Research, Survey
Research, Survey
ADMINISTRATION AND INTERVIEWING
Survey research is a methodological process by which social scientists convert theoretical concepts into numbers. The quantification of theoretical concepts via the survey instrument allows for the modeling of relationships between variables such as education and socioeconomic status. Surveys are used by researchers across all of the social sciences as one of the primary means of data collection. For example, in order to measure a theoretical concept such as religiosity, the social scientist can use an indicator or set of indicators of religiosity, and the aggregate responses can then be coded numerically. A common example of such an indicator is the question “How important is religion in your life?” The subjects being surveyed would typically select one of the following responses: “Very Important,” “Pretty Important,” or “Not Too Important.” The three response categories are then coded as follows: 1= “Very Important,” 2= “Pretty Important,” and 3= “Not Too Important.” In this way, social scientists are able to quantify theoretical concepts.
EPISTEMOLOGICAL FOUNDATIONS
The philosophical underpinnings of survey research rest squarely on the side of positivism, as opposed to interpretivism. Positivist social scientists seek to emulate the level of quantification that is characteristic of the natural sciences, particularly physics. The aim of survey research, from a purely positivist viewpoint, is to discover the universal laws operating in society. It is thought that such laws are best uncovered through a deductive, scientific method, whereby data is collected through a survey instrument in order to test a theory. Interpretive theorists, with a more qualitative focus on issues such as meaning and historicity, are critical of positivist assumptions and tend to prefer a more inductive, theory-building approach to data collection.
QUESTIONNAIRE CONSTRUCTION
The successful execution of survey research depends in great measure on the proper construction of the questionnaire. Good questionnaires provide measures of theoretical concepts, also referred to as variables, that are concomitantly reliable (i.e., able to be reproduced) and valid (i.e., accurate). The questionnaire should be clear and concise in presentation and wording. The researcher can choose between single items or multiple items in order to measure a variable. For example, another question that measures religiosity might be: “How often do you attend religious services?” Possible response categories could include: 1= “Often,” 2= “Not Too Often,” and 3= “Never.” This is an example of a closed-ended question, in which response categories are predetermined by the researcher. Such categories should be exhaustive (i.e., they should cover all possible responses) and mutually exclusive (i.e., there should be no overlap among the responses).
One of the most common sets of response categories is the Likert scale, which measures how strongly a person agrees or disagrees with a statement. One way to measure happiness, for example, is with a four-point Likert scale in which the response categories range from “Strongly Agree” to “Strongly Disagree.” Consider the following indicator of happiness: “On the whole, I am happy with my life.” A standard Likert scale would code the responses as: 1= “Strongly Agree,” 2= “Agree,” 3= “Disagree,” and 4= “Strongly Disagree.”
When multiple items are used to measure a single concept like happiness, it is sometimes useful to invert some items, such as: “On the whole, I am unhappy with my life.” The inversion of items can disrupt potential patterns in which the respondent “Strongly Agrees” or “Agrees” with all of the items.
In addition to closed-ended questions, which tend to be comparatively high on reliability, survey instruments can also be constructed with open-ended questions, which tend to be comparatively high on validity. Open-ended items probe the respondent’s knowledge on a given issue, but they do not offer a predetermined set of response categories. For example, a social scientist interested in studying the nature and determinants of trust might include the following item in a questionnaire: “What are the most important factors in your decision to extend trust?” The obvious advantage of such open-ended items is that the researcher does not need to be concerned with the development of exhaustive and mutually exclusive response categories. The major disadvantage is that open-ended responses, as a result of their variability, are difficult to code, leading to reduced reliability.
ADMINISTRATION AND INTERVIEWING
The three most common methods of administering questionnaires are by mail, by telephone, and in face-to-face interviews. Since the 1990s, there has also been an interest in the use of the Internet to administer electronic surveys. The principal advantage of administration via post and the Internet is the comparatively low cost of the research vis-à-vis telephone and personal interviews. The disadvantages of survey research conducted via post and the Internet are the comparatively lower response rates and the overall length of the research process.
Telephone interviews have the obvious advantage of being cheaper to conduct than face-to-face interviews. In terms of the length of the research process, a second clear advantage is that questionnaires can be administered and data collected more quickly than with face-to-face interviews. The third advantage of telephone interviews is that a random sample can be readily selected using random digit dialing (RDD), whereby a computer generates random telephone numbers. However, the fact that not every household has a telephone can bias the sample. Furthermore, telephone interviews are generally not suitable for long questionnaires.
The consensus is that face-to-face interviews yield the highest quality data. In general, there are three main advantages of face-to-face interviews. First, the instrument can be longer than if it were administered by telephone, post, or the Internet. Second, interviewers can use visual aids to assist respondents. Third, it is easier for interviewers to clarify questions and items that are unclear, although it is possible that clarification on the part of the interviewer can bias the results. The main disadvantage of personal interviews is the high cost involved in the research.
STRENGTHS AND WEAKNESSES
In short, survey research has several strengths vis-à-vis the more qualitative methodological approaches. First, the results of survey research are, on balance, more generalizable than the results of qualitative research. In other words, because of the greater number of respondents, survey research typically yields findings that are more representative of the population being studied. Second, survey instruments yield more reliable results than qualitative interviews. Third, questionnaires are useful in terms of both testing theories and establishing correlations between variables.
There are four main weaknesses of survey research. First, it can be potentially very expensive to conduct. Second, it is rigid and inflexible, especially in the absence of open-ended items. Third, it is open to the criticism of having a top-down bias; that is, it seldom allows for theory construction from the bottom-up. Fourth, although it is strong in terms of establishing statistical correlations, it is weak in terms of proving causality between two variables. In spite of these weaknesses, survey research is one of the most widely used methodological approaches.
SEE ALSO Methods, Research; Surveys, Sample
BIBLIOGRAPHY
Babbie, Earl. 1999. The Basics of Social Research. Belmont, CA: Wadsworth Publishing.
Dillman, Don A. 2000. Mail and Internet Surveys: The Tailored Design Method. 2nd ed. New York: Wiley.
Judd, Charles M., Eliot R. Smith, and Louise H. Kidder. 1991. Research Methods in Social Relations. 6th ed. Fort Worth, TX: Holt, Rinehart, and Winston.
Andrew R. Timming