OrientationObservationIn-depth interviewsDocument analysis and semiologyConversation and discourse analysisSecondary Data
SurveysExperimentsEthicsResearch outcomes
Conclusion
8.3.10.2 Interviewing Interviews are a far more personal form of research than questionnaires. In the personal interview, the interviewer works directly with the respondent. Unlike mail surveys, the interviewer has the opportunity to probe or ask follow-up questions. Interviews are generally easier for the respondent, especially if what is sought is opinions or impressions.
Interviews can be very time consuming and they are resource intensive for the researcher.
The interviewer is considered a part of the measurement instrument and interviewers have to be well trained in how to respond to any contingency.
The interviewer's job is to locate the respondent, obtain an interview, ask the questions and record the answers as accurately as possible. This is not always as easy as it seems.
8.3.10.2.1 Locating the respondent
Interviewers need to be given clear instructions about who to ask, how many times they should try to locate particular respondents and who, if anybody, they may substitute for missing respondents.
A set of interviewer instructions ensures that the same procedures are used throughout the interview stage (see, for example, the interviewer instructions in CASE STUDY Sandwell Skills Audit).
The exact nature of these instructions will depend on the nature of the sample and the adequacy of the sampling frame. If, for example, the sample consists of names drawn from the electoral register, the researcher must decide whether to allow the interviewer to substitute new occupants of houses for people in the sample who have moved out of the area.
In some cases there is no formal identified sample and the researcher has to use initiative to gather a sample. MacKellar et al. (2002) compared recent risk behaviors and HIV seroconversion among young men who have sex with men. There would have been no sampling frame and people were approached in locations where the type of respondent was likely to be found. The research reported a sample of 3430 male adolescents and young men aged 15 to 22 years were selected at random from 194 gay-identified venues in seven U.S. cities between 1994 and 1998. It isn't clear how the 'random selection' was made and this study has an unusually lengthy data collection period.
8.3.10.2.2 Obtaining an interview
Having located a potential respondent, the interviewer has to persuade that person to answer the questions. It helps if the interviewer adopts a pleasant demeanour and tone and if the interview seems to be about something relevant and interesting to the respondent.
If respondents find the topic threatening or are suspicious or are unclear about the researcher's motives or unclear about the organisation that the interviewer represents then an interview is unlikely to result.
The interviewer should assure respondents that the information they provide as individuals is confidential, anonymous and will not be passed on to anyone else. Only aggregated data will be reported.
For example, in the CASE STUDY Sandwell Skills Audit, which included questions on employment, the interviewers had to make it clear that they were from Sandwell College and not from the Department of Employment nor, given that the (now defunct) 'poll tax' had just been announced, that they were not from the local authority. Further, assurances were given that no personal information would be passed on to either of these organisations.
The interviewer's opening statement when approaching someone for an interview is therefore very important. It should summarise these points very briefly and seek to reassure respondents and gain their co-operation (see the interviewer's introduction in the CASE STUDY Sandwell Skills Audit).
8.3.10.2.3 Asking the questions
The interview should resemble as closely as possible a structured conversation in which the interviewer leads the dialogue. Interviewers need to be familiar with the interview schedule, both the content and sequence of the questions, so that the interview flows without artificial breaks caused by the interviewer stumbling over questions or not being sure which question to ask next.
A well-designed schedule that clearly indicates which question to ask next and that has clear instructions to the interviewer makes the process of asking the questions much easier than one that is poorly designed.
The interviewer should attempt to establish 'rapport' so that the respondent feels at ease in answering the questions. Conventionally, it is argued that the best way of doing this is for the interviewer to talk to people in their homes with the respondent seated facing the interviewer (so that the respondent cannot read the schedule) (see for example, Atkinson, 1967; Moser and Kalton, 1971).
However, it is not always possible or desirable to conduct interviews in the respondent's home.
For security reasons, when conducting house-to-house interviews interviewers should ensure that someone else knows where they are going. Interviewers may also prefer to work in pairs rather than alone in some circumstances. This might involve two people doing the interview together (although the respondent might feel threatened by this) or it might mean two people working together in the same street, keeping an eye on the movements of each other.
Never be afraid of terminating an interview and leaving if the situation becomes uncomfortable or threatening. This also applies to situations where respondent are clearly wasting the interviewer's time by not taking the interview seriously.
8.3.10.2.4 Telephone interviews Some interviews are conducted remotely via telephone or Internet. Almost everyone is familiar with the telephone interview, if only as part of unwanted sales techniques!.
Telephone interviews enable a researcher to gather information rapidly. Most of the major public opinion polls that are reported were based on telephone interviews.
Telephone interviews involve some personal contact between the interviewer and the respondent, with consequent opportunity to ask-follow up questions or seek clarification and, therefore, differ from impersonal questionnaires.
There are disadvantages in undertaking telephone interviews.
First, many people do not have publicly-listed telephone numbers, which restricts the potential sample (furthermore, some landlines are little used or have been given up in favour of a personal mobile telephone).
Second, some people object to strangers calling them unannounced. Overcoming this requires pre-contact in some other form, be it letter, email or SMS text.
Third, it is sometimes hard to guage the reaction of the interviewee as in conventional telephone conversation there are only aural and not visual cues.
Fourth, telephone interviews normally have to be relatively short because people will hang up if they have had enough.
In their study of non-standard work practices Olsen & Kalleberg (2004)undertook telephone surveys of managers in 802 establishments in the United States and 2130 establishments in Norway. The study showed that establishments in the public sector are more likely to use direct-hired temporary workers and less apt to use contractors; this pattern is particularly striking in Norway but is also evident in the United States. Furthermore, highly unionised establishments tend to have the lowest use of non-standard arrangements in both countries. [See Section 8.3.12 for details of Olsen and Kalleberg's analysis]
8.3.10.2.5 Asynchronous interviews
Information technology now makes it possible to conduct interviews asynchronously, using such things as email, personal telephone messaging and social media.
Asynchronous communication in this sense is the exchange of messages that doesn't take place in real time with responses immediately following each other as in a normal conversation. Instead, the interviewer sends the questions one at a time or in batches and respondents reply when it is convenient and safe for them to do so.
The advantage of this is that the respondent is in control of the exchange and also has time to reflect on the question before responding. Asynchronous interviewing allows for clarification on items and makes it easy to do follow-up interviews. Asynchronous interviews are good when the researcher is probing an issue in depth or needs the respondent to provide information they may not have to hand.
The disadvantages are that the interviewer is not in control of the interview situation; the respondent can easily terminate the exchange (without any notice) and the interviewer may be waiting in vain for a response. The respondent also has time to contrive answers that may put the respondent in a better light rather than the instant raw response that the interviewer might be seeking. If the interviewer wants a 'first impression' response rather than a considered one, then asynchronous communication is probably not the best method.
8.3.10.2.6 Recording the answers
The point of the interview is to get the respondent's answers. The interviewer must try to record these as accurately as possible and should seek clarification of ambiguous answers rather than simply presuppose or guess what they mean.
Similarly, the interviewer should avoid leading the respondent by suggesting answers. The interviewer must be honest and under no circumstances write in answers that were not provided by the respondent.
8.3.10.2.7 Interview bias
Irrespective of the expertise of the interviewer there is still an issue of interview bias. An artificial situation such as an interview will have an impact on the information respondents are prepared to provide about their attitudes and opinions (Oakley, 1981).
Despite attempts at rapport, interviews will result in selective information, the part that the respondent is prepared to make available and that she or he thinks will be of interest to the researcher. This is known as interview effect.
Interviewing on a large scale that involves employing interviewers, who may be less committed to the enquiry than the researcher, leads to problems called the hired-hand effect. This includes the accidental mis-checking of an item on a schedule; the failure of adequately recording open answers; deliberately questioning the wrong people out of convenience; and even outright fraudulent completion of schedules by the interviewer without interviewing anyone.
Hired-hand effect, especially the minor accidental mistakes, is often difficult to spot.
Deming (1944) noted that interviewer bias was a well-established source of error in surveys. He looked at a study by Stuart Rice that showed the response of an interviewee can be highly influenced by the interviewer. Two thousand 'destitute men' were interviewed and their destitution was attributed to the 'evils of alcohol' by those interviewed by a teetotaller and to 'industrialisation' by those interviewed by a socialist. Deming suggests that the educational, economic status, environmental background, political, religious and social beliefs of interviewers will all affect the interview situation and thus the results.
A further early study in America (Williams, 1964) showed that, in a survey where all the respondents were black and interviewers were black or white, the largest potential source of bias occurred when questions were 'threatening', that is, they challenged accepted social norms. In particular, middle-class white interviewers obtained a significantly higher percentage of conservative responses from lower-class blacks than did middle-class black interviewers on these highly threatening questions.
.
