Formulating the Hypothesis

After reviewing earlier research and drawing on the contributions of sociological theorists, the researchers may then formulate the hypothesis. A hypothesis is a speculative statement about the relationship between two or more factors known as variables. Income, religion, occupation, and gender can all serve as variables in a study. We can define a variable as a measurable trait or characteristic that is subject to change under different conditions.

Researchers who formulate a hypothesis generally must suggest how one aspect of human behavior influences or affects another. The variable hypothesized to cause or influence another called the independent variable. The second variable is termed the dependent variable because its action depends on the influence of the independent variable.

Our hypothesis is that the higher one's educational degree, the more money one will earn. The independent variable that is to be measured is the level of education. The variable that is thought to depend on it – income- must also be measured.

Identifying independent and dependent variables is a critical step in clarifying cause-and-effect relationships. Causal logic involves the relationship between a condition or variable and a particular consequence, with one event leading to the other. For instance, being less integrated into society may be directly related to, or produce a greater likelihood of, suicide. Similarly, the time students spend reviewing material for a quiz may be directly related to, or produce a greater likelihood of, getting a high score on the quiz.

A correlation exists when a change in one variable coincides with a change in the other. Correlations are an indication that causality may be present; they do not necessarily indicate causation. For example, data indicate that working mothers are more likely to have delinquent children than mothers who do not work outside the home. But this correlation is actually caused by a third variable: family income. Lower-class households are more likely to have a full-time working mother; at the same time, reported rates of delinquency are higher in this class than at other economic levels. Consequently, while having a mother who works outside the home is correlated with delinquency, it does not cause delinquency. Sociologists seek to identify the causal link between variables; this causal link is generally described in the hypothesis. 

Collecting and Analyzing Data

How do you test a hypothesis to determine if it is supported or refuted? You need to collect information, using one of the research designs described later in the chapter. The research design guides the researcher in collecting and analyzing data.

Selecting the Sample. In most studies, social scientists must carefully select what is known as a sample. A sample is a selection from a larger population that is statistically representative of that population. There are many kinds of samples, but the one social scientists use most frequently is the random sample. In a random sample, every member of an entire population being studied has the same chance of being selected. Thus, if researchers want to examine the opinions of people listed in a city directory, they might use a computer to randomly select names from the directory. The results would constitute a random sample. The advantage of using specialized sampling techniques is that sociologists do not need to question everyone in a population. It is all too easy to confuse the careful scientific techniques used in representative sampling with the many nonscientific polls that receive much more media attention. For example, television viewers and radio listeners are often encouraged to e-mail their views on headline news or political contests. Such polls reflect nothing more than the views of those who happened to see the television program (or hear the radio broadcast) and took the time, perhaps at some cost, to register their opinions. These data do not necessarily reflect (and indeed may distort) the views of the broader population. Not everyone has access to a television or radio, time to watch or listen to a program, or the means and/or inclination to send e-mail. Similar problems are raised by the "mail-back" questionnaires found in many magazines and by "mall intercepts," in which shoppers are asked about some issue. Even when these techniques include answers from tens of thousands of people, they will be far less accurate than a carefully selected representative sample of 1,500 respondents.

Ensuring Validity and Reliability . The scientific method requires that research results be both valid and reliable. Validity refers to the degree to which a measure or scale truly reflects the phenomenon under study. A valid measure of income depends on the gathering of accurate data. Various studies show that people are reasonably accurate in reporting how much money they earned in the most recent year. Reliability refers to the extent to which a measure produces consistent results. Some people may not disclose accurate information, but most do. In the General Social Survey, only 5 percent of the respondents refused to give their income or indicated they did not know what their income was. That means 95 percent of the respondents gave their income, which we can assume is reasonably accurate (given their other responses about occupation and years in the labor force).

Developing the Conclusion

Scientific studies, including those conducted by sociologists, do not aim to answer all the questions that can be raised about a particular subject. Therefore, the conclusion of a research study represents both an end and a beginning. It terminates a specific phase of the investigation but should also generate ideas for future study.

Major Research Designs

An important aspect of sociological research is deciding how to collect the data. A research design is a detailed plan or method for obtaining data scientifically. Selection of a research design is often based on the theories and hypotheses the researcher starts with (Merton 1948). The choice requires creativity and ingenuity, because it directly influences both the cost of the project and the amount of time needed to collect the data. Research designs that sociologists regularly use to generate data include surveys, observation, experiments, and existing sources.

Surveys

Almost all of us have responded to surveys of one kind or another. We may have been asked what kind of detergent we use, which presidential candidate we intend to vote for, or what our favorite television program is. A survey is a study, generally in the form of an interview or questionnaire, that provides researchers with information about how people think and act. Among the United States' best-known surveys of opinion are the Galiup poll and the Harris poll. As anyone who watches the news during presidential campaigns knows, these polls have become a staple of political life.

When you think of surveys, you may recall seeing many person-on-the-street interviews on local television news shows. Although such interviews can be highly entertaining, they are not necessarily an accurate indication of public opinion. First, they reflect the opinions of only those people who happen to be

at a certain location. Such a sample can be biased in favor of commuters, middle-class shoppers, or factory workers, depending on which street or area the newspeople select. Second, television interviews tend to attract outgoing people who are willing to appear on the air, while they frighten away others who may feel intimidated by a camera. As we've seen, a survey must be based on precise, representative sampling if it is to genuinely reflect a broad range of the population.

In preparing to conduct a survey, sociologists must not only develop representative samples; they must exercise great care in the wording of questions. An effective survey question must be simple and clear enough for people to understand. It must also be specific enough so that there are no problems in interpreting the results. Open-ended questions ("What do you think of the programming on educational television?") must be carefully phrased to solicit the type of information desired. Surveys can be indispensable sources of information, but only if the sampling is done properly and the questions are worded accurately and without bias.

There are two main forms of the survey: the interview, which a researcher obtains information through face-to-face telephone questioning, and the questionnaire, in which the researcher uses a printed or written form to obtain information from a respondent. Each of these has its own advantages. An interviewer can obtain a higher response rate because people find it more difficult to turn down a personal request for an interview than to throw away a written questionnaire. In addition, a skillful interviewer can go beyond written questions and probe for a subject's underlying feelings and reasons. On the other hand, questionnaires have the advantage of being cheaper, especially in large samples.

Studies have shown that the characteristics of the interviewer have an impact on survey data. For example, women interviewers tend to receive more feminist responses from female subjects than do male researchers, and African American interviewers tend to receive more detailed responses about race-related issues from Black subjects than do White interviewers. The possible impact of gender and race indicates again how much care social research requires (D. W. Davis 1997; Huddy et.al. 1997).

The survey is an example of quantitative research, which collects and reports data primarily in numerical form. Most of the survey research discussed so far in this book has been quantitative. While this type of research can make use of large samples, it can't offer great depth and detail on a topic. That is why researchers also make use of qualitative research, which relies on what is seen in field and naturalistic settings, and often focuses on small groups and communities rather than on large groups or whole nations. The most common form of qualitative research is observation, which we consider next.