# Essentials of Sampling in Introduction to Sociological Research

Sampling is an essential element of all social investigation. It is a means by which a selection is made from the basic unit of study and is traditionally associated with survey research. “Attributes” and “Variables” are the technical expressions used in Sampling. Attribute refers to the characteristic that considers whether an individual possess or donot possess certain characteristics. A variable can always be transformed into an attribute by a broad grouping and an attribute can be changed into a variable by allocating the score “1” to all who possess the attribute and “0” to those who donot. “Statistic” and “Parameter” are the other terms that need explanation. Statistic refers to a summary value of a variable calculated from a sample whereas parameter refers to a summary value of the variable in the population that one is trying to estimate. Sample surveys have two objectives: – the main purpose is to estimate certain population parameters and then to test a statistical hypothesis about a population. The estimation and testing hypotheses are accomplished by the standard error. Statements based on sample results are always probability statements.

Sampling in field research involves the selection of a research site, time, people and events .the kind of data gathered is influenced by these factors. According to Spradley (1980) while selecting a research site, a series of factors needs to be considered among which is research mobility. The research site should be such that it provide situations and subsides for investigation, have a degree of access and allow the researchers to participate in the activities that they plan to study. Anthropologists argue that sources of data for research get selected when certain individuals volunteer their help, some groups extend welcome, and some techniques happen to be accessible to observation.

Researchers have to consider the time dimension in all field situations as activities may vary within a social organisation with time. A structured approach to time sampling has been used to study children in hospitals by Stacey (1969). Another dimension to time sampling occurs in organisations such as hospitals and factories, where shift systems operate. By using time sampling, the field researcher can gather detailed systematic data in a social setting that can be compared with observational material and with data gathered by other methods

After deciding where and when to do the research, decisions have to be made about the people who are to be researched. Researchers use snowball sampling if they find difficulty concerning which they should study. Here they use informants to introduce them to other members of their group. Another major facet of field research is the way in which data is collected from particular informants in some depth. According to Spradley (1979) an individual can be selected as a key informant by keeping in view his involvement in the social setting, representation of a cultural scene, non-analytic abilities and the details that they can provide.

It is important for the field researcher to distinguish between three discrete sets of events: the routine, the special and the untoward. Routine events are situations that occur regularly, special events are defined as situations that are fortuitous, but nevertheless anticipated, while untoward events are defined as emergency situations.

The proper application of statistical techniques is dependent upon the extent to which the researcher adheres to or satisfies certain rules or assumptions. All statistical tests and techniques have several general assumptions that pertain to

(1). Levels of Measurement

(2). Sampling Procedures

(3). Sample Size

Statistical Procedures involve manipulation of numbers .coding is a process whereby data of any kind are transformed into numerical quantities. Numbers mean different things depending on what they stand for which affects the statistical procedures accordingly. The levels of measurement include (1) Nominal (2) Ordinal (3) interval and (4) Ratio.

Numbers that are applied to data of the Nominal level stand for categorical or classified information. It includes variables such as Political Affiliation, Sex, Race, Religious preference. The Ordinal level of measurement implies that numbers assigned to subclasses of any variable may be rank-ordered or gradated according to some low to high arrangement. The interval level contains all the properties and characteristics of the Nominal and Ordinal levels of measurement and provides for equal spacing between numbers. Ratio level is seldom referred to or used in statistical work. It includes all properties and Characteristics of the nominal. Ordinal, interval levels and has an absolute zero associated with it. The level of measurement assumption is important because it is directly connected to the arithmetic operations involved in statistical tests and Procedures.

A second assumption underlying the appropriate application of all statistical tests is that the researcher needs to obtain a random sample of elements. A random sample is defines as a sample drawn in such a way so that each element has an equal and an independent chance of being included. Here, a distinction is to be made between probability sampling and non-probability sampling. Probability sampling designates a method that specifically intends every unit in the universe under study to have the same known probability of being studied. It follows definite rules and samples are drawn at random; hence the familiar name for such sampling is Random Sampling. Non-Probability Sampling plans are those that donot use randomness as the primary control for the inclusion of elements.

There are four Sampling plans that fall under Probability Sampling

(1) Simple Random Sampling

(2) Proportionate Stratified Random Sampling

(3) Disproportionate Stratified Random Sampling

(4) Cluster or Area Sampling

In Simple Random Sampling, all elements or nearby all elements are identified by the researcher and a desired number of them is selected randomly for subsequent study.

If a researcher wants to select persons with particular attributes such as “year in school”, “race”, or “Sex”, a stratified random Sample will be required .Proportionate Stratified random Samples are those in which certain characteristics of interest to the researcher are represented in the same proportion that they exist in the Population. Disproportionate stratified random samples are randomly selected in such a way that certain characteristics are not distributed in the same proportion as they are distributed in the population. The key control governing the inclusion of elements for both kinds of Stratified random Samples is the procedure of Randomness. Ecologists and Demographers who study larger Populations, such as the Population of India, Canada use cluster Samples. This Sampling is of Practical Advantage as it saves considerable time and travel expense.

Non-Probability Sampling includes (1) Accidental Sampling (2) Quota Sampling and (3) Judgemental or Purposive Sampling

Accidental Sampling occurs whenever the Researcher takes a sample from whatever is available at the time. Quota samples are designed to ensure the inclusion of certain types of persons in a subsequent draw of elements. A judgemental sample is one that is handpicked by someone who knows the population fairly well and can be relied upon to designate persons who represents a reasonable sampling of viewpoints on a given issue. Probability and non-Probability samples manifest similar characteristics and results when the universe is more homogenous. However Probability sampling is advantageous as it can conveniently and confidently answer questions concerning the frequency with which features are distributed in a large Population. On the other hand it turns out to be inappropriate if the aim of research is to understand a social or cultural system to whose operation or dynamics individual actors or artefacts offer only clues. It is especially appropriate to problem oriented research, where it helps in defining the crucial variable s which, in turn are often few enough to allow an adequate sampling frame to be efficiently constructed.

The third assumption is about the sample size as the sample-size requirements vary depending upon a person’s area of Specialisation. For eg: a person studying small groups and small group interactions will require samples of fewer than 25 elements whereas a demographer might require a sample of 300 to 400 or even more for demographic research purpose.

Thus, we see for a successful research objective proper sampling involving different methods are required. However, it is not always possible to use statistical sampling procedures. While adopting sampling strategies, the researchers should keep the focus on the sociological characteristics of the groups and individuals that are studied

They need to understand the principles involved in sampling strategies and the way in which they can be combined.

**Author::**

Mrinal Saikia

New Delhi

B.Tech (electric and communication) &

M.Tech( VLSI design and embeded system)

Karnataka University

**Bibliography:**

- Burgess Robert G (chapter 1) elements of Sampling in Field research, (Chapter-12) sampling in Ethnographic Fieldwork.
- Moser, C.A and G. Kalton ,1971 survey methods in social investigations , London ,Heinemann Educational Books ,Chapter -4
- Champion, D.J 1981 Basic Statistics for Social research. New York : Macmillan Chapter-2.