SAMPLING

 Sampling is a technique of data collection. It is a process of drawing a sample from a ‘population’ or ‘universe’. A sample is a portion that is selected from the ‘universe’ or ‘population’. ‘Universe’ or ‘Population’ in this case refers to all the items in any field of inquiry.

In statistics, universe or population refers to all the individuals, events, things which a particular study wants to cover.

When a statistical investigation is carried out, it is known as a sample survey (enquiry). If the population is infinite, its complete study is impossible. Even if the population is finite, it is often possible to determine the characteristics of the population only on the basis of a sample. 

Objectives of sampling

a. Selecting a sample of adequate size.

b. Collecting the information

c. Making inferences (conclusions) about the population

Advantages of sampling

a. A sample survey is time saving and less expensive compared to a census survey.

b. A sample survey requires small administration organization because the field of survey is small, and also the staff needed is small as the amount of information to be collected and processed is small.

c. The results obtained from sampling method are accurate and reliable

d. Since the coverage is limited, detailed information can be obtained.

e. If the population is very large, sampling is generally the only method of study applicable.

Disadvantages of sampling

a. This method is not suitable where a high degree of accuracy is required

b. In the absence of expert investigators, the results obtained through this method cannot be relied upon.

c. This method is not suitable when there is lot of heterogeneity in the population

d. If due care is not taken in selection of sampling, the conclusions derived will be misleading.

Sampling method

There are two broad categories in sampling:

A. Probability Sampling

B. Non-Probability Sampling

a. PROBABILITY SAMPLING

This type of sampling is also known as ‘random sampling’ or ‘chance sampling’. Under this sampling design, every item of the universe has an equal chance of being selected or being included in the sample. Under probability sampling we have the following types of sampling:

1. Random Sampling

This is the best known form of probability or chance of being selected regardless of the similarities or differences among them, as long as they are members of the same universe.

All that is required to conduct a random sample is to select persons without showing any bias for any kind of personal characteristics.

Accuracy of random sample depends on the accuracy of the sample frame. If some people are listed more than once, they have a greater probability of being selected. If other persons are omitted from the list, they will not be selected at all. In either of theses cases, the sampling will not be random.

Random sampling is used to obtain a sample that is most likely to be a representative of the population. In statistics, random has a technical meaning. It does not mean haphazard or unplanned. When items are selected by chance and not by choice, it is termed as a random sample. In other words, selection depends on chance, and for this reason it is known as chance selection.

The following methods are used for obtaining a random sample:

i. Lottery Method

ii. Table of random numbers

iii. Arrangement of all numbers in the same order, and every 5^th, 10^th, 100^th, etc. unit is selected.

One important property of random sampling is that the larger size of the sample, the more likely it will be closer to population.

Merits of random sampling

1. Since the selection of units in the sample depends entirely on chance, there is no possibility of personal bias affecting the results.

2. Compared to judgment sampling, random sampling represents the universe in a better way.

3. The margin of excess can be calculated because sampling error follow the principle of chance.

Thus random sampling has the advantage of cancelling out biases.

Limitations of random sampling

1. It requires a complete cataloged universe from which a sample is drawn. However, it is difficult to get such a list

2. At times the field of survey is wide, and therefore, it is a restriction geographically for random sampling

3. The size of the sample required to ensure satisfactory and reliable results is usually larger for random sampling rather than under stratified sampling.

2. Stratified sampling

In this method, the population is divided into different groups or classes called strata and a sample is drawn from each stratum at random.

The purpose of stratification is to increase the efficiency of sampling by dividing a heterogeneous population in such a way that there is a great homogeneity within each strata and a marked difference between different strata.

A stratified sample is controlled so that it reflects exactly some known characteristic of the population. In stratified sampling everything is not left to chance. For eg in a public opinion poll the sample selected should reflect all different strata in the population – Hindus, Christians, Parsis, etc.

Stratified samples are rank ordered, such as professors are categorized into associate professors, full time professors, part time professors, assistant professors, etc. time and money are saved in stratified sampling.

Stratified sampling is not limited to only one variable; one can stratify two or more variables simultaneously. For e.g. one can also look into age group, gender, income group etc.

3. Systematic sampling

Systematic sampling is formed by selecting one unit at random and then selection additional units at evenly spaced intervals, until the sample has been formed. This method is popularly used in those cases where a complete list of population is available. The list may be prepared in alphabetical, geographical, numerical or some other order. The most practical way is to select every 15^th name on the list or every 10^th house on the side of the street, etc. an element of randomness is usually introduced into this kind of sampling by using random numbers to pick up the unit with which to start.

Systematic sampling design is simple and convenient to adopt. The time and labour involved is relatively less and the results obtained are often satisfactory. If the population is sufficiently larger, systematic sampling can often yield satisfactorily results.

The main limitation of this method is that it becomes less representative if we are dealing with a population having hidden properties. Another limitation is that if the population is ordered, at times the characteristic investigator is interested in, may not be included.

4. Multi stage sampling or cluster sampling

Cluster sampling is also known as area sampling. It is mainly concerned with a particular geographical area or a particular aspect of population. Under this method a random selection is made of primary, intermediate and final unit from a given population. This method is carried out at different stages. At the first stages, units are selected by some suitable sampling method, then a sample is selected fro the first stage unit by some other sampling technique. Like this, there are stages which are added. E.g. if we want to take students from Delhi University, we may select colleges at the first stage, then select departments at the second stage and choose students at the third and final stage.

Merits

1. This method is used when the area of inquiry is wide

2. It saves time and money

Demerits

1. This method is complicated

2. It is less accurate.

B NON-PROBABILITY SAMPLING

Non probability sampling is that sampling procedure which does not afford any basis for estimating the probability that each item in the population has of being included in the sample. Here the selection of the sample is based on the choice of the investigator. In such sampling, personal element has a great chance of entering into the selection of the sample. The investigator may select a sample which shall yield favorable results to his point of view, and this can be a great disadvantage of this method. The advantage of non probability sampling is that it is less complicated, less expensive and can be done on spur of the moment basis. The various types of non probability sampling are as follows:

1. Convenience sampling or Accidental Sampling

In this type of sampling, the investigator chooses those people who are readily and easily accessible. For e.g. in a study on teenagers and their shopping preferences, the investigator could take a sample of teenagers staying in the vicinity. This method is not very accurate, but it is time and money saving.

Self Selection Sampling is used if you want people to participate on their own accord. 

2. Quota sampling

This is a type of judgment sampling, where the investigator uses his / her own judgment for selecting the sample. In this, quotas are set up according to some specified characteristic like occupation, education, age etc. Each interviewer is then asked to interview a certain number of persons. Here the selection of sample depends upon personal judgment for e.g. in a survey on people who listen to the radio, the interviewer may be told to interview 500 people living in a certain area, and out of every 100 person of the interview, 60 are supposed to be people who are service workers, 25 housewives, 15 youngsters etc.

The cost involved in quota sampling is relatively less, but there are many chances of personal biases which may depend on integrity, honesty and competence of the investigator.

3. Purposive or Judgment Sampling

In this method the investigator has complete freedom to choose the sample according to his wishes. The investigator chooses those units which he thinks are more representative of the universe. For e.g., if a sample of 20 students is to be selected from a class of 80 then the investigator will select those 20 students who in his opinion are most representative of that class.

Judgment sampling is used only when a small sample is to be selected, or when we want to study some unknown traits of the population or need to solve some sort of everyday problem or to make policy decisions. Judgment sampling can be used to arrive at solutions and \make decisions. Judgment sampling can be used to arrive at solutions and make decisions.

A disadvantage of this method is that even though it is simple, it is not scientific. Personal biases and prejudices of the investigator may come in the way of the study.

The success of this method depends upon how excellent is the judgment of the investigator. A good judgment will result in good representation of the population.

4. Snowball sampling

In recent years snowball sampling has become very important. This method of sampling is conducted in stages. In the 1^st stage a few persons having the required characteristics are identified and interviewed. These persons are then used as informants to identify other such people who qualify to be included in the sample.

 At the 2^nd stage, these persons who were suggested by the 1^st group are interviewed and who will in turn lead to the next some more people who will be interviewed for the 3^rd stage and so on.

For e.g. In a study on married women who work from home, in the first stage some such women will be selected, and later they will suggest other such married women who work from home to be interviewed for the next stage. This continues till the desired sample size gets collected by the researcher.

The term ‘snowball’ comes from the analogy of a snowball which begins small, but becomes bigger and bigger as it rolls downward.