Sampling – UGC NET – Notes

Introduction

• Geography, dealing with the man and environment relationship, is essentially a social science. One of the main problems that geographers meet in their pursuit of research is the abundance of data.
• As a matter of fact, in the recent decades, an ‘explosion of data’ has taken place in every sphere of life, providing an enormous source of valuable information in the form of numerical facts for the quantification of socioeconomic problems in space and time.
• The increased quantity of data, though useful for the formulation of hypotheses and their testing, has created the problems of data processing, its plotting on maps and analysis. The task of researchers has, thereby, been rendered arduous, costly and time consuming.
• Almost all the branches of geography, e.g., geomorphology, climatology, oceanography, pedology, demography, economic, agricultural and industrialn geography, urban and rural land use planning, transport, urban settlement, electoral and medical geography have all turned to more precise numerical data in their attempts to render a more realistic and objective assessment of geographical phenomena.
• Moreover, now geographers are increasingly cooperating with scientists of other disciplines. Application of sound and sophisticated statistical techniques to geographical data has, therefore, become essential.
• ‘Sampling’ is a useful technique for the processing of data. It is frequently used by geographers in their studies. Sampling of data in itself is a tedious job which requires utmost care on the part of researcher to arrive at reliable results.
• The essence of sampling lies in the fact that a large number of items, individuals, or locations (statistical population) may, within specified limits of statistical probability, be presented by a smaller group of items (a sample) selected from the larger group (a parent population).
• Out of the huge population, if a limited selection of items or cases is made, it is called a ‘sample’. The limited ‘sample’ is generally sufficient for making a generalization about the whole population. In many cases, the numbers of individuals in the population, e.g., the average yield of all the plots of an agricultural region or the pebbles on a sea-beach are so numerous that measuring of all of them would be almost impossible from a practical point of view.
• But, if through sampling, a limited selection of fields for the measurement of yields would enable the observer to obtain the average yield of the fields in the entire region, similarly, a limited selection of pebbles on the sea-beach will be sufficient for making a generalization about the pebbles on that coast.
• Sampling, thus, represents a more efficient use of our energy, still allowing us to make reliable statements about the whole population. Public opinion polls announce how a nation intends to vote, or analyze people’s attitudes on current issues, but their conclusions are obtained from a sample consisting of a few hundred questionnaires, rather than by consulting everyone in the country. Complete enumeration of population in most of the cases is almost impracticable.
• An appropriate sampling in geographical research is highly desirable as it saves time, efforts and cost appreciably and gives reliable results which can be used for the purpose of generalization and forecasting. The problem of choosing the right size of sample, however, is a little more complicated.
• The simplest rule is that the larger the size of sample, the more likely it is to give a reliable picture of the parent population. As a further rough guide, it can be said that the size of sample should be at least 5 per cent to 15 per cent of the total for satisfactory results. The decisions on defining parent population and choosing the best sampling method, however, depend to a large extent on commonsense.
• Some of the commonly known and frequently used methods of sampling are: random sampling, purposive sampling, systematic sampling, stratified sampling and multistage sampling.