2.1 An Overview of Research Methods
Quantitative research is research that translates the social world into numerical data that can be treated mathematically. It often tries to find cause-and-effect relationships.
Qualitative research is research working with nonnumerical data such as texts, field notes, interview transcripts, photographs, and tape recordings. It often tries to understand how people makes sense of their world.
The Scientific Method
Identify a problem. / Ask a question.
Define variables, provide precise operational definitions, and form a hypothesis.
Choose a research design or study method.
Analyze data: Evaluate the accuracy or inaccuracy of the hypothesis in predicting the outcome.
Repeat experiment. Replicability is an important step for scientific work.
One limit to the scientific method is spurious correlation. This arises due to an intervening variable that makes it difficult to distinguish between correlation and causation.
Paradigm shifts occur when new data generated by research force us to look at the world in a different way.
Which method to use?
All methods have their pros and cons. The choice of method is guided by what the researcher wants to accomplish sociologically, what methods the researcher is competent in, how much time the project is allotted, resources and funding, and access to people.
2.2 Ethnography/Participant Observation
Ethnography is not only a qualitative method based on studying people in their own environment but also the term for the written report resulting from this study.
Ethnographers can be covert or overt, but they must always establish a good rapport with their subjects in order to gain access to them, since research is conducted through participant observation.
Data are collected through detailed field notes featuring thick description.
Ethnographers should always be aware of reflexivity, as their presence, personal feelings, and close ties to subjects can affect the study.
To analyze data, ethnographers use grounded theory, an inductive approach whereby identifying relationships between specific data categories allows for the building of broader theoretical propositions.
Example: Promises I Can Keep: Why Poor Women Put Motherhood before Marriage, by Kathryn Edin and Maria Kefalas.
Advantages: Allows study of subcultures; may challenge taken-for-granted notions; reshapes stereotypes and public policy; allows for innovation.
Disadvantages: Lacks replicability; limits in degree of representativeness; prone to bias.
Interviews are face-to-face, information-seeking conversations that gather qualitative data directly from respondents.
Interviewers must select a sample of respondents who are representative of the study’s target population. Interviews can be conducted one-on-one or in a focus group, but all participants must give their informed consent to be in the study.
Interviewers usually use a combination of closed-ended and open-ended questions to gain as much information as possible. Leading questions and double-barreled questions should be avoided.
For data analysis, interviews are transcribed so they can be examined for patterns of similarities and differences between responses.
Example: The Second Shift, by Arlie Russell Hochschild.
Advantages: Respondents can speak in their own words, revealing their own thoughts, feelings, and beliefs; dispels preconceptions and highlights issues that might have otherwise been overlooked.
Disadvantages: Respondents are not always forthcoming or truthful; limits in degree of representativeness.
Surveys are questionnaires administered to a sample of respondents selected from a target population. Samples are chosen through probability sampling; the most basic type is the simple random sample.
Question formats can be yes-or-no, based on a Likert scale, or open-ended. Leading questions, negative questions, double-barreled questions, and bias should be avoided. Questionnaire construction is also important. A pilot study can help refine a survey design before it is used with a larger group.
A survey is only valid if it has a sufficiently high response rate.
Responses are coded into numerical data for computerized statistical analysis, which helps establish relationships between certain variables.
Example: The State of Our Unions: The Social Health of Marriage in America, from the National Marriage Project at Rutgers University.
Advantages: Best method for gathering original data on large populations; quick; economical; reliable; less concern over bias.
Disadvantages: Doesn’t allow respondents a full range of expression; weak on validity; can be subject to sampling errors; can be manipulated to support a claim of point of view rather than for pure science.
2.5 Existing Sources
Existing sources are materials produced for some other reason that can be used as data for sociological research; considered unobtrusive measures.
Comparative historical research seeks to understand relationships between elements of society in various regions and time periods by analyzing cultural artifacts such as literature, paintings, newspapers, and photographs.
Content analysis is an approach that looks for recurrent themes or counts the number of times specific variables appear in a series of sources, then analyzes those variables and the relationships among them.
Example: Anxious Parents: A History of Modern Childrearing in America, by Peter Stearns.
Advantages: Researchers can work with information they could not obtain firsthand; allows for the comparison of different time periods and cultures; replicable.
Disadvantages: Limited to existing data; does not reveal how the messages were interpreted by the original consumers.
2.6 Experimental Methods
Experiments are formal tests of specific variables and effects that are performed in a setting in which all aspects of the situation can be controlled. An independent variable is made to differ between the experimental group and the control group to see what changes are reflected in the dependent variable. All other conditions are kept as consistent as possible between the two groups.
Advantages: Researchers can manipulate and control the social environment; best methods for establishing causality; replicable.
Disadvantages: Only applicable to certain types of research; examines an oversimplified version of the world.
2.7 Social Network Analysis
Social network analysis (SNA) is a tool for measuring and visualizing the structure of social relationships between two or more people. A questionnaire is used to map a respondent’s ties to different people in a given community, and the data can be used to study everything from disease transmission to corporate behavior.
SNA predates modern online social networks and was used in the 1960s by Stanley Milgram in his study showing the “six degrees of separation” phenomenon.
Advantages: Can trace the route of almost anything (especially useful for epidemiology); contributes to the production of “big data,” which is valuable for academia and corporations.
Disadvantages: Can gloss over important qualitative details and diversity; “big data” trends don’t necessarily hold up in smaller, more connected networks.
2.8 Issues in Sociological Research
Nonacademic uses of research methods
Opinion polls can be used to shape public opinion as well as to reflect it.
Market research is used to target specific populations for selling products and services.
Values, objectivity, and reactivity
Values: Most sociologists believe that research should be value free, meaning that personal beliefs should not influence research. This notion is challenged by Marxist sociologists, who believe that social research should influence social action, and by symbolic interactionists, who believe that researchers can never be completely value free. There is an ongoing debate over whether sociologists should engage in basic research or applied research.
Objectivity: The assumption is that researchers should be impartial in order to make accurate observations. However, all research is somewhat biased because each person’s experience and knowledge of the word is personal and idiosyncratic.
Reactivity: The ways people respond to being studied. Researchers may affect the results simply by being present. The term “Hawthorne effect” refers to research data in which results are skewed to the point of reflecting only the effect of being studied.
Researching human beings raises moral issues, including questions about the researcher’s responsibilities to the research subjects.
Each academic discipline has its own code of ethics. In addition, research universities have an institutional review board (IRB) that reviews the ethics of each research proposal.
2.1 An Overview of Research Methods