In survey research, data collection instruments can be administered either as questionnaires in which there is no intervention between the data collection instrument and the subject or as an interview conducted by a human being. The intervention of a human interviewer between the survey instrument and the subject's responses has both advantages and disadvantages. In particular, interviewer bias and interviewer effects can impact the quality of the data obtained from the subject. However, interviewing techniques - in particular cognitive interviewing that employs probing or "think aloud" methodologies - can improve the quality of the data gathered in a survey. These methods also help survey designers develop better survey instruments and help researchers better understand and predict human behavior. When properly used, interviewing techniques can give researchers both a depth and breadth of information that cannot be gathered through other data collection techniques.
From a research point of view, interviews are part of the survey research methodology. In this type of research, data about the opinions, attitudes, or reactions of the members of a sample are gathered using a survey instrument. Surveys can be administered in one of two ways. The data collection instrument can be presented in the form of a questionnaire that the subject responds to without the intervention of another human being (e.g., researcher, interviewer). Surveys presented as questionnaires can be administered as paper-and-pencil instruments (e.g., through the mail or distributed by hand) or electronically (e.g., on a website or by e-mail). Surveys administered as interviews employ the same data collection instrument, but are administered by an interviewer asking questions of the subject either in person or over the telephone.
In interviews, the interviewer directs the conversation with the subject for the purpose of gathering specific information. Interviews can range from highly structured formats that use questions that are specifically worded and administered in a prescribed order from which the interviewer may not deviate, to very unstructured formats in which interviewers only follow a general form and are allowed great latitude in what specific data are collected or what follow-up questions are asked. As compared to questionnaire administration, surveys can result in a higher response rate because people have more difficulty turning down a person asking for a few minutes of their time than they do in throwing out a questionnaire. In addition, an interviewer can probe for further information whereas a questionnaire cannot. On the other hand, interviews are more expensive than questionnaires, particularly when one is attempting to collect data from a large sample.
Although survey research is used in social science research, it typically does not produce the same type of quantitative data as is generated by experimental research. Sometimes surveys are designed so that the results can be expressed in numerical form and, therefore, can potentially be analyzed using inferential statistics. However, even for those surveys that are designed so that the subjects' responses can be quantified (e.g., "on a scale from 1 to 10, rate what you think about X"), the results often violate the assumptions (e.g., true zero, interval, or ratio scale) of many of the inferential statistical tools that would be used to analyze them. In general, survey research belongs to the realm of qualitative research in which observations cannot be or are not quantified (i.e., expressed in numerical form).
This does not mean, however, that surveys cannot be used as part of the scientific method. The questions investigated in behavioral and social science research are often extremely complex, and survey research methodology can be employed as part of the inductive reasoning process to better understand the importance of and relationship between variables observed in the real world. The results of this process can often later be used to experimentally test scientific theories as part of the scientific method. Further, there are often ethical considerations that prevent social and behavioral scientists from intentionally manipulating variables and collecting data that can be analyzed using inferential statistics (e.g., to better understand the relationship between personal support systems and recovery after the death of a loved one, it would be both unethical and illegal to experimentally manipulate whether or not the loved ones of the subjects lived or died). Survey research can be used to ethically gather information that could not otherwise be obtained. Also, as shown in Figure 1, it can be used with the paradigms of the scientific method and theory building process to help the researcher better understand observed behavior.
Like any research methodology, interviews have both advantages and disadvantages and are subject to various pitfalls. As opposed to a survey instrument that is administered as a questionnaire which, therefore, presents questions to all subjects in a standardized format, the interview paradigm introduces an additional extraneous variable that can negatively or positively influence the results: the interviewer. This influence can come from a variety of sources. Interviewer bias occurs when the individual administering the interview has certain expectations, beliefs, prejudices, or other attitudes that may affect the interview process and the subsequent interpretation of data collected through the interview process. If, for example, an interviewer thinks that women are unable to give thoughtful, long responses to interview questions, he or she may ask the questions of women in such a way that they are unlikely to give long responses but ask questions of men in a different way that encourages them to do the opposite. The possibility of interviewer bias is of particular concern when using an unstructured interview. In such interviews, questions are open-ended rather than forced choice in order to allow the interviewer to probe for more information or to allow the subject to think aloud. Although the purpose of the unstructured interview is to gather additional information and higher quality data, a biased or untrained interviewer can effectively produce the opposite effect.
Specific biases on the part of the interviewer are not the only things that can impact the quality of the data engendered in an interview. Interviewer effects are the influence of the interviewer's behaviors and attributes on the subject's response in an interview situation. For example, the appearance, demeanor, training, age, gender, and ethnicity of the interviewer may all affect the way that a subject perceives the interview or responds to the questions during an interview. In some cases, the subject may try to please the interviewer by giving responses that he or she thinks the interviewer may want to hear or in other cases may give non-responsive answers in order to negatively impact the value of the data collected by an interviewer that he or she does not like. Research has found, for example, that female subjects give more feminist responses to female interviewers than they do to male interviewers. Likewise, African Americans tend to give more detailed responses concerning race-related issues to African American interviewers than to those of other races.
The way that the interviewer interacts with the subject can have an impact on a structured interview, as well. A surly or condescending interviewer can easily create a hostile environment in which the subject is unlikely to give additional information or, sometimes, even to answer basic survey questions honestly. However, the same can be true for an interviewer who is too friendly. Although friendliness can lead to a situation in which the interviewer can gather more valuable information from the subject, it can also lead to a situation in which the subject attempts to give information that he or she thinks will please the interviewer rather than giving responses that truly represent his or her real opinions. In many instances, this problem can be overcome through interviewer training in which interviewers learn how to ask questions in a neutral way, effectively probe for additional information, and not let their own beliefs or opinions bias the interview results.
Because of these potential problems with collecting data via interviews, it is important to take the characteristics of both the interviewer and the subjects into account when designing a research...
(The entire section is 3809 words.)