Techniques of data collection




3 B. Techniques of data collection Primary Sources: Case Studies: A case study involves the detailed examination of a single example of something and is therefore bound to lack external validity. It could involve the study of a single institution, community, or social group, an individual person, a particular historical event, or a single social action. Case studies are not representative and instead attempt to arrive at a comprehensive understanding of the group under study. They can be used to falsify a general theory, produce typologies (a set of categories defining types of a social phenomenon), and generate new hypotheses which can then be tested in later studies. Robert K. Yin distinguished between five types of case studies:

  1. The critical case, where a particular example is useful for testing a hypothesis. Example: a study of a religious cult examining how its members reacted when the world failed to end on the day predicted by their leader.
  2. The unique case, where there is only one known example of something. Example: Gough and the Nayar society.
  3. The revelatory case, where the researcher can gain access to an aspect of social life which was previously inaccessible. Example: Elliot Liebow in the USA in the 1960s.
  4. The longitudinal case provides a chance to study the case at two or more points in time, making it possible to see the effects of social change. Example: Beverley Skeggs - studying working-class women in England over a period of twelve years.
  5. The representative or typical case or exemplifying case which illustrates a common social phenomenon. It is not possible to generalize based on the findings of a case study. Bryman suggests conducting a number of case studies of the same type of phenomenon, preferably by the same researcher and at the same time period, to improve validity. Example: Shoshana Zuboff carried out case study research in eight organizations in order to try to make generalizations about the impact of IT. Life Histories: Life histories are a particular type of case study - the whole of a case study is also sometimes known as the biographical method. Life histories can be carried out using a variety of methods but most frequently use extended, unstructured interviews. Sometimes they also make considerable use of personal documents. Examples: a study of the life of a Polish peasant conducted by Gordon Allport, a study of an aging woman, and extreme or unique cases by Margaret Mead, Robert Bogdan, Thomas and Znaniecki.

Life histories can falsify existing theories, inspire new ones, or help formulate a hypothesis. Feminist researchers argue that life histories help women understand their situation, and once they gain consciousness and awareness of their own exploitation by encouraging them to reflect upon the factors that have shaped their life experiences. Postmodernists have also used life histories. Example: Judith Stacey.

Pilot Studies: Having selected a research method and chosen a method of selecting a sample, some sociologists carry out a pilot study before embarking upon the main research project. A pilot study is a small-scale preliminary study conducted before the main research to check the feasibility or improve the design of the research. They are generally carried out before large-scale quantitative research to avoid time and money being wasted on an inadequately designed project. A pilot study is usually carried out on members of the relevant population but not on those who will form part of the final sample, as that might influence their later behavior.

Pilot studies have the following uses:

  1. If interviews or questionnaires are to be used, the questions can be tested to make sure that they make sense to respondents, are unambiguous, and produce the required information. This can help improve the reliability and response rate of the research.
  2. It can help researchers develop ways of building rapport and developing full cooperation of the subjects, so as to garner open and honest answers.
  3. It can help develop the research skills, especially of amateur interviewers.
  4. It may determine whether or not a research goes ahead. Funding organizations may demand the result of a pilot before green-lighting the whole research. If a pilot study is unsuccessful, the full study may be abandoned.

Quantitative Primary Research - Social Surveys: Social surveys can be defined as research projects that collect standardized data about large numbers of people. The data are usually in a statistical form, and the most practical way of collecting such data is through the use of questionnaires. Stephen Ackroyd and John A. Hughes distinguished three main types of surveys:

  1. The factual survey is used to collect descriptive information. Such surveys have been used to collect data on poverty and social exclusion. Example: government census.
  2. The attitude survey attempts to discover the subjective states of individuals. Example: attitude towards government policies or towards a political party.
  3. The explanatory survey tries to test theories and hypotheses or to produce new theories. Most sociological surveys contain some explanatory element. Example: Marshall et al. tested the theory that routine white-collar workers had become proletarianized.

Researchers usually want to be able to generalize from social surveys, and so surveys are based on carefully selected samples, and the success of any survey depends on the quality of data it produces.

Quantitative Primary Research - Questionnaires: A questionnaire consists simply of a list of pre-set questions. In questionnaire research, the same questions are usually given to respondents in the same order so that the same information can be collected from every member of the sample.

Administering questionnaires:

  1. Structured interview - here the interviewer himself asks the questions. These have the advantage of having a trained interviewer on hand to make sure that the questionnaire is completed properly and to clarify any ambiguities. However, they may involve the problem of interviewer bias.
  2. Postal - it is mailed to respondents, with a stamped addressed envelope. The response rate is usually low, and there may be systematic differences between people who return the questionnaire, and those who do not, thus leading to biased results.
  3. Group - such as a classroom of students or workers at a union meeting. This is a less expensive method, but care should be taken to ensure that no discussions take place among the respondents before answering.
  4. Telephone - it is relatively difficult to build a rapport or ask sensitive questions. No visual aids can be used.
  5. Email - the sample may be unrepresentative due to the digital divide.

Producing questionnaires and analyzing the data: Questionnaires tend to be used to produce quantitative data while testing a hypothesis. Some idea of what factors are important is needed before constructing a questionnaire.

In the process of choosing questions, researchers have to operationalize concepts - abstract concepts have to be translated into concrete questions that make it possible to take measurements relating to those concepts. First, an operational definition is established by breaking the concept down into various components or dimensions. Then, indicators are selected for each component. And finally, indicators of each dimension are put into the form of a series of questions that will provide quantifiable data for measuring each dimension. Questions may be open-ended, allowing the respondent to compose their answers. However, this may be difficult to classify and quantify. If the researcher wants the data to be in a statistical form, coding is necessary. It involves identifying a number of categories into which answers can be placed. Closed questions require a choice between a number of given answers. Such questions are easy to classify and quantify. Questions can also be direct or indirect (ask about other people).

Primary questions elicit information directly related to the research topic. Secondary questions elicit information not directly related to the topic but guards the truthfulness of the respondents. Tertiary questions only establish a framework that allows convenient data collection and building of rapport.

Questions can be nominal (fall in only one of the categories), ordinal (ranking), or interval. Once the data has been collected and classified, it is analyzed. Statistical tools and multivariate analysis are used here.

Advantages of questionnaires:

  1. Practical way to collect large quantities of data from considerable numbers of people over a relatively short period of time.
  2. Results are easily quantifiable and can be analyzed more scientifically and objectively than qualitative data.
  3. Can be used inductively or deductively, to try and establish cause-effect relationships through multivariate analysis.

Disadvantages of questionnaires:

  1. Interactionists see statistical data as inadequate for producing sociological explanations of human behavior.
  2. Phenomenologists see the data produced as an artificial creation of the researcher.
  3. Reliability is high but validity is questionable.
  4. Respondents may interpret the wordings of a question differently from what the researcher assumes, and as there is no opportunity to qualify meaning, it may mean different things to different social groups.
  5. Researcher imposition, as he assumes to know what is important and is limited to testing theories he has already thought of.
  6. Operationalizing of concepts and coding of answers from open-ended questionnaires also involve researcher imposition, thus producing a distorted picture of the social world.
  7. Validity may be reduced due to the inability or unwillingness of the respondents to answer, or they may simply lie or suffer from faulty memory or lack the relevant information. Also, it cannot be assumed that stated attitudes will be translated into actual behavior.

Interviews: Types: A completely structured interview is simply a questionnaire administered by an interviewer just like reading a script, while at the other extreme, a totally unstructured interview takes the form of a conversation with no predetermined questions. Most interviews fall somewhere in between these two extremes. Similarly, there can be standardized or unstandardized interviews, self-administered or other-administered interviews, personal or non-personal interviews, unique or panel interviews.

Styles: Non-directive - the interviewer refrains from offering opinions or expressing approval or disapproval. Howard Becker states that such a passive style may inhibit the interviewer, and instead suggests a more aggressive and active approach. This can provide fuller data. Example: he adopted such tactics in his interviews with Chicago school teachers, to understand the way they classified and evaluated students in terms of their class and ethnic backgrounds - information they would have preferred to keep hidden for fear of being accused of prejudice and discrimination. Ann Oakley prefers an empathetic approach rather than an aggressive one.

Numbers: It is normal for a single interviewer to interview a single respondent as it helps in building rapport and ensures confidentiality and no distraction or influence. However, sometimes group interviews are preferred. Example: Paul Willis, in his study of education, interviewed several of the group relations and allowed him to observe their interactions. James Holstein and Jaber Gubrium state that multivocality in group interviews broadens interviews and can make the participants more reflexive. Alan Bryman defines focus group as having several members discussing a topic that has been carefully specified, with the aim of using the discussion to construct meaning as a group, symbolic interactionist theory as it allows the researcher to observe how a group of people, through interaction with each other, arrive jointly at meaning and understanding. David Morgan suggests that this method can be used with groups of individuals who have a shared interest or area of expertise as they can stimulate each other, and will need minimum intervention by the interviewer. Sue Wilkinson egalitarian than one-to-one interviews, the interviewer dominates them less (feminist ethics), but are less useful for making systematic comparisons between social groups. Fran Tonkiss forums or chat rooms.

Advantages:

  1. Can be used on larger samples than participant observation
  2. Because there is usually some degree of structure in an interview, it is easier to make direct comparison than it is by using data from participant observation
  3. Unlike questionnaires, can be used to generate new hypotheses and theories which the researcher would not have thought of
  4. More practical, flexible, can produce valid data about suppressed views and sensitive issues, allow opportunity for critical reflection
  5. Can be used to carry out research into groups who might not otherwise consent to being the subject of research, as trust can be gained during interviews. Example: Laurie Taylor interviews with professional criminals in Britain

Disadvantages:

  1. Responses may not be accurate and may not reflect real behavior, may lie, forget or lack information. Example: some criminals lied to Laurie Taylor by making up fanciful stories about their escapades, social desirability bias among the answering participants
  2. Interviewees may be influenced by the presence of the interviewer and the interview setting. Example: William Labov found that young black American children responded differently to a white interviewer in a formal setting and to a black interviewer in an informal setting. Age, race, sex, clothing, and accent of the interviewer may affect the respondents.
  3. Interviewer bias, especially in unstructured interviews. Respondents may consciously or unconsciously give the sort of answers they believe the interviewer wants to hear. Example: By Stuart A. Rice - two thousand destitute men were asked to explain their situation. There was a strong tendency for those interviewed by a supporter of Prohibition to blame their decline on alcohol, but those interviewed by a committed socialist were much more likely to explain their plight in terms of the industrial situation.

The interviewer must be aware of the social conventions of those being interviewed to interpret the responses correctly. For example, Bruce Dohrenwend conducted research in New York to find the relationship between mental health and ethnicity. Respondents were asked whether or not they had experienced a list of symptoms associated with mental illness. Compared to Jews, Irish, and Black individuals, Puerto Ricans reported experiencing more of the symptoms. However, this was because they were more willing to admit them than other groups who found the symptoms undesirable.

5. Hammersley and Gomm argued that internal feelings, motives, intentions, and so on cannot be adequately expressed through verbal responses.

6. Interviewing lacks the rigor of scientific research and is less objective or systematic. Ecological validity may be a problem.

Observation: Positivists believe that the social world can be objectively observed, classified, and measured. Qualitative social researchers like interactionists and phenomenologists also favor observations. For example, J. Maxwell Atkinson observed the process of decision-making by coroners. However, there are limits to sit data produced. There are many social situations in which the presence of an observer is prohibited. It is not likely to be an option when studying processes whereby individuals adopt new behaviors which you are not likely to be able to predict. For example, Alan Bryman says a researcher could not become vegetarianism.

Ethnography: Geoff Payne and Judy Payne study how people in a social setting lead their lives based upon systematic and long-term observation of, and.

Participant Observation: Observation is the purposeful utilization of vision as a means of collecting data. Participant observation involves the investigator becoming a part of the social environment that he intends to study. The observations he makes may also be supplemented by interviews and questionnaires. Overt participant observations include M.N. Srinivas and Sudhir Venkatesh. Nigel Fielding a covert participant observation. Sometimes, researchers choose to be partially open but do not provide those being studied with the full story - William Foote Whyte in his study of an Italian American Slum introduced himself as a writer, without elaborating further. Examples include Howard Parker's study of Liverpool delinquents, Teela Sanders' study of sex workers, a study of Glasgow gangs, and Simon Holdaway - a cop who became a sociologist, providing a true insider view.

Advantages:

  • Less researcher imposition, higher levels of ecological validity
  • Can capture non-verbal elements like body language, group dynamics, etc.
  • More difficult for the subjects to lie or mislead the observer
  • Intimate and informal setup
  • Useful for generating new hypotheses or for falsifying theories

Criticism:

  • The observer may become so involved that he loses objectivity in the observation
  • Not all dimensions of a phenomenon can be observed simultaneously
  • Difficult to replicate, hence low reliability based on subjective interpretation of the observer
  • Very time-consuming and demanding, ethical or legal issues, especially in covert observation
  • Only a small sample can be studied, generalizations cannot be made
  • Validity of the data may be affected by the presence of the observer
  • Sometimes may be difficult to gain entry into the group to be studied, for example, drug gangs

Member validation, triangulation, and testing the hypotheses generated in follow-up studies are methods to refine the findings of participant observation.

Critical Ethnography: This is the sort of ethnography advocated by supporters of critical social science. They believe that ethnography can be used both, to develop and to test theories, including theories that examine the structure of society as a whole. It is seen as a method for researchers to understand how oppression is experienced by the oppressed by sharing some of the same experiences. An example is Paul Willis, who relied on interviews, but other critical ethnographers have also used participant observation and other methods. Mac an Ghaill tries to develop theories of masculinity by studying two student groups - one heterosexual and one homosexual, in the same area and belonging to the same social class. In the course of the study, he had some success in encouraging the gay students to value positively their conceptions of masculinity than being defensive in the face of hostility from heterosexuals. Circular arguments - the ethnographic description is used both for developing theory and for testing it. Experience comes to be interpreted in terms of the theory, yet the experience is also used to confirm the theory. The criticisms of critical social science methodology apply here too.

Longitudinal Research: Some researchers study a group over an extended period, collecting data on them at regular intervals. Such studies are known as longitudinal or panel studies. Quantitative as well as qualitative studies can be conducted using this method. Longitudinal studies originated as extended public attitude surveys to assist in policy decisions. Since then, they have been used in a variety of researches. An example is West and Farrington who followed 411 London schoolboys from age 8 to 18. Their major advantage is the ability to pick up change. They are also more likely to provide valid data than other types of research. Wall and Williams point out that retrospective studies that ask people to report on past events in their lives rely upon fallible human memories. Longitudinal studies help overcome this problem, and also, avoid the pitfall of events being reinterpreted in the light of subsequent consequences. Quantitative longitudinal studies often examine a large number of variables because researchers are unsure what data may prove to be important or required later in the research. However, it is necessary to select people who are accessible and willing to cooperate over an extended period. Furthermore, the size of the sample is liable to fall as some individuals become unwilling to continue to take part or prove impossible to trace. Rutterford comments that it is hard to disentangle age effects from cohort effects. Also, subjects of the research are conscious of the fact that their behavior is being studied. This may influence them to change their behavior because they think more carefully about their actions. Variants of panel studies - cohort studies, multi-stage studies, multi-phase studies.

Secondary Sources: These consist of data that have already been produced. Sociologists often use secondary data produced by the government. Organizations such as trade unions, companies, and charities are a useful source of data, as are documents such as letters, diaries, and autobiographies produced by individuals. They may be contemporary or historical, and the data available from them may be primarily qualitative or quantitative. However, their reliability and validity remain open to question.

Official Statistics: Government data such as the census, birth and death rates, marriage and fertility patterns, divorce, crime statistics, unemployment rates, inflation, etc. Positivists like Durkheim accepted official statistics uncritically. He believed the suicide statistics were sufficiently reliable and valid to measure the extent and social distribution of suicide. Ethnomethodologists and phenomenologists reject the use of statistics for measuring or determining the causes of the social facts to which they claim to refer. They adopt a constructionist position, arguing that data are socially constructed rather than a description of an external reality. Cicourel and Atkinson believe that statistics are a product of the meanings and taken-for-granted assumptions of those who construct them. Conflict sociologists argue that official statistics are neither hard facts nor subjective meanings. Instead, they consist of information that is systematically distorted by power structures in society. Miles and Irvine say that

Chrono-malleable
5. Multi-modal
6. In the context of social construction, it seems to produce a sense of meaning and identity that comes to be seen as real. Teela Sanders used the internet extensively in her study on sex workers. She got ethnographic information about the sex industry from using the website Punternet, which provided message boards for their clients. She also used emails to contact and then interview some of the sex workers.

Methodological Pluralism: Today, in sociological research, it is difficult to see quantitative methods (favored by positivists) and qualitative methods (favored by anti-positivists) as mutually exclusive. Combining different methods of research is becoming increasingly common, and this is known as methodological pluralism. As Alan Bryman points out, quantitative data tends to produce a static picture which allows researchers to examine and discover overall structures and patterns in society as a whole, while qualitative data allows a richer and deeper understanding of the process of change in social life, and the meanings and motives at the level of the actors. Example: In her study of the Unification Church, Eileen Barker used participant observation, questionnaires, and in-depth interviews. This allowed her to see how the Church as a whole was organized and also how it influenced the day-to-day actions and interactions of its members. Martyn Hammersley suggests three approaches to methodological pluralism: Triangulation, where each method is used to cross-check the findings of other methods; Facilitation, where one method assists the use of another method; Complementarity, where each method studies a different aspect of the subject being studied. Anthony Giddens' structuration theory also bridges the gap between macro and micro methodology. However, Bryman cautions that plural methodologies require the same careful use as any single methodology and have their own limitations. He elaborates on ten ways in which multi-strategy research can be used: 1. The logic of triangulation 2. Qualitative research facilitates quantitative research 3. Quantitative research facilitates qualitative research 4. Filling in the gaps 5. Static and processual features 6. Perspectives 7. The problem of generality 8. Qualitative research facilitating the interpretation of the relationship between variables 9. Studying different aspects of a phenomenon 10. Solving a puzzle


Also Watch Video on Youtube

Techniques of data collection