with respect to characteristics relevant to the variables of the study.

NON PROBABILITY SAMPLINGIn non-probability sampling, the selection of a population element to be part of the sample is based in some part on the judgment of the researcher. There is no known chance of any particular element in the population being selected.

Therefore, it is impossible to calculate the sampling error that has occurred. There is no way to determine whether or not the sample estimates calculated from a non probability sample are accurate or not.

CONVENIENCE SAMPLINGConvenience samples are selected, as the name implies, on the basis of the convenience of the researcher. Examples are:(1) asking for people to volunteer to test products and using these people as respondents, (2) stopping people in a shopping mall to get their opinion, (3) using students for conducting an experiment, (4) having people in the streets interviewed by a television interviewer, etc. In each instance, the sampling unit or element is self-selected or has been selected because it was easily available. In all cases, it is unclear as to what population the actual sample is drawn from. The television interviewer may state that her sample represents the community. Clearly, she is wrong. Most members of the community had no chance of being selected. It is only those who happened to be where the interviewer was at that time of the show who had a chance of being selected. Even the exact chance of these people being selected is unknown.

In such cases, the difference between the population value of interest and the sample value is unknown, in terms of both size and direction.

We cannot measure sampling error, and clearly we cannot make definitive or conclusive statements about the results from such a sample. However, convenience samples can be most easily justified at the exploratory stage of research , as a basis for generating hypotheses and for conclusive studies where the manager is willing to accept the risk that the study results might have great inaccuracies. Convenience sampling is extensively used in practice.

JUDGEMENT SAMPLINGJudgment samples or purposive samples are selected on the basis of what some experts thinks those particular sampling units or elements will contribute to answering the particular research question at hand. For example, in test marketing, a judgment is made as to which cities constitute the best ones for testing the marketability of a new product.

In industrial marketing research, the decision to interview a purchasing agent about a given product constitutes a judgment sample. He/she must be regarded as a representative of the company by the person who draws the sample. Other examples could include an instructors choice of someone to start a class discussion: expert witnesses presenting their views in court, and the selection of stores in an area to try out a new display.

Again, the degree and direction of error are unknown and definitive statements are not meaningful. However, if the expert judgment is valid, the sample will be a better one than is a convenience sample is used.

Judgment sampling is moderately used in practice.