Quantitative surveys are usually comprised of standardized interviews that are conducted using a questionnaire. The term “standardization” describes the predetermination of the course of the interviews. In a fully standardized questionnaire each respondent is presented with the same stimulus, i.e., an equal question. Therefore, the reaction (i.e., the answer) is comparable to that of another respondent. The aim of a standardized interview is to avoid disturbing factors as far as possible. Usually, however, not all bias can be eliminated, which is why one should at least try to keep it constant. In standardized quantitative surveys, this means that the wording of all questions and response items is identical for all respondents. Furthermore, the sequence of the questions is specified exactly and the social situation should be constant in every interview.
Standardization aims at comparability of results as an important prerequisite for generalization and representativity of the whole survey. Quantitative surveys are able to produce results that are considered representative for the part of the population that participated in the interview. Besides using a specific sampling technique, this is achieved by a high degree of standardization and by complying with specific criteria regarding selection, training, and control of the interviewers. In media and communication research, quantitative interviews are a central survey instrument because representative results are often sought. Quantitative surveys are used for large samples and to make (representative) statements about the population as a whole or specific parts of it. This, for instance, may be the case in the field of political communication (e.g., election research) as well as in reception and effects research.
Mode
Several modes or methods of conduct are possible when implementing a quantitative interview. The main differentiation is the communication channel used. Quantitative interviews can be conducted orally, either in person (face-to-face interview) or via telephone. They can also be carried out in writing. All three forms may be supported by use of the computer. The interview situation will change significantly according to the use of the different modes.
Interview Situation
In a standardized survey, researchers collect socio-scientific data using a highly formalized instrument. The resulting communication situation has a functional character that is strictly defined beforehand. This leads to an artificial social situation in this type of interview that differs distinctly from other kinds of everyday communication. As a rule, strangers meet in an asymmetrical communication relationship – one person asking questions, the other one providing answers. Moreover, both the situation itself and the respondent’s answers are socially without consequences. Also, because of the required degree of standardization, the interviewer acts in a highly functional and formalized manner and, thus, quite unnaturally. The wording of the questions and response items is accurately defined, as is the question sequence, and the use of lists, card games (to support the answering process) and the like.
Interviewer And Respondent
Compared to qualitative interview forms, in a quantitative interview the interviewer has hardly any scope. This is necessary to make sure the standardization serves its purpose. To narrow the scope even more, researcher and interviewer are usually not one and the same person. Interviewers should scarcely know about the purpose of the survey so that they will not be able to influence the respondent (interviewer bias). Instead, they are restricted to the role of a professional interviewer who is able to motivate the respondents to participate and to handle the questionnaire technically well. Their purpose is to set the stimulus. The question then directly activates a multilevel cognitive process consisting of at least five steps (Schwarz & Oyserman 2001, 129). The interviewee has to (1) understand the question, (2) recall the relevant behavior, (3), make inferences and estimations concerning this behavior, (4) adapt his or her answer to fit the response format, and (5) edit the answer for reasons of social desirability. This process implies that “people know what they do and can report on their behavior with candor and accuracy, although they may not always be willing to do so” (Schwarz & Oyserman 2001, 129).
Steps Of A Standardized Survey
A quantitative survey is normally based on a particular research question and conducted according to a specific pattern. The question is divided into dimensions and those relevant for the study are selected. By using indicators, these dimensions are operationalized and transferred into the questionnaire. The final version of the questionnaire is subject to a pre-test and afterwards submitted to the selected sample. Subsequently, the data are transferred into a data processing program and statistically analyzed.
Each of these steps is prone to error and has to be conducted meticulously to guarantee methodological quality. The operationalization of the research question, i.e., the wording of the questionnaire, is a particularly important step. In a quantitative interview, formalized questions are asked to collect information. It is important that all respondents should understand these questions in the same way. Normally, though, different parts of the population have a different understanding of the same term. Also, different frames of reference are possible. Additionally, in most cases one cannot directly ask what the researcher is interested in but one has to use the indicators defined beforehand.
Therefore, particular attention has to be paid to the question phrasing in a quantitative survey. The research question has to be properly “translated.” First, this requires the predefinition of the indicators that have to be collected. Second, they have to be phrased in a language comprehensible to all respondents. In a good questionnaire, the researcher’s frame of reference – i.e., the dimensions of the question that are to be considered – is transferred into questions that are adequate to the respondents’ frame of reference. In methodological research, this aspect is the subject of many studies because the survey’s reliability and validity depend on the quality of the questions and the respondents’ understanding of them.
Unwanted Effects In Quantitative Surveys
When wording questions, unwanted effects may arise. These effects can be the result of response order and so-called “non-opinions.” An order effect occurs when respondents choose an answer because of a formal reason instead of a content-related one. For instance, they tend to prefer either the first (primacy effect) or the last (recency effect) item in a long list. The unwanted effect called nonor pseudo-opinion occurs when respondents tend to answer questions although they do not have (or cannot have) an opinion about the subject. This problem can only present itself when using closed questions, of course, because multiple-choice answers enable the respondents to venture a guess. When they are asked to give evaluations with the help of a scale, they often tick a medium position. If no middle position is given, new problems of pseudo-opinions may arise when respondents choose an answer at random.
Unwanted effects may also arise due to the sequence of the items because questions (and their answers) can influence the answering behavior in subsequent questions (context effects).
Types Of Questions
Survey questions have to be clear and concrete, explicit, and nonsuggestive. Generally, there are two types of questions: open-ended (also “free answer”) and closed ones. In open-ended questions, no alternative answers are offered; the question itself provides the frame of reference. Within this frame, respondents can express themselves freely. Such questions should be used sparingly in a standardized interview. They may be useful if the range of possible answers is too broad to be given (e.g., “year of birth”) or if the respondents’ exact choice of words is to be maintained, if the range of answers or spontaneity is important. One has to be aware that analyzing open questions in a standardized interview is very time-consuming. Also, open-ended questions are laborious for both respondents and interviewers, and the responses obtained are not always significant.
Therefore, mostly closed questions are used in standardized interviews. Unlike openended ones, all alternative answers are presented next to the question itself. Closed questions enable the researcher to compare respondents as the answers define the frame of reference. Thus, these categories guide the respondents more strongly and constrain them at the same time. Therefore they are regarded as more valid and reliable. Considering the cognitive process described above, the response items should be worded at least as meticulously as the questions. Respondents infer the respective significance from the questions and response items, and use them as a source of information on “usual” behavior (Schwarz & Hippler 1987).
There are various forms of closed questions. They may be simple alternative questions, multi-choice questions with a range of possible answers, ranking scales to measure order, or rating scales to measure intensity. The scale types can be classified as nominal, ordinal, or interval according to the data level they produce. Scales are used to measure attitudes, evaluations, and opinions but also activities. Using scaling techniques allows the measurement of non-observable states “within” a person, such as hypothetical constructs or emotions, preferences or evaluations. Rating scales may be presented as verbal scales, numeral scales, or visualized versions.
References:
- Fink, A. (ed.) (1995). The survey kit, vols. 1–9. Thousand Oaks, CA: Sage.
- Möhring, W., & Schlütz, D. (2001). Die Befragung in der Kommunikationswissenschaft [Standardized interviewing in communication research]. Wiesbaden: Westdeutscher.
- Schwarz, N., & Hippler, H.-J. (1987). What response scales may tell your respondents: informative functions on response alternatives. In H.-J. Hippler, N. Schwarz, & S. Sudman (eds.), Social information processing and survey methodology. New York: Springer, pp. 163–177.
- Schwarz, N., & Oyserman, D. (2001). Asking questions about behavior: Cognition, communication, and questionnaire construction. American Journal of Evaluation, 22, 127–160.
- Sudman, S., & Bradburn, N. M. (1983). Asking questions: A practical guide to questionnaire design. San Francisco, CA: Jossey-Bass.
- Sudman, S., Bradburn, N. M., & Schwarz, N. (1996). Thinking about answers: The application of cognitive processes to survey methodology. San Francisco, CA: Jossey-Bass.