subject based on information obtained from a carefully chosen representative set of people. Such people are usually referred to as ‘respondents’ — they respond to questions asked of them by the researchers. Survey research is usually done by large teams consisting of those who plan and design the study (the researchers) and their associates and assistants (the latter are called ‘investigators’ or ‘research assistants’). The survey questions may be asked and answered in various forms.
Often, they are asked orally during personal visits by the investigator, and sometimes through telephone conversations. Responses may also be sought in writing, to ‘questionnaires’ brought by investigators or sent through the post. Finally, with the increasing presence of computers and telecommunication technology, these days it is also possible for surveys to be conducted electronically. In this format, the respondent receives and responds to questions by email, the Internet, or similar electronic medium.
The survey’s main advantage as a social scientific method is that it allows us to generalise results for a large population while actually studying only a small portion of this population. Thus a survey makes it possible to study large populations with a manageable investment of time, effort and money. That is why it is such a popular method in the social sciences and other fields. The sample survey is able to provide a generalisable result despite being selective by taking advantage of the discoveries of a branch of statistics called sampling theory.
The key element enabling this ‘shortcut’ is the representativeness of the sample. How do we go about selecting a representative sample from a given population? Broadly speaking, the sample selection process depends on two main principles. The first principle is that all the relevant sub-groups in the population should be recognised and represented in the sample.
Most large populations are not homogenous — they belong to distinct sub-categories. This is called stratification (Note that this is a statistical notion of stratification which is different from the sociological concept of stratification that you have studied in Chapter ). For example, when considering the population of India, we must take