Research Methods and Analysis PDF

4 RESEARCH METHODS AND ANALYSIS Field A systematic and scientific research is essential for sociology to claim itself as a scientific discipline. However, whatsoever be the theoretical perspective or methodological orientation of the researcher, ultimately, he has to carry out his research in the f...

4 RESEARCH METHODS AND ANALYSIS Field A systematic and scientific research is essential for sociology to claim itself as a scientific discipline. However, whatsoever be the theoretical perspective or methodological orientation of the researcher, ultimately, he has to carry out his research in the field. So, let’s try to understand what ‘field’ implies in a socio- logical enquiry and what are the problems or challenges associated with the fieldwork. In the context of sociological research, the term ‘field’ refers to the members of a social group which is the prime object of study for a social scientist. In its early phase, Malinowski and Radcliffe-Brown laid the foundations of intensive fieldwork among anthropologists in Britain. However, in the Indian context, it was M.N. Srinivas who strongly advocated for the ‘field-view’ of Indian society in place of the ‘book-view’. The book-view of Indian society was largely cham- pioned by Indologists like B.K. Sarkar, G.S. Ghurye, Radhakamal Mukerjee and Irawati Karve. Indologists have claimed that the Indian society could be understood only through the concepts, theories and frameworks of Indian civi- lisation. They believed that an examination of the classical texts, manuscripts, archaeological artefacts, etc. should be the starting point for the study of the present (Beteille, 2009). Srinivas was critical of the ‘book-view’ of Indian society. He argued that the book-view gave a distorted picture of society by dwelling on the ideals of the past from which the present reality departed considerably. The book-view of Indian society presented an idealised picture of its institutions – marriage, family, kin- ship, caste and religion – dwelling more on what they were supposed to be than on how they actually worked. For example, the book-view had represented caste in terms of the invariant and immutable scheme of the four varnas. Field studies DOI: 10.4324/9781003291053-5 142 Research Methods and Analysis shifted attention away from the fourfold scheme of the varnas to the operative units of the system, namely the innumerable jatis. They also drew attention to the ambiguities of caste ranking and the very distinctive process of caste mobil- ity. Thus, the field-view revealed the gap everywhere between the ideal and the actual. By bringing to attention ambiguities, contradictions and conf licts, it paved the way for a better understanding of the dynamics of social change. Thus, the idea of an unchanging and immutable society began to give way, and the field-view changed not only the perception of India’s present, but to some extent also the perception of its past (Beteille, 2009). However, like every other method, fieldwork too is marked by its own set of challenges and problems in conducting genuine sociological research. First and foremost, the researcher faces the problem of the choice of the ‘field’ to carry out his fieldwork as no typical field exists in reality. As stated earlier, unlike natural sciences, sociology cannot study any particular social phenomena in a laboratory by the experimental method due to certain moral and ethical reasons. As a result, social research takes place in the open, where, unlike a scientific experiment, it is extremely difficult to control the extraneous variables. Hence, it becomes increasingly challenging for the social scientist to establish a cause– effect relationship between the variables stated under the hypothesis. After hav- ing identified the field for his research, the researcher faces the challenge of entry into the field. This implies that unless the researcher is able to establish a good rapport with the natives, he would find it hard to carry out his research. Thus, in order to seek the cooperation of the native population for his data collection, the researcher must gain their acceptability. In this, the social background of the researcher also plays an important role. Further, since the researcher can carry out only a limited study of any given social phenomenon, the problem of holism looms large. Since holistic study appears to be impractical in the study of complex societies, the researcher should keep in mind that the segment he is studying is a part of the larger and complex whole and should look for interrelationships. A researcher may also face problems in the formulation of hypothesis and might have to reformulate or modify his hypothesis because hypotheses cannot be formulated in vacuum, without the knowledge of the field. Further, the issues of objectivity and ethical neutrality also need to be addressed. The researcher should be aware of his biases and preju- dices and try to make certain that they do not inf luence his collection and inter- pretation of data. Though some of these challenges are endemic to any social science enquiry, yet they can be dealt with a cautious and informed approach on the part of the researcher. Since the fieldwork basically involves dealing with people, the researcher must be empathetic and f lexible in his approach and employ the services of well-trained field workers. In the ultimate analysis, it may be argued that in any field research, the soci- ologist is an integral part of the research process. The data so collected have no existence independent of the researcher. His data are ‘constructions, not ref lec- tions of facts or relationships alone. In the process of knowing, external facts  Research Methods and Analysis 143 are sensorily perceived and transformed into conceptual knowledge’. Thus, the sociologist as a researcher is an active factor in the creation of knowledge and is not just a mere passive recipient. The importance of his perception makes a sociologist as integral part of the research process as the data he observes (Srinivas et al., 2002). Refexivity and the Changing Notion of Field and Fieldwork Practices You now know that ref lexivity offers an alternate perspective to study social reality. Ref lexivity in social anthropological research implies that the ideas about ‘field’ and ‘practices’ of fieldwork are constantly examined and reformed in the light of new developments, thus continuously altering their character. Ref lexivity challenges the conventional notions of anthropological research with regard to field and practices of fieldwork. Early social anthropological research was largely concerned with the study of small-scale societies in their natural state or surroundings. Hence, the term ‘field’ came to denote a distinct social group which was to be studied in its unique sociocultural and geographical setting. Early anthropological research was largely based on the dichotomy of subject and object. In other words, it was based on the separation of the social scientist (subject) from those ‘others’ whom he observes (object). It was based on the assumption that over-involvement of the social scientist with his object of study (social group) may contaminate the research findings. The idea about ‘otherness’ remained remarkably central to the fieldwork practices of Malinowski, A.R. Radcliffe-Brown, etc. Amory shows how these ideas about ‘otherness’ and taking for granted of a white subject have shaped the field of African studies in the United States. She shows that African American scholars were discouraged from working in Africa, on the grounds that they were ‘too close’ and would not manage to be ‘objective’, while white scholars were judged to have the appropriate distance from the black ‘other’. This helps to explain why the contemporary field of African studies con- tains remarkably few black American scholars. Kath Weston too in her study of gay and lesbian communities in United States arrived at a similar conclusion. She argued that her position as a native ethnographer itself blurs the subject/object distinction on which ethnography is conventionally founded. She calls native ethnographer a ‘virtual anthropologist’. Akhil Gupta and James Ferguson also question the conventional notion of field and argue that in the light of new developments, there is a need for reconstruction of field and fieldwork practices. They argue that processes such as decolonisation and globalisation, accompanied by processes of diffusion and acculturation, have challenged the traditional definition of field and the very idea of a clearly demarcated space of ‘otherness’. They argue that the conventional notion about the ‘field’ in terms of a homogenous social group with its unique culture and geographical surroundings has come to be questioned in the wake 144 Research Methods and Analysis of globalisation. Social groups are no longer tightly territorialised or spatially bounded. Further, the processes of diffusion and acculturation have significantly altered the homogenous character of social groups, and today cultural heteroge- neity is more common. Gupta and Ferguson further question the ‘field/home dichotomy’ in social anthropological research. They question the traditional notion of field which rested on the idea that different cultures exist in discrete and separate places. They argued that the ‘location’ of the field should not merely be seen in geo- graphical sense alone. They advocated retheorising of fieldwork from spatial sites to social and political locations in terms of unequal power relations. For example, subaltern approach in sociology has significantly contributed towards a better understanding of various socio-economic and political processes in India, which were until now largely studied from an elitist perspective. Gupta and Ferguson argue that with decolonisation, there is proliferation of domestic research led by the natives. As a result, today, the very idea of ‘otherness’, which was central to the early anthropological fieldwork, is subjected to review. Hence, there is a need to modify the practices of fieldwork accordingly. Further, Gupta and Ferguson also question the fundamental premise of early anthropological fieldwork practices that only professionally trained observers could be trusted to collect ethnographic data. Paul Radin in his study found that his untrained native research assistants proved to be better than the academically and professionally qualified observers in terms of gathering valuable data. This is because, as Radin argues, such professionals are socially separated from those whom they study by their very training. The training of the professional observ- ers erects an undesirable barrier between themselves and the persons to be inter- rogated. It may lead to difficulty in establishing direct and immediate contact and building rapport with their sources of information. On the other hand, the native research assistants or local intellectuals are better positioned at least for certain sorts of data collection (Gupta and Fergusson, 1997). Thus, ref lexivity has significantly contributed in reconstruction of the ideas about field and fieldwork practices in social anthropology. Such a rethinking of the idea of the ‘field’ coupled with an explicit attentiveness to ‘location’ might open the way for a different kind of anthropological knowledge and a different kind of anthropological subject. Let us now discuss some of the fundamental characteristics of quantitative as well as qualitative research methods. Quantitative and Qualitative Research Methods Methodology is an integral aspect of any scientific discipline. Every discipline requires a methodology to reach its conclusions. In other words, it must have ways of producing and analysing data so that theories can be tested, accepted or rejected. In the absence of a systematic way of producing knowledge, the findings of a subject can be dismissed as guesswork or mere common sense assumptions.  Research Methods and Analysis 145 You know that sociology first developed in Europe in the nineteenth century when industrialisation resulted in massive social changes. Accompanying these social changes were intellectual changes during which science started to enjoy a higher reputation than ever before. Many of the founders of sociology believed it would be possible to create a science of society based upon the same principles and procedures as that of the natural sciences. However, not all sociologists have agreed that it is appropriate to adopt the methodology of the natural sciences. These sociologists believed that the subject matter of the social and natural sciences is fundamentally different. While natu- ral sciences deal with matter, social sciences deal with man and his behaviour. Since human beings are conscious beings, their action is meaningful. Hence, for these sociologists, methods of natural sciences alone would be inadequate. They instead emphasised on the interpretive methodology to understand the subjective meanings that underlie social action. Thus, on the basis of the above discussion, two broad traditions within sociol- ogy could be identified. 1. Those who advocated the use of scientific and quantitative methods (positivists) 2. Those who supported the use of more humanistic and qualitative methods (anti-positivists) However, in recent years, the new-generation sociologists have started ques- tioning such rigid divisions between quantitative and qualitative methodologies. Most sociologists instead advocate a combination of both quantitative and quali- tative methods in sociological research. Quantitative Research Methods Quantitative research in sociology is largely associated with the ‘positivist tra- dition’. Early sociologists belonging to the positivist tradition, such as Comte, Spencer, Durkheim, etc., believed that the methods and procedures of natural sciences could be adopted in sociology as well. Quantitative research is associated with a number of techniques of data collection such as survey, questionnaire, structured interview and secondary sources of data, etc. Some of the features of quantitative research in sociology are discussed next. 1. Social Facts: Early sociologists like Comte and Durkheim, who were positivists, placed particular emphasis on behaviour that can be directly observed. It argues that those aspects of behaviour which are not directly observable, such as meanings, feelings and purposes, are not particularly important and can be misleading. Durkheim went to the extent of stat- ing that social facts should be treated as ‘things’. This means that the belief systems, customs and institutions of society – the facts of the social world – 146 Research Methods and Analysis should be considered as things in the same way as the objects and events of the natural world. 2. Statistical Data: The second aspect of quantitative approach as advocated by positivists is the use of statistical data. Positivists believed that the behav- iour of man, just like the behaviour of matter, can be objectively measured and, thus, it was possible to classify the social world in an objective way. For example, Durkheim collected data on social facts such as the suicide rate and the membership of different religions. 3. Correlation: The third aspect of positivist methodology entails looking for correlations between different social facts. A correlation is a tendency for two or more things to be found together, and it may refer to the strength of the relationship between them. For example, in his study of suicide, Durkheim found an apparent correlation between a particular religion, Protestantism, and a high suicide rate. 4. Causation: The fourth aspect of positivist methodology involves a search for causal connections. Positivists believed that on the basis of the data col- lected through quantitative methods, patterns could be identified and a cause-and-effect relation between two or more variables or social phenom- ena could be established. For example, Durkheim in his study of suicide had explained that low solidarity among the Protestants was the causal factor for high suicide rate among them. 5. Generalisation and Replicability: The quantitative researcher is invari- ably concerned to establish that his result of a particular investigation can be generalised to the larger population. Positivists like Comte and Durkheim believed that just as natural sciences could arrive at universal laws with regard to matter, laws of human behaviour can also be discovered in social sciences. They believed that laws of human behaviour can be discovered by the collection of objective facts about the social world, by the careful analysis of these facts and by repeated checking of the findings in a series of contexts (replication). Positivism is based upon an understanding of science that sees science as using a mainly inductive methodology. An inductive methodology starts by collect- ing the data. The data are then analysed, and out of this analysis theories are developed. Once the theory has been developed, it can then be tested against other sets of data to see if it is confirmed or not. If it is repeatedly confirmed (replicated), then positivists like Comte, Durkheim, etc. assume that they have discovered a law of human behaviour (Haralambos and Holborn, 2014). Qualitative Research Methods As mentioned earlier, qualitative research methods in sociology are largely advo- cated by sociologists belonging to social action approach who emphasise on the interpretive understanding of human social behaviour. These sociologists prefer  Research Methods and Analysis 147 sacrificing a certain precision of measurement and objectivity in order to get closer to their subjects, to examine the social world through the perspective of the people they are investigating. They sometimes refer to quantitative researchers as those who ‘measure everything and understand nothing’. Qualitative research fundamentally refers to that approach to the study of the social world which seeks to describe and analyse the culture and behaviour of humans and their groups from the point of view of those being studied. As discussed earlier, quantitative data are data in a numerical form, for example, official statistics on crime, suicide and divorce rates. By comparison, qualitative data are usually presented in words, e.g. an ethnographic account of a group of people living in poverty, providing a full and in-depth account of their way of life, or a transcript of an interview in which people describe and explain their attitude towards and experience of reli- gion. Compared to quantitative data, qualitative data are usually seen as richer, more vital, as having greater depth and as more likely to present a true picture of a way of life, of people’s experiences, attitudes and beliefs. Participant observa- tion, unstructured interview, focus group discussion, life history or case study method are some of the major methods or techniques of data collection in quali- tative research. The main intellectual undercurrents which tend to be viewed as providing qualitative research with their distinct methodology are Weberianism, symbolic interactionism, phenomenology, ethnomethodology, etc. Some of the features of qualitative research in sociology are discussed next. 1. Empathetic Description of Social Reality: Since social action theorists believe that social action is meaningful, they focus on the interpretation of the meanings that social actors have probably given to their act. The most fundamental characteristic of qualitative research is its express com- mitment to viewing events, actions, norms, values, etc. from the perspective of the people who are being studied. This is explicitly stated by Weber in his Verstehen methodology. 2. Contextualism: Social action theorists believe that social action does not exist in isolation, rather it is expressed in a given socio-economic, politi- cal and historical context. Hence, qualitative research exhibits a preference for contextualism in its commitment to understanding events, behaviour, etc. in their respective context. It is almost inseparable from another theme in qualitative research, namely ‘holism’, which entails an undertaking to examine social entities – schools, tribes, firms, slums, delinquent groups, communities or whatever – as wholes to be explicated and understood in their entirety. 3. Emphasis on Processual Dimension: Qualitative research views social life in processual and dynamic terms, rather than static terms. The emphasis on process can be seen as a response to the qualitative researcher’s concern to ref lect the reality of everyday life which, they tend to argue, takes the form of streams of interconnecting events. The general image that quali- tative research conveys about the social order is one of interconnection 148 Research Methods and Analysis and change. For example, symbolic interactionists explain social order as a ‘negotiated order’. 4. Flexibility: Qualitative researchers tend to favour a research strategy which is relatively open and unstructured. Such strategy allows them access to unexpectedly important topics which may not have been visible to them had they foreclosed the domain of study by a structured, and hence, potentially rigid strategy. However, some sociologists, in recent years, have questioned the need for such a rigid division between quantitative and qualitative methodology and have advo- cated combining the two approaches. Alan Bryman has suggested a number of ways in which a plurality of methods – a practice known as triangulation – can be useful. 1. Quantitative and qualitative data can be used to check on the accuracy of the conclusions reached on the basis of each. 2. Qualitative research can be used to produce hypotheses which can then be checked using quantitative methods. 3. The two approaches can be used together so that a more complete picture of the social group being studied is produced. 4. Qualitative research may be used to illuminate why certain variables are statistically correlated. For example, Durkheim concluded in his study on suicide that the rate of suicide varies from religion to religion because of their varying degree of solidarity. Bryman believes that both quantitative and qualitative research have their own advantages. Neither can produce totally valid and completely reliable data, but both can provide useful insights into social life. He argues that each has its own place and they can be most usefully combined. Generally, quantitative data tend to produce rather static pictures, but they can allow researchers to examine and discover overall patterns and structures in society as a whole. Qualitative data are less useful for discovering overall patterns and structures, but they allow a richer and deeper understanding of the process of change in social life (ibid.). Many of the debates about the merits of particular research methods focus on questions of ‘reliability’ and ‘validity’. In the natural sciences, data are seen to be ‘reliable’ if other researchers using the same methods of investigation on the same material produce the same results. By replicating an experiment, it is pos- sible to check for errors in observation and measurement. Once reliable data have been obtained, generalisations can then be made about the behaviour observed. No sociologist would claim that the social sciences can attain the standards of reliability employed in the natural sciences. Many would argue, however, that sociological data can attain a certain standard of reliability. Generally speaking, quantitative methods are seen to provide greater reliability. They usually pro- duce standardised data in a statistical form: the research can be repeated and the  Research Methods and Analysis 149 results checked. Qualitative methods are often criticised for failing to meet the same standards of reliability. Such methods may be seen as unreliable because the procedures used to collect data can be unsystematic, the results are rarely quantified and there is no way of replicating a qualitative study and checking the reliability of its findings. Further, data are considered ‘valid’ if they provide a true picture of what is being studied. A valid statement gives a true measurement or description of what it claims to measure or describe. It is an accurate ref lection of social reality. Data can be reliable without being valid. Studies can be replicated and they produce the same results, but those results may not be a valid measure of what the researcher intends to measure. For instance, statistics on church attendance may be reliable but they do not necessarily give a true picture of religious commitment. Supporters of qualitative methods often argue that quantitative methods lack validity. Statistical research methods may be easy to replicate, but they may not provide a true picture of social reality. They are seen to lack the depth to describe accurately the meanings and motives that form the basis of social action. They use categories imposed on the social world by sociologists, categories that may have little meaning or relevance to other members of society. To many inter- pretive sociologists, only qualitative methods can overcome these problems and provide a valid picture of social reality. Researchers are sometimes attracted to quantitative methods because of their practicality. Quantitative methods are generally less time-consuming and require less personal commitment. It is usually possible to study larger and more repre- sentative samples which can provide an overall picture of society. Qualitative research often has to be confined to the study of small numbers because of prac- tical limitations. It is more suited to providing an in-depth insight into a smaller sample of people. Hence, most sociologists today tend to combine both quantita- tive and qualitative methods in their social research to enhance its reliability as well validity. Concepts and Hypothesis In our discussion on ‘Scientific Method and Sociological Research’ in Chapter 3, we have learned that a social scientist starts his research by defining precisely what it is that he wants to know. This he does by formulating a clear and verifi- able hypothesis. Let us now discuss hypothesis and its relevance in sociological enquiry in detail. According to Theodorson and Theodorson, a hypothesis is a tentative state- ment asserting a relationship between certain facts. Bailey has also said that a hypothesis is a proposition stated in a testable form which predicts a particular relationship between two or more variables. Since statements in hypothesis have to be put to empirical investigation, the definition of hypothesis excludes all such statements which are merely opinions or value judgements, for example, ‘All politicians are corrupt’. In other words, a hypothesis carries clear implications for 150 Research Methods and Analysis testing the stated relationship, that is, it contains variables that are measurable and also specifies how they are related. A statement that lacks variables or that does not explain how the variables are related to each other is no hypothesis in sci- entific sense. Variables are measurable phenomena whose values can change. In social sciences, variables may be understood as the social characteristics that can be converted into measurable forms and analysed. This point will become clearer in our subsequent discussion on operationalisation of concepts (Ahuja, 2008). In social sciences, the social scientists tend to study and explore the various aspects of social reality and interrelationship among them. Since social reality is infinite, social scientists make sense of this infinite social reality through logi- cal abstractions. These logical abstractions or mental constructs are nothing but the ‘concepts’. Hence, when a social scientist is carrying out research to test his hypothesis, he is actually exploring the relationship between the two concepts. Thus, in general, variables in social sciences are nothing but the concepts which are the part of the research. However, in particular, variables are described as the specific characteristics or attributes of the more general concepts. You will soon learn that in order to carry out research, the variables or the concepts used in the hypothesis should be clearly defined and operationalised. This point will be dis- cussed in detail with examples in our subsequent discussion on operationalisation of concepts. Further, the terms ‘independent variable’, ‘dependent variable’ and ‘extraneous variable’ used commonly in research have already been discussed in our earlier discussion on the ‘scientific method’. A few examples of hypothesis are cited below: suicide rates vary inversely with social integration urbanisation leads to proliferation of nuclear families literacy rate is directly related to average marital age children from broken homes more likely tend to be delinquents Hypothesis formulation is of fundamental importance in any research. A hypoth- esis looks forward. It provides direction to research. Without it, research is unfo- cused and may be reduced to a random empirical wandering. Results of such unguided research would be of little use. The hypothesis is the necessary link between theory and the investigation that leads to the discovery of additions to knowledge. Goode and Hatt argue that theory and hypothesis are very closely interrelated. Hypotheses are the deduced propositions from the existing theory. These hypotheses, when tested, by means of empirical investigations, are either proved or disproved. Hence, in turn, hypothesis testing leads to either revalida- tion or reformulation of the theory (Goode and Hatt, 2006). Let us now discuss a few essential characteristics of a good hypothesis. The most fundamental of them all is that a hypothesis must be conceptually clear. This means that the variables or the concepts used in the hypothesis should be clearly defined and operationalised. Operationalisation of concepts refers to  Research Methods and Analysis 151 the process of defining the concepts in terms of those attributes which could be empirically observed during research. In other words, operationalisation refers to the process of converting concepts in their empirical measurements. For example, the concept of anomie, in general terms, is defined as a ‘state of normlessness’ in society. Now, anomie could be observed in social, economic and political systems of a given society. For instance, the concept of anomie in social sphere may be operationalised by identifying the quantifiable attributes such as incidence of suicide, crime, honour killings, etc. on which empirical data can be collected. Similarly, in political sphere, the concept of anomie may be operationalised when it is explained in terms of the attributes like stability of the government, corruption in the government, people’s perception about the government, etc. In the economic sphere, the concept of anomie may be opera- tionalised when it is defined in terms of attributes like economic inequality, poverty, unemployment, etc. Second, the variables or the concepts used in hypothesis should be commonly accepted and communicable. In simpler words, it implies that there should be uniformity in the definition and meaning of the concepts used in hypothesis. Once the concepts have been clearly defined and operationalised by means of their empirical referents, a hypothesis must also specify the relationship between the variables. A hypothesis that does not explain how the concepts are related to each other is no hypothesis in scientific sense. For example, suicide rates vary inversely with social integration. Another important characteristic of a good hypothesis is that it should be related to the available techniques of data collection and interpretation. In other words, a hypothesis must be so formulated keeping in mind the availability as well as applicability of the techniques of data collection in the respective field (socio-geographical area identified to carry out the research). Further, a good hypothesis must be related to a body of theory. As stated earlier, one of the key features of any scientific discipline is its cumulative nature. Likewise, sociologi- cal theories are built upon one another, extending and refining the older ones and producing the new ones. This would be possible only if the hypotheses are related to a body of theory (Ahuja, 2008). Dear Learner, so far we have discussed the meaning of hypothesis, its rel- evance in sociological enquiry and some characteristics of a good hypothesis. We have discussed that a hypothesis is a tentative statement asserting a relation- ship between two or more concepts or variables. What do you understand by concept? What is the relevance of concepts in sociology? What are the problems associated in defining the concepts in sociology? Let us now try to understand and answer these questions. Concepts are the logical abstractions or mental constructs created from sense impressions, percepts or experiences. Concept formation is an essential step in the process of sociological reasoning. Concepts are the tools with which we think, criticise, argue, explain and analyse. We build up our knowledge of the 152 Research Methods and Analysis social world not simply by looking at it but through developing and refining concepts which will help us make sense of it. Concepts, in that sense, are the building blocks of human knowledge. Concepts help in comprehending the real- ity that a science is engaged in studying. Concepts act as mediums of short-cut communication among those associated with the enquiry (social scientists). Concepts and hypotheses are the core of social research. For any social research to be fruitful, it is important that the concepts or variables mentioned in the hypothesis are operationalised. As discussed earlier, operationalisation of concepts refers to the process of defining concepts in terms of those attributes which could be empirically observed during research. In other words, opera- tionalisation refers to the process of converting concepts in their empirical measurements. For example, the concept of alienation is generally explained in terms of powerlessness, meaninglessness, normlessness, social isolation and self- estrangement. Now, in a given workplace, powerlessness could be empirically measured in terms of the indicators such as participation in the administration, degree of control over the decision-making process, grievance redressal mecha- nism, etc. Let us now discuss the various problems in defining concepts in sociology. Social reality is dynamic in character, and so are the concepts in sociology. Sociology being a relatively young discipline relies more and more on empiri- cal research for verification and validation of the existing theories and concepts. Hence, new findings lead to modification of the established concepts and theo- ries. In other words, it leads to re-conceptualisation or re-specification of a con- cept. For example, earlier the personality differences between men and women were explained in biological terms. However, later-day research by anthropolo- gists like Margaret Mead, who in her work Sex and Temperament in Three Primitive Tribes studied three tribes, namely Arapesh, Mundugumor and Tchambuli (in the western Pacific), concluded that personality patterns were culturally determined rather than biologically. In brief, her comparative study revealed a full range of contrasting gender roles. Among the Arapesh, both men and women were peaceful in temperament and neither men nor women engaged in fight. Among the Mundugumor, the opposite was true: both men and women were of violent temperament. Among the Tchambuli, gender role reversal was found. While the men ‘primped’ and spent their time decorating themselves, the women worked and were the practical ones. Similarly, you can also discuss here how the concepts of class, caste ( jati), etc. have undergone changes with new findings put forward by later day researches. Second, sociology as a discipline is rapidly attaining maturity with the contri- butions of several established and highly reputed schools like the British school, the American school, the French school and the German school. But the con- tributions from diverse schools of thought give rise to the problem of ensuring uniformity in the definition and meaning of the concepts. Concepts develop from a shared experience. Since each school of sociological thought puts for- ward its own set of concepts and defines the concept in the context of its unique  Research Methods and Analysis 153 social setting, it gives rise to the problem of communication. For example, the concepts of Gemeinschaft and Gesellschaft, which were coined by the German soci- ologist Ferdinand Tonnies, have no English equivalent. The terms Community and Association, which are English translations of these words, do not convey the particular sociological meaning of these two German words. Third, due to the very subject matter of sociology, the terms used to denote scientific concepts may also have meanings in other frames of reference. For instance, the term ‘bureaucracy’, which implies a particular type of social struc- ture, may either be seen as a rationally designed authority structure or as an administrative institution characterised by red-tapism, corruption and official disregard for the public interest. Another problem associated with defining the concepts in social sciences is that the same term may refer to different phenomena. For example, the term ‘function’ in one sense may be used to denote the contributions which a given practice or belief makes towards the continued existence of society. However, the term function may also be used to denote the causal relation- ship between two variables; for example, in determining to what extent one variable (proliferation of nuclear families) is the function of another variable (industrialisation). Further, in social sciences, different terms or concepts may be used to refer to the same or similar phenomena. For example, the terms like formal–informal, organic–mechanistic, primary–secondary, community–association, etc. overlap to a great degree in their meaning. Another problem with regard to concepts in social sciences is that a given concept may have no immediate empirical referent. For example, the concepts like social system, social structure, etc. have no imme- diate empirical referents or quantifiable attributes. At best they can be studied by observing the patterns of relations among the members of a given society (Goode and Hatt, 2006). Thus, a social scientist must define the concepts as precisely as possible and operationalise the concepts in order to conduct a meaningful and result-oriented research. Techniques of Data Collection We have earlier discussed some of the fundamental characteristics of quantitative (positivist) and qualitative (anti-positivist) research methodologies. However, in practice, the distinctions between positivist and anti-positivist research method- ologies are not as clear-cut as the previous sections have shown. They have been placed at opposite ends of the spectrum for purposes of emphasis and illustration. A large body of sociological research falls somewhere between the two extremes. In the same way, the methods of data collection discussed in the following sections cannot be neatly categorised as aspects of positivist or anti-positivist methodologies. However, certain methods are regarded as more appropriate by supporters of one or other of these perspectives. 154 Research Methods and Analysis Let us now discuss some of the important techniques employed by the social scientists for collecting data from the field. The need for adequate and reliable data is ever increasing for taking policy decisions in different fields of human activity. There are two ways in which the required information may be obtained: 1. complete enumeration survey (also known as the census method) 2. sample method Under the complete enumeration survey method or census method, data are collected for each and every unit (person, household, field, shop, factory, etc., as the case may be) belonging to the population or universe, which is the com- plete set of items that are of interest in any particular situation. Since every unit is covered, this method ensures greater accuracy. However, it is a highly time- and money-consuming exercise, and that is why it is used very rarely and selectively for some specific purposes only, such as census. This is more true of underdeveloped countries where resources constitute a big constraint. Also, if the population is infinite, the method cannot be adopted. ‘Population’ here refers to ‘all those people with the characteristics which the researcher wants to study within the context of a particular research problem’. A population could be all students in a school, all patients in a hospital, all prisoners in a prison, etc. Hence, in modern times, very little use is made of complete enumeration survey. How to collect the data then? It is through the adoption of sampling technique that a large mass of data pertaining to different aspects of human activity are collected these days (Gupta, 1990). Sampling As mentioned earlier, when the population is relatively large or widely dispersed, researchers survey only a sample. Sampling is simply the process of learning about the population on the basis of a sample drawn from it. In the sampling technique, instead of every unit of the universe, only a part of the universe is studied and the conclusions are drawn on that basis for the entire universe. A sample is not studied for its own sake. The basic objective of the sample study is to draw inference about the entire population which it claims to represent. In other words, sampling is only a tool which helps to know the characteristics of the universe or population by examining only a small part of it. Features of a Good Sample Since conclusions are drawn on the basis of the study of a small part of the entire universe or population, it is necessary that a sample possesses certain essential characteristics. First, a sample should be representative enough. In other words, a sample should be so selected that it truly represents the universe, otherwise the results obtained may be misleading. Second, the size of sample should be  Research Methods and Analysis 155 adequate in relation to the universe it tends to represent, otherwise it may not represent all the characteristics of the universe. Third, all items of the sample should be selected independent of one another, and all items of the universe should have equal chance of getting selected in the sample. Last but not the least, there should be homogeneity between the units of the sample and that of the universe. Types of Sampling There are basically two types of sampling: probability sampling and non-prob- ability sampling. Probability sampling (also known as random sampling) is one in which every unit of the population has an equal probability of being selected for the sample. It offers a high degree of representativeness. This implies that the selection of sample items is independent of the person making the study – that is, the sampling operation is controlled so objectively that the items will be chosen strictly at random. Hence, it provides estimates which are essentially unbiased. However, this method is expensive, time-consuming and relatively complicated since it requires a large sample size and the units selected are usually widely scat- tered. Also, it requires a very high level of skill and experience for its use. Non- probability (or non-random) sampling makes no claim for representativeness, as every unit does not get the chance of being selected. It is the researcher who decides which sample units should be chosen. Probability sampling today remains the primary method for selecting large, representative samples for social science and business researches. Some of the important sampling designs or methods under this category are simple random sampling, stratified random sampling, systematic (or interval) sampling, cluster sampling and multi-stage sampling. Simple Random Sampling Random sampling refers to the sampling technique in which each and every item of the population is given an equal chance of being included in the sample. The selection is, thus, free from personal bias because the investigator does not exercise his discretion or preference in the choice of items. Since selection of items in the sample depends entirely on chance, this method is also known as the method of chance selection. Some people believe that randomness of selec- tion can be achieved by unsystematic and haphazard procedures. But this is quite wrong. However, the point to be emphasised is that unless precaution is taken to avoid bias and a conscious effort is made to ensure the operation of chance factors, the resulting sample shall not be a random sample. Random sampling is sometimes referred to as ‘representative sampling’. If the sample is chosen at random and if the size of the sample is sufficiently large, it will represent all groups in the universe. A random sample is also known as a ‘probability sample’ because every item of the universe has an equal opportunity of being selected 156 Research Methods and Analysis in the sample. To ensure randomness of selection, one may adopt any of the fol- lowing methods. 1. Lottery Method: This is a very popular method of taking a random sample. Under this method, all items of the universe are numbered on separate slips of paper of identical size and shape. These slips are then folded and mixed up in a container or drum. A blindfold selection is then made of the number of slips required to constitute the desired sample size. The selection of items, thus, depends entirely on chance. 2. Table of Random Numbers: The lottery method discussed above becomes quite cumbersome to use as the size of population increases. An alternative method of random selection is that of using the table of random numbers. Three such tables are available: (i) Tippett’s table of random numbers, (ii) Fisher and Yate’s numbers and (iii) Kendall and Babington Smith’s numbers. The merits of random sampling lie in the fact that since the selection of items in the sample depends entirely on chance, there is no possibility of personal bias affecting the results. Further, as the size of the sample increases, it becomes increasingly representative of the population. However, the use of random sam- pling necessitates a completely catalogued universe from which to draw the sam- ple. But it is often difficult for the investigator to have up-to-date lists of all the items of the population to be sampled. This restricts the use of random sampling method (Gupta, 1990). Stratifed Random Sampling In this sampling method, the population is divided into various strata or classes, and a sample is drawn from each stratum at random. For example, if we are interested in studying the consumption pattern of the people of Delhi, the city of Delhi may be divided into various parts or zones, and from each part a sam- ple may be taken at random. However, the selection of cases from each stratum must be done with great care and in accordance with a carefully designed plan as otherwise random selection from the various strata may not be accomplished. Stratified sampling may be either proportional or disproportional. In pro- portional stratified sampling, the cases are drawn from each stratum in the same proportion as they occur in the universe. For example, if we divide the city of Delhi into four zones A, B, C and D with 40, 30, 20 and 10 per cent of the total population, respectively, and if the sample size is 1,000, then we should draw 400, 300, 200 and 100 cases, respectively, from zones A, B, C and D, i.e. sample is proportional to the size in the universe. In disproportional stratified sampling, an equal number of cases is taken from each stratum, regardless of how the stra- tum is represented in the universe. Thus, in the above example, an equal number of items from each zone may be drawn, that is, 250. This approach is obviously inferior to the proportional stratified sampling.  Research Methods and Analysis 157 The most important merit of the stratified random sampling is that it is more representative. Since the population is first divided into various strata and then a sample is drawn from each stratum, there is little possibility of any essential group of the population being completely excluded. A more representative sam- ple is thus secured. Stratified sampling is frequently regarded as the most effi- cient system of sampling. However, utmost care must be exercised in dividing the population into various strata. Each stratum must contain, as far as possible, homogenous items as otherwise the results may not be reliable. However, this is a very difficult task and may involve considerable time and expense. Systematic or Interval Sampling This method is popularly used in those cases where a complete list of the popu- lation from which a sample is to be drawn is available. It involves obtaining a sample of items by drawing every nth item from a predetermined list of items. In other words, it involves randomly selecting the first respondent and then every nth person after that; ‘n’ is the sampling interval. For example, if a complete list of 1,000 students of a college is available and if we want to draw a sample of 200, then this means we must take every fifth item (i.e. n = 5). The first item between one and five shall be selected at random. Suppose it comes out to be three. Now we shall go on adding five and obtain numbers of the desired sample. Thus, the second item would be the 8th student, and the third would be the 13th student and so on. Systematic sampling differs from simple random sampling in that in the latter the selections are independent of each other; in the former, the selec- tion of sample units is dependent on the selection of a previous one. The sys- tematic sampling is more convenient to adopt than the random sampling or the stratified sampling method. The time and work involved in sampling by this method are relatively smaller. It is a rapid method and eliminates several steps otherwise taken in probability sampling. However, critics of this method argue that it ignores all persons between two nth numbers with the result that the possibility of overrepresentation and underrepresentation of several groups is greater. Cluster Sampling This sampling implies dividing the population into clusters and drawing ran- dom samples either from all clusters or from selected clusters. This method is used when (a) cluster criteria are significant for the study, and (b) economic considerations are significant. In cluster sampling, initial clusters are called pri- mary sampling units; clusters within the primary clusters are called secondary sampling units; and clusters within the secondary clusters are called multi-stage clusters. When clusters are geographic units, it is called area sampling. For exam- ple, dividing one city into various wards, each ward into areas, each area into 158 Research Methods and Analysis neighbourhoods and each neighbourhood into lanes. We can take an example of a hospital. The issue is to ascertain the problems faced by doctors, patients and visitors in different units and to introduce some reformative programmes. Administratively, it will not be viable to call all doctors from all units, nor a large number of patients admitted in different units like cardiology, neurology, orthopaedic, gynaecology and so on. Treating each unit as a cluster, randomly selected doctors and patients – say 2 doctors and 3 patients, or about 50 people all together – from all units may be invited for discussion. On the basis of such discussion, suggestions by the stakeholders may be submitted to higher authori- ties for necessary action. The advantage of cluster sampling is that it is much easier to apply this sam- pling design when large populations are studied or when a large geographical area is studied. Further, the cost involved in this method is much less than in other methods of sampling. The disadvantages of this sampling method are that each cluster may not be of equal size, and hence the comparison so done would not be on an equal basis. The chances of sampling error are greater as there could be homogeneity in one cluster but heterogeneity in other. Multi-stage Sampling As the name implies this method refers to a sampling procedure which is carried out in several stages, but only the last sample of subjects is studied. Suppose it is decided to take a sample of 5,000 households from the state of Uttar Pradesh. At the first stage, the state may be divided into a number of districts and a few dis- tricts are selected at random. At the second stage, each district may be subdivided into a number of villages, and a sample of villages may be taken at random. At the third stage, a number of households may be selected from each of the villages selected at the second stage. In this way, at each stage the sample size becomes smaller and smaller. The merit of multi-stage sampling is that it introduces f lex- ibility in the sampling method, which is lacking in the other methods. It enables existing divisions and subdivisions of the population to be used as units at various stages. It permits the fieldwork to be concentrated despite covering a large area. Another important advantage in this sampling design is that it is more repre- sentative. Further, in all cases, complete listing of population is not necessary. This saves cost. However, a multi-stage sample is in general less accurate than a sample containing the same number of final stage units which have been selected by some suitable single stages process. Let us now discuss about the non-probability sampling. In many research situations, particularly those where there is no list of persons to be studied (e.g. widows, alcoholics, migrant workers), probability sampling is difficult and inap- propriate to use. In such a research, non-probability sampling is the most appro- priate one. Non-probability sampling procedures do not employ the rules of probability theory, do not claim representativeness and are usually used for quali- tative exploratory analysis. Some of the important sampling designs under this  Research Methods and Analysis 159 category are convenience sampling, purposive or judgement sampling, quota sampling, snowball sampling and volunteer sampling. Convenience Sampling This sampling is also known as ‘accidental’ or ‘haphazard’ sampling. In this sam- pling, the researcher studies all those persons who are most conveniently availa- ble or who accidentally come in his contact during a certain period of time in the research. For example, the researcher engaged in the study of university students might visit the university canteen, library, some departments, playgrounds and verandahs and interview certain number of students. Another example is of elec- tion study. During election times, media personnel often present man-on-the- street interviews that are presumed to ref lect public opinion. In such a sampling, representativeness is not significant. The most obvious advantage of convenience sample is that it is quick and economical. But it may be a very biased sample. The possible sources of bias could be: (i) the respondents may have a vested interest to serve in cooperating with the interviewer, and (ii) the respondents may be those who are vocal and/or want to brag. Convenience samples are best utilised for exploratory research when additional research will subsequently be conducted with a probability sample. Purposive Sampling Purposive sampling is also known as judgement sampling. In this sampling, the choice of sample items depends exclusively on the discretion of the investigator. In other words, the investigator exercises his judgement in the choice and includes those items in the sample which he thinks are most typical of the universe with regard to the characteristics under investigation. For example, if a sample of 10 students is to be selected from a class of 60 for analysing the spending habits of students, the investigator would select 10 students who, in his opinion, are repre- sentative of the class. This method, though simple, is not scientific because there is a big possibility of the results being affected by the personal prejudice or bias of the investigator. Thus, judgement sampling involves the risk that the investigator may establish foregone conclusions by including those items in the sample which conform to his preconceived notions. For example, if an investigator holds the view that the wages of workers in a certain establishment are very low, and if he adopts the judgement sampling method, he may include only those workers in the sample whose wages are low and thereby establish his point of view, which may be far from the truth. Since an element of subjectiveness is possible, this method cannot be recommended for general use. However, because of simplicity and easy adaptability, this method is quite often used by businessmen in the solu- tion of everyday problems. Indeed, if applied with skill and care, the judgement method can be of great help to businessmen. At least, it helps deriving somewhat better solutions to the problems than could be obtained without it. 160 Research Methods and Analysis Quota Sampling Quota sampling is a type of judgement sampling. In a quota sample, quotas are set up according to given criteria, but within the quotas the selection of sample items depends on personal judgement. For example, in a radio listening survey, the interviewers may be told to interview 500 people living in a certain area and that out of every 100 persons interviewed, 60 are to be housewives, 25 farmers and 15 children. Within these quotas, the interviewer is free to select the people to be interviewed. The cost per person interviewed may be relatively small for a quota sample, but there are numerous opportunities for bias which may invalidate results. Because of the risk of personal prejudice and bias entering the process of selection, the quota sampling is rarely used in practical work. Snowball Sampling In this technique, the researcher begins the research with the few respondents who are known and available to him. Subsequently, these respondents give other names who meet the criteria of research, who in turn give more new names. This process is continued until ‘adequate’ number of persons are interviewed or until no more respondents are discovered. For instance, in studying wife batter- ing, the researcher may first interview those cases whom he knows, who may later on give additional names and who in turn may give still new names. This method is employed when the target population is unknown or when it is dif- ficult to approach the respondents in any other way. Reduced sample sizes and costs are a clear advantage of snowball sampling. Bias enters because a person known to someone (also in the sample) has a higher probability of being similar to the first person. If there are major differences between those who are widely known by others and those who are not, there may be serious problems with snowball sampling. Volunteer Sampling In this sampling, the respondent himself volunteers to give information he holds. The success of the research is dependent on the ‘rich’ information given by the respondents. However, there is a possibility that the informants may not truly represent the population, i.e. they may not have the aggregate characteristics of the population. Further, the personal leanings of the researcher of being preju- diced against certain types of persons, say, untouchables or religious minorities, may also affect the validity of the findings (Ahuja, 2008). On a review of the pros and cons of the various methods of sampling, it is clear that stratified sampling and systematic sampling methods based on ran- dom principle are more reliable, and hence, these methods are more widely used than others. Let us now brief ly discuss the merits of the sampling procedure in  Research Methods and Analysis 161 general. The sampling technique has the following merits over the complete enumeration survey: 1. Sampling is essentially useful in cases where the universe if large and scattered. 2. It consumes less time. Since sampling is a study of a part of the population, considerable time and labour are saved when a sample survey is carried out. Time is saved not only in collecting data but also in processing it. 3. Less cost is incurred in sampling compared to other techniques of data col- lection in terms of the cost involved. This is a great advantage particularly in an underdeveloped economy where much of the information would be difficult to collect by the census method for lack of adequate resources. 4. Sampling yields more reliable results. This is because more effective pre- cautions can be taken in a sample survey to ensure that the information is accurate and complete. Moreover, it is possible to avail of the services of experts and to impart thorough training to the investigators in a sample survey which further reduces the possibility of errors. Follow-up work can also be undertaken much more effectively in the sampling method. 5. Since the sampling technique saves time and money, it is possible to collect more detailed information in a sample survey. For example, if the popula- tion consists of 1,000 persons in a survey of the consumption pattern of the people, the two alternative techniques available are as follows: (a) We may collect the necessary data from each of the 1,000 people through a questionnaire containing, say, ten questions (census method), or (b) We may take a sample of 100 persons, i.e. 10 per cent of population, and prepare a questionnaire containing as many as 100 questions. The expense involved in the latter case may be almost the same as in the former, but it will enable nine times more information to be obtained (Gupta, 1990). However, despite the various advantages of sampling, it is not altogether free from limitations. Some of the difficulties involved in sampling are stated next. 1. A sample survey must be carefully planned and executed, otherwise the results obtained may be inaccurate and misleading. 2. Sampling generally requires the services of experts for the proper planning and execution of the survey. In the absence of qualified and experienced persons, the information obtained from sample surveys cannot be relied upon. 3. If the information is required for each and every unit in the domain of study, sample method cannot be adopted. To appreciate the need for sample surveys, it is necessary to understand clearly the role of sampling and non-sampling errors in complete enumeration and sam- ple surveys. The errors arising due to drawing inferences about the population 162 Research Methods and Analysis on the basis of few observations (sample) is termed as sampling errors. Clearly, the sampling error in this sense is non-existent in a complete enumeration sur- vey, since the whole population is surveyed. However, the errors mainly arising at the stages of ascertainment and processing of data which are termed non- sampling errors are common both in complete enumeration and sample surveys. Sampling Errors Even if utmost care has been taken in selecting a sample, the results derived from the sample may not be representative of the population from which it is drawn, because samples are seldom, if ever, perfect miniatures of the population. This gives rise to sampling errors. Sampling errors are, thus, due to the fact that sam- ples are used and to the particular method used in selecting the items from the population. Sampling errors are of two types: biased and unbiased. Biased errors are those which arise from any kind of bias in selection, estimation, etc. Bias may arise either due to a faulty process of selection or faulty method of analysis. Unbiased errors, on the other hand, arise due to chance differences between the members of the population included in the sample and those not included. The simplest and the only certain way of avoiding bias in the selection process is for the sample to be drawn either entirely at random or at random, subject to restric- tions, which, while improving the accuracy, are of such a nature that they do not introduce bias in the results. Non-sampling Errors When a complete enumeration of units in the universe is made, one would expect that it would give rise to data free from errors. However, in practice it is not so. For example, it is difficult to completely avoid errors of observation or ascertainment. Similarly, in the processing of data tabulation errors may be committed affecting the final results. Errors arising in this manner are termed non-sampling errors, as they are due to factors other than the inductive process of inferring about the population from a sample. Thus, the data obtained in a census by complete enumeration, although free from sampling errors, would still be subject to non-sampling errors, whereas the results of a sample survey would be subject to sampling errors as well as non-sampling errors. Non-sampling errors can occur at every stage of planning and execution of the census or survey. Such errors can arise due to a number of causes, such as defective methods of data collection and tabulation, faulty definitions, incom- plete coverage of the population or sample, etc. More specifically, non-sampling errors may arise from one or more of the following factors. 1. Data specification being inadequate and inconsistent with respect to the objectives of the census or sample survey 2. Inappropriate statistical unit  Research Methods and Analysis 163 3. Inaccurate or inappropriate methods of interviews, observation or measure- ment with inadequate or ambiguous schedules, definitions or instructions 4. Lack of trained and experienced investigators 5. Lack of adequate inspection and supervision of primary staff 6. Errors due to non-response, i.e. incomplete coverage in respect of units 7. Errors in data processing operations such as coding, punching, verification, tabulation, etc. 8. Errors committed during presentation and printing of tabulated results These sources are not exhaustive but are given to indicate some of the possible sources of error. In some situations, the non-sampling errors may be large and deserve greater attention than the sampling error. While, in general, sampling error decreases with increase in sample size, non-sampling error tends to increase with the sample size. In the case of complete enumeration, non-sampling error and in the case of sample surveys both sampling and non-sampling errors require to be controlled and reduced to a level at which their presence does not vitiate the use of final results (Gupta, 1990). Survey The most common type of empirical, or quantitative, research in sociology is the survey, which consists of systematically questioning people about their opinions, attitudes or behaviours. A social survey involves the collection of standardised information from a sample selected as being representative of a particular group or population. The group from which the sample is drawn may be the population as a whole, a particular class, ethnic, gender or age group, etc., depending upon the objective of the researcher. In a survey, standardised information is obtained by asking the same set of questions to all members of the sample. Questionnaires and structured interviews are two important and popular techniques of data col- lection in a social survey. Social surveys are broadly divided into two categories: descriptive and ana- lytical. Descriptive surveys are used to describe the world as it is. In other words, descriptive surveys are concerned with description rather than explanation. It aims to provide an accurate measurement of the distribution of certain char- acteristics in a given population. For example, a survey may be conducted in a city or town to measure the extent of poverty in the given population. Here the researcher and his team would be interested in collecting the data on average per capita income of the working-class stratum or the population below poverty line, etc. In other words, they aim to measure the extent of poverty in a given popula- tion rather than to explain the causes of poverty. Analytical surveys, on the other hand, are concerned with explanation. They are designed to test hypotheses about the relationships between a number of factors or variables. For example, an analytical survey may seek to discover possible relationships between social class and educational attainment, gender and occupation, etc. Analytical surveys 164 Research Methods and Analysis are not simply concerned with discovering relationships but also with explaining them (Haralambos and Heald, 2006). Analytical surveys are usually designed to test the effects of a number of variables or factors on some other variable. For example, a researcher may suggest that social class differences in some way cause or determine variations in educational attain- ment. However, there may be other factors also affecting educational attainment, and they must also be considered if the inf luence of social class is to be accurately assessed. For example, variables such as caste, gender and ethnicity may account for some variation in educational attainment. As a result, researchers usually gather data on a range of factors which might inf luence the variable in question. The method used to analyse relationships between variables is known as ‘multivari- ate’ or ‘variable analysis’. With the aid of various statistical techniques, the analyst attempts to measure the effects of a number of variables upon other variables. This method was pioneered in sociology by Emile Durkheim in his study of suicide. Official statistics revealed significant variations in suicide rate between European societies. Durkheim’s research indicated that predominantly Protestant societies had a higher rate of suicide in comparison to societies where Catholicism was the majority faith. But before a causal relationship could be claimed between religion and suicide rates, it was necessary to eliminate other possibilities. For example, could variations in suicide rates be the result of differences in national cultures? To test this possibility, Durkheim held the variable of national cul- ture constant by examining differences in suicide rates between Catholics and Protestants within the same society. The relationship still held. Within the same society Protestants had higher suicide rates than Catholics. To ensure greater objectivity in his research, Durkheim then went a step further and examined the possibility that regional differences rather than religion might account for variations in suicide rates. He found, for example, that Bavaria had the lowest suicide rate of all the states in Germany and it also had the highest proportion of Catholics. Yet, might the suicide rate be due to the peculiarities of Bavaria as a region rather than its predominantly Catholic population? To test this possibil- ity Durkheim compared the suicide rates and the religious composition of the various provinces within Bavaria. He found that the higher the proportion of Protestants in each province, the higher the suicide rate. Again, the relationship between religion and rates of suicide was confirmed. By eliminating variables such as national culture and region, Durkheim was able to strengthen the rela- tionship between religion and suicide rates and provide increasing support for his claim that the relationship is a causal one (ibid.). Let us brief ly discuss some of the major steps normally involved in survey research. First, before a survey is begun, the issues to be explored must be clearly defined. At the same time, the target population to be interviewed is selected. The target population might be identified on the basis of the characteristics that the researcher is interested in examining. It could either be a particular gender or age group, or any specific socio-economic section of the society, etc. This first step is crucial, for if the population is not correctly specified, the results of the  Research Methods and Analysis 165 survey may be meaningless. For example, if the aim of the research is to predict the results of an election, it is very important that the population chosen consist only of those persons who will actually vote in that election. Second, if the population is large, time and cost will almost always make it impractical to interview the entire population. So, the second step in surveying is to pick an appropriate sample of the population to interview. A sample is a limited number of selected cases drawn from a large group. Careful procedures have been established for selecting samples. The better the sampling procedure, the more closely the sample will resemble the entire population and the more accurate will be the generalisations or predictions. In other words, if generali- sations are to be made from the findings of a social survey, it is essential that the sample is representative. This is often accomplished by means of a ‘random sample’. We have already discussed the various types of sampling techniques and their merits and limitations. Once the researcher is satisfied that he has obtained a representative sample, he can begin the survey proper and feel some justification in generalising from its findings. Once the sample is selected, the third step in survey research is to interview or administer the questionnaire to the selected people and to collect the data. At this point a major consideration is the precision of the questions. Do the ques- tions really pinpoint the issues concerned? Are they phrased in such a way that they will be interpreted correctly and similarly by each person interviewed, i.e. the respondents. In addition to being precise and unambiguous in meaning, a survey question must also be neutrally stated. For most accurate results, the entire sample must be interviewed, particularly if the sample is small. If some people refuse to answer or are unavailable for inter- viewing, the sample is no longer representative and, consequently, the accuracy of the data may be reduced. Non-response is frequently a serious problem when questionnaires are sent by mail, for refusals to respond to mailed questionnaires tend to be high. Replies often come only from those who have some interest in the particular issue, thus introducing a bias into the survey findings. To assure maximum response, most major attitude surveys and public opinion polls are conducted through personal interviews. These interviews range from the highly structured to be highly unstructured. A structured interview consists of a set of questions and answers which are always stated in the same way and in the same sequence. The answers are, thus, easily compiled and generalised. Most public opinion polls use structured inter- views. For other research purposes, where more extensive information about individual attitudes or behaviour is desired, the unstructured interview has many advantages. An unstructured interview may consist of open-ended questions (How do you feel about the inter-caste marriages or caste-based reservations?) or even just a list of topics to be discussed. It is possible for the interviewer to introduce bias into the survey. He may, for example, use expressions or make comments that encourage the respondent to answer in a certain way. In an unstructured interview he may inf luence the answers by the way he phrases the 166 Research Methods and Analysis questions. It is important that interviewers be suited to their task and that they be well trained in the techniques of interviewing. The final step in survey research is the tabulation, analysis and interpretation of the data. In all but the smallest surveys, this step normally involves the use of computers. There are several possible sources of error in survey results. Sampling error is the degree to which the selected sample misrepresents the population as a whole. Other major sources of error arise from problems in observation and measure- ment, processing the data and analysing the findings. A basic problem with all surveys is that what people say may not always agree with how they act. People sometimes conceal their attitude purposely. An individual prejudiced against lower castes in India, for example, may act in a discriminatory fashion towards them, but because he knows that this prejudice is socially disapproved of, he will not admit it to an interviewer. Research that is well designed and carried out can help to overcome these difficulties, but the sociologist must be constantly aware that attitudes expressed in interviews are not always perfect expressions of under- lying values, and that actions do not always ref lect stated attitudes. The success of any survey is, however, ultimately dependent on the quality of its data. At the end of the day a social survey stands or falls on the validity of its data. Case Study Case study method is an ideal methodology when a holistic, in-depth investi- gation is needed. Frederic Le Play is reputed to have introduced the case study method into social science. He used it as a handmaiden to statistics in his studies of family budgets. Herbert Spencer was one of the first sociologists to use case materials in his ethnographic studies. Case study is an intensive study of a case which may be an individual, an institution, a system, a community, an organisation, an event or even the entire culture. Robert K. Yin has defined case study as ‘an empirical inquiry that investigates a contemporary phenomenon within its real-life context, when the boundaries between phenomenon and context are not clearly evident, and in which multiple sources of evidence are used’. It is, thus, a kind of research design which usually involves the qualitative method of selecting the sources of the data. It presents the holistic account that offers insights into the case under study. It is worth noting that while a case study can be either quantitative or qualitative, or even both, most case studies lie within the realm of qualitative methodology. It is the preferred strategy when ‘how, who, why and what’ questions are being asked, or when the focus is on a contemporary phenomenon within a real-life context (Ahuja, 2008). Case studies have been used in varied social investigations, particularly, in sociological studies, and are designed to bring out the details from the viewpoint of the participants by using multiple sources of data. It is, therefore, an approach to explore and analyse the life of social unit, be it a person, a family, an insti- tution, a culture group or even an entire community. Its aim is to determine  Research Methods and Analysis 167 the factors that account for the complex behaviour patterns of the unit and the relationships of the unit to its surroundings. Case data may be gathered, exhaus- tively, on the entire life cycle or on a definite section of the cycle of a unit but always with a view to ascertain the natural history of social unit and its relation- ship to the social factors and forces involved in its environment. In other words, through case studies researchers attempt to see the variety of factors within a social unit as an integrated whole. When attention is focused on the development of the case, it is called ‘case history’. For example, how a particular boy became a juvenile delinquent because of lack of parental control, impact of peers, lack of attention by teachers and money earned through cheap means, and then became an adolescent thief and a sex criminal and ultimately a professional pickpocket is tracing criminality through case history method. Data, for case studies, can be collected by primary as well as secondary sources. Two main sources of primary data collection are interviews and obser- vations. Interviews may be structured or unstructured. Most commonly, it is the unstructured interview which is used by the investigators. The observation method used could be either participant or non-participant, while the secondary data can be collected through a variety of sources like reports, records, newspa- pers, magazines, books, files, diaries, etc. Sjoberg has identified some essential characteristics of case study method, which are as follows: 1. The case study ‘strives towards a holistic understanding of cultural systems of action’. Cultural systems of action refer to sets of interrelated activities engaged in by the actors in a social situation. 2. Case study research is not sampling research. However, selection of the items or sources must be done so as to maximise what can be learned, in the lim- ited period of time available for the study. 3. Because they are intensive in nature, case studies tend to be selective, focus- ing on one or two issues that are fundamental to understanding the system being examined. 4. ‘Case studies are multi-perspectival analyses’. This means that the researcher considers not just the voice and perspective of the actors but also of the rel- evant groups of actors and the interaction between them. According to Black and Champion, some of the advantages of case study design are: 1. Case study makes holistic and in-depth study of the phenomenon possible. 2. It offers f lexibility with respect to using methods for collecting data, e.g. questionnaire, interview, observation. 3. It could be used for studying any specific dimension of the topic in detail. 4. It can be conducted in practically any kind of social setting. 5. Case studies are relatively inexpensive. 168 Research Methods and Analysis However, practically, the case study method is very time-consuming and demand- ing of the researcher. The possibility of becoming involved emotionally is much greater than in survey research, thus making detached and objective observation dif- ficult and sometimes impossible. Another problem in the use of case study method is that since only one example of a social situation or group is being studied, the results may not be representative of all groups or situations in the category. In other words, the particular mental hospital ward, slum or suburb may not be typical of all mental hospital wards, slums or suburbs. Critics of the case study method believe that the study of a small number of cases can offer no grounds for establishing reli- ability or generality of findings. Some dismiss case study research as useful only as an exploratory tool. Yet, researchers continue to use the case study research method with success in carefully planned and crafted studies of real-life situations, issues and problems. In comparison to survey, the case study method is more intensive, while survey research is more extensive in nature. In other words, surveys are usually con- ducted on a fairly large scale in contrast to case studies that tend to be more intensive but on a smaller scale. Case study is done in terms of limited space and broader time, whereas survey is done in terms of limited time with broader space. Case study is inward looking, while survey method is outward looking. Interviews Interviews are one of the most widely used methods of gathering data in soci- ology. They consist of the researcher asking the interviewee or respondent a series of questions. Bingham and Moore have described interview as ‘a conver- sation with a purpose’ (Ahuja, 2008). According to Goode and Hatt (2006), ‘Interviewing is fundamentally a process of social interaction’. Interviews can be classified as ‘structured’ or ‘unstructured’ though many fall somewhere between these two extremes. In a structured interview, there is a set of predetermined questions which the interviewer is required to put before the interviewee (respondent) to collect the required information. In this type of interview, the wording of the questions and the order in which they are asked remains the same in every case. The result is a fairly formal question and answer session. However, unstructured interviews are more like an informal conversation. In an unstructured interview, there are no specifications in the wording of the questions or the order of the questions. The interviewer usually has particular topics in mind to cover but few, if any, preset questions. There are no specifications in the wording of the questions or the order of the questions. He has the freedom to phrase questions as he likes, ask the respondent to develop his answers and probe responses which might be unclear and ambiguous. This freedom is often extended to the respondent, who may be allowed to direct the interview into areas which interest him (Haralambos and Heald, 2006). Data from structured interviews are generally regarded as more reliable. Since the order and wording of questions are the same for all respondents, it is more  Research Methods and Analysis 169 likely that they will be responding to the same stimuli. Thus, different answers to the same set of questions will indicate real differences between the respondents. Different answers will not, therefore, simply ref lect differences in the way ques- tions are phrased. Thus, the more structured or standardised an interview, the more easily its results can be tested by researchers investigating other groups. The structured interview reduces the interviewer’s bias to the minimum. This form of interview is largely employed in quantitative research. By comparison, data from unstructured interviews are seen as less reliable. Questions are phrased in a variety of ways, and the relationship between the interviewer and the respondent is likely to be more intimate. It is unclear to what degree the answers are inf lu- enced by these factors. Differences between respondents may simply ref lect dif- ferences in the nature of the interviews. It is therefore more difficult to replicate an unstructured interview but the greater f lexibility of unstructured interviews may strengthen the validity of the data. They provide more opportunity to dis- cover what the respondent ‘really means’. Ambiguities in questions and answers can be clarified and the interviewers can probe for shades of meaning (ibid.). Structured interviews are largely employed in quantitative research. Such interviews are regarded as appropriate for obtaining answers to questions of ‘fact’ such as the age, sex and income of the respondent. Unstructured interviews, on the other hand, are seen as more appropriate for eliciting attitudes, opinions and interests. Interview data are often taken as indications of respondents’ attitudes and behaviour in everyday life, although what a person says in an interview may have little to do with his normal routines. Even if the respondent does his best to provide honest answers, he may be unaware of the taken-for-granted assump- tions which he employs in everyday life. Various studies have suggested, though, that interviews pose serious prob- lems of reliability and validity. This is partly due to the fact that interviews are interaction situations. Thus, the results of an interview will depend in part on the way the participants define the situation, their perceptions of each other and so on. Most studies have been concerned about the effects of interviewers on the respondents. The significance of what has come to be known as ‘inter- viewer bias’ can be seen from the research conducted by Katz. The classic study of socio-economic status of interviewers was conducted by Katz in 1942 in a lower-class area in Pittsburgh. Katz compared the results obtained by a group of interviewers who were blue-collar industrial workers with the results obtained by middle-class interviewers. Katz found that low-income industrial workers consistently gave more radical answers on labour issues to the blue-collar lower- class interviewers than to the middle-class interviewers. Katz concluded that lack of rapport between lower-class respondents and middle-class interviewers led to bias in response (Bailey, 1994). In a similar study, J. Allan Williams Jr. concluded that the greater the status differences between interviewer and respondent, the less likely the respondent will be to express his true feelings. In a series of inter- views organised in 1960s, Williams found drastic difference in the responses of the respondents when interviewed by black and white interviewers at different 170 Research Methods and Analysis points of time. Williams found that on issues such as civil rights demonstrations and school desegregation, black respondents often tended to give the answers they felt that white interviewers wanted to hear. These findings suggest that when status differences are wide, as is often the case with middle-class sociolo- gists interviewing members of the lower working class, interview data should be regarded with caution (Haralambos and Heald, 2006). Interviewers, like everybody else, have values, attitudes and expectations. However, no matter how much the interviewer tries to disguise his views, they may well be communicated to the respondent. This is particularly likely in the more informal situation of the unstructured interview. As a result, the inter- viewer may ‘lead’ the respondent whose answer will then ref lect something of the interviewer’s attitudes and expectations. This can be seen from a study con- ducted by Stuart A. Rice in 1914: 2,000 destitute men were asked, among other things, to explain their situation. There was a strong tendency for those inter- viewed by a supporter of prohibition to blame their demise on alcohol, but those interviewed by a committed socialist were much more likely to explain their plight in terms of the industrial situation. To counter this problem, interviewers are often advised to be ‘non-directive’, to refrain from offering opinions, to avoid expressions of approval and disapproval. It is suggested they establish ‘rapport’ with their respondents, that is, a warm, friendly relationship which implies sym- pathy and understanding, while at the same time guarding against communicat- ing their own attitudes and expectations (ibid.). However, despite these limitations, interviews do have certain advantages. They are less costly and time-consuming and can cover much larger samples. Further, the response rate of the interview method is high, particularly when compared to mailed questionnaires. Most importantly, the validity of the infor- mation can be checked. Since the respondent’s confidence can be sought through personal rapport, in-depth probing is possible. The interviewer can explain dif- ficult terms and clear up any confusion and misunderstandings. He gets the opportunity to observe the non-verbal behaviour of the respondent, which thus enables him to record the responses in the right perspective. Questionnaires A questionnaire consists of a list of preset questions to which respondents are asked to supply answers. Questionnaire poses a structured and standardised set of questions, either to one person or to a small population, or most commonly to respondents in a sample survey. Structure here refers to questions appearing in a consistent, predetermined sequence and form. Researchers who use question- naires regard them as a comparatively cheap, fast and efficient method for obtain- ing large amounts of quantifiable data on relatively large numbers of people. Questionnaires may be administered in a number of ways. Questionnaires may either be distributed by mail or by hand, through arrangements such as the ‘drop- off ’, where a fieldworker leaves the questionnaire for respondents to complete by  Research Methods and Analysis 171 themselves, with the provision either for mailing the complete form back to the research office or for a return call by the fieldworker to collect the questionnaire. A questionnaire administered in a face-to-face interview, or over the telephone (growing in popularity among researchers), is generally termed a ‘schedule’. Often, they are given to individuals by interviewers, in which case they take the form of structured interviews. This method was used by Goldthorpe and Lockwood in the aff luent worker study and by Young and Willmott in their survey of fam- ily life in London conducted in 1970. This method has the advantage of having a trained interviewer on hand to make sure that the questionnaire is completed according to instructions and to clarify any ambiguous questions. However, ques- tionnaires administered by interviewers involve the problem of ‘interviewer bias’. Further, this method is expensive in comparison to other alternatives available such as postal questionnaire. As its name suggests, a postal questionnaire is mailed to respondents with a stamped addressed envelope for return to the researcher. It is a cheaper method of data collection, especially if the respondents are dispersed over a large geographical area. However, the response rate of postal questionnaires is low. In deciding upon one of these methods, the researcher balances the cost, probable response rate and the nature of the questions to be posed. It is important to note that the set of structured questions in which answers are recorded by the interviewer himself is called interview schedule, or simply the schedule. It is distinguished from the questionnaire in the sense that in the questionnaire the answers are filled in by the respondent himself. Though the questionnaire is used when the respondents are educated, schedule can be used for both the illiterate and the educated respondents. The questionnaire is espe- cially useful when the respondents are scattered in a large geographical area but the schedule is used when the respondents are located in a small area so that they can be personally contacted. The wording of the questions in the questionnaire has to be simple, since the interviewer is not present to explain the meaning and importance of the question to the respondent. In the schedule, the investigator gets the opportunity to explain whatever the respondent needs to know. Questionnaires could broadly be classified into three types: standardised ques- tionnaire, open-ended questionnaire and close-ended questionnaire. Standardised ques- tionnaires are those in which there are definite, concrete and pre-ordained questions with additional questions limited to those necessary to clarify inad- equate answers or to elicit more detailed responses. The questions are presented with exactly the same wording and in the same order to all the respondents. The reason for standardised questions is to ensure that all the respondents are replying to the same set of questions. Here the respondents or the researcher mark certain categories of reply to the questions asked, for instance, ‘yes/no/don’t know’ or ‘very likely/likely/unlikely/very unlikely’. Standardised questionnaires have the advantage that responses are easy to compare and tabulate, since only a small number of categories are involved. On the other hand, because the standardised questions do not allow for subtleties of opinion or verbal expressions, the infor- mation they yield is likely to be restricted in scope. 172 Research Methods and Analysis Open-ended questions allow the respondent to compose his own answers rather than choosing between a number of given answers. For example, ‘What’s your view on the reservation policy in India?’ Open-ended questionnaires are designed to permit a free response from the subject rather than one limited to certain alternatives. This may provide more valid data since he can say what he means in his own words. However, this kind of response may be difficult to clas- sify and quantify. Answers must be carefully interpreted before the researcher is able to arrive at certain conclusion. Close-ended or fixed-choice questions, on the other hand, require the respondent to make a choice between a number of given answers. For exam- ple, ‘Do you agree with the reservation policy in India?’ The answer choices given are, ‘yes’, ‘no’, and ‘partly’. From the point of view of the interpretation of questionnaires, the closed question is preferable. The results are unambiguous and comparable. With an open question, the heterogeneous answers must first be ordered into classes (codified) before they can be interpreted. Constructing classes in this way is sometimes very laborious and a challenging task. From the point of view of the reliability of interview data also, the closed question is pref- erable. This is because that the response to an open-ended question is subjected to the perception and linguistic ability of the respondent and under certain cir- cumstances this can produce serious distortions. Although the content of questionnaires is governed by the purpose of the study, many problems of communication may still arise on all surveys regardless of the content. Much careful attention and experimentation are needed to pro- duce effectively worded questions. The language should be concise and directed towards producing uniformity of understanding among the respondents. Great care is therefore needed in designing a questionnaire. Sometimes the main survey is preceded by a ‘pilot study’, which involves giving the questionnaire to a group similar to the population to be surveyed. This helps to clear up any ambiguity in the wording of questions and to ensure their relevance to future respond