Geoforum. Vol. 24, No. 3. pp.

339-355, 1993 Printed in Great Britain 0

oolfF7185/93 $h.00+0.00 1993 Pergamon Press Lid

Interviewing Business Owners and Managers: a Review of Methods and Techniques

MICHAEL

J. HEALEY,* Coventry, U.K., and MICHAEL RAWLINSON,? Cardiff, U.K.

B.

Abstract: Interviews with business owners and managers are a prime source of information for economic geographers, yet very little has been written about this technique. Although the issues involved in business surveys are not unique, several points about research methods are highlighted when undertaking business interviews, including the link between interview techniques and research design, identifying respondents and obtaining access, preparing for interviews and obtaining accurate answers, and ways of conducting and recording interviews. This paper draws upon both published sources and a survey of the interview methods used by Anglo-American economic geographers in order to review the methods and techniques of interviewing. This paper concludes by arguing that there is no one best way of interviewing business owners and managers. Methods vary for different situations, depending on a range of factors, including the research design, the kind and amount of information required, the resources available, and the size, organizational structure, sector and location of the businesses to be approached. There is an urgent need for a greater priority to be given to experimenting with different survey methods and reporting the results.

Business Interviews

and Geographical

Research

Business organizations have been an important focus for geographical research for a long time. In the late 1950s and 1960s a combination of dissatisfaction with the treatment of the firm as a black box in neoclassical location theory and the increasing concentration of economic power in large businesses led to the emergence of the behavioural approach and an analysis of the geography of enterprise (KRUMME, 1969; McNEE, 1960). To begin with most attention was given to large industrial enterprises [e.g. WATTS

*Division of Geography, mental Sciences, Coventry U.K. tCardiff Business School, Cardiff, Cardiff CFl 3EU,

School of Natural and EnvironUniversity, Coventry CV15FB, University U.K. of Wales College of

(1980)]. Although later interest continued in the role of multiplant and multinational organizations in influencing spatial patterns of human activity [e.g. HEALEY and WATTS (1987) and TAYLOR and THRIFT (1982)], attention also widened to include small enterprises [e.g. KEEBLE and WEVER (1986) and MASON (1989)] and the relationships between small and large businesses [e.g. TAYLOR and THRIFT(1983) and RAWLINSON (1991)]. Furthermore, interest spread from a concern with enterprises primarily involved in manufacturing to those engaged in other sectors, such as producer services [e.g. THRIFT et al. (1988)], distribution [e.g. McKINNON (1985)], and agriculture [e.g. ILBERY (1987)]. Interviews with owners and managers are a prime source of information on the activities of businesses, 339

340 whether behaviour the purpose is to understand the spatial or as a process contain

Geoforum/Volume

24 Number

3/1993

of the organizations

themselves

of undertaking research in organizations also helpful insights [e.g. BRYMAN (1988, discussions of surveying in the industrial market HAGUE (1987a), HART BLEASDALE (1987), (1972), RAWNSLEY

means of analysing the economic geography of a particular sector or area. It needs to be recognized, however, managers material that interviews with business owners and will provide the research worker with from only one particular viewpoint and it

1989)]. More general businesses are contained research literature [e.g. (1987), JOBBER and MACFARLANE-SMITH (1978) and WILSON

will often be desirable to obtain other viewpoints, such as those of the unions and various government bodies. demic Where detachment conflicting evidence collected. different and owners emerges in assessing methods managers, acaand is of will be required

(1968)], but most of this work is,

not surprisingly, concerned with market assessment and industrial buying, and not all of this experience may types be directly of business transferable survey. 1984; 1992; to undertaking With 1983. other 199la; 1991. a few exceptions

interpreting concerned interviewing

the information only with reviewing business

This paper

(BACHTLER. MCDOWELL,

HEALEY,

SCHOENBERGER.

although many of the points made here may be of relevance to gathering information from representatives of a wide range of organizations, and several issues overlap with those discussed in social surveys. Business interviews range from asking the same list of questions in an identical form to all respondents to discussions of a series of topics in which the questions asked vary from respondent to respondent, depending on the answers previously given and the nature of the interaction process in the interview. The choice of standardized or non-standardized interviews is linked to the research design and theoretical framework adopted. However, the relationship is not exact and many combine different methods of interviewing as appropriate within the same research design. Interviews may take place over the telephone or, more frequently, face-to-face. Many books have been written on survey research methods [e.g. DIXON and LEACH (1978), GARDNER (1978) and HOINVILLE et al. (1978)] and the methodology of qualitative research [e.g. EYLES and SMITH (1988) and SILVERMAN (1985)]. Some focus specifically on interviewing [e.g. BRIGGS (1988) and GORDEN (1975)], but most are concerned with social surveys rather than investigations of businesses. Although the general advice in these texts is relevant to the economic geographer. for the most part the books lack examples of business interviews and discussion of the specific problems involved. Some useful hints arc contained in the literature on elite interviewing [e.g. DEXTER (1970) and MOYSER and WAGSTAFFE (1987)], but, with the exception of KINCAID and BRIGHT (1957a. b), little involves economic elites. Discussions of the

1992), discussions by economic geographers of interviewing business managers are restricted to the methodology sections of research theses and monographs, and brief comments on business surveys in a few text books [e.g. ILBERY (1985a)). This paper attempts to bring together previously scattered comments on interviewing business owners and managers with a coherent framework. Relatively few issues of research method are unique to studies of businesses (BULMER, 1988). Several issues are, however. sharply highlighted in business surveys. Business organizations arc bounded institutions to which the research worker has to seek, negotiate and gain access. Once entry has been secured a workable relationship needs to be established with the representative(s) of the business in which to gather information by interview. Various ethical issues, concerned with explaining adequately the purpose of the investigation, the confidentiality of the responses, and the rights of respondents to comment on what is written about them, also arise in business surveys. The aim of this paper is to discuss these and other issues by reviewing the literature on interviewing business owners and managers, and illustrating the points raised with examples of the methods used by economic geographers. It builds on an earlier broader survey of methods of obtaining information from businesses by one of the authors (HEALEY, 199la). To supplement the limited number of published examples, 20 colleagues in the U.K., North America and Australasia were invited to give us illustrations of their experiences of interviewing business owners and managers. These contacts provided many useful

GeoforumNolume

24 Number

3/1993

341 Le Heron (personal communication, 1991) emphasizes that, in two recent projects, one on the restructuring of farm businesses in New Zealand and the other on the internationalization of large Australian and New Zealand companies in the 198Os, as much effort went into establishing the context of state policy and investment conditions as into interviewing. Interviews are thus rarely used as the only source of information and are employed to complement other data sources, such as company accounts and aggregate In many industry situations, statistics. however, the information and

examples and emphasised the need to illuminate the way interviews are undertaken so as to highlight successful practices and to improve the standard of interviewing. The debate on the nature of the links between the methods used on the ground and the underlying philosophy developed elsewhere 1985 ; MCDOWELL, SCHOENBERGER, sis is on interviewing gathering information or epistemological (MASSEY and 1992; base is well MEEGAN,

1992; 1992). In this paper the emphaas a method or technique of once it has been chosen by a

SAYER,

research worker. Discussion of how the data, once collected, is analysed, though important, is outside the remit of this paper. This review of successful practices should be useful as a guide to research workers unfamiliar with undertaking interviews of business owners and managers, and should also help experienced researchers to reflect on the strengths and weaknesses of a widely used technique. It is important to appreciate the problems and potentials of interviewing if the findings of research using the method are to be understood fully. Although this paper is illustrated primarily by the experiences of economic geographers, many of the points made should be helpful to other social scientists planning to interview business owners and managers. Following a brief discussion of the situations when interviews are necessary, the relationship between interview techniques and research design is discussed. Methods of identifying respondents and obtaining interviews are then examined. This is followed by sections, first, on preparing for interviews and obtaining accurate answers and, second, on conducting and recording interviews. This paper concludes by arguing that there is an urgent need for a greater priority to be given to experimenting with different survey methods and reporting the results. When to Interview Obtaining information direct from business owners and managers is a time-consuming and labourintensive method of collecting information, and it is therefore usually sensible first to examine other sources. Much general information exists, at least about large businesses and the industries in which firms operate (although the quality of the former is sometimes debatable), and this should be examined before interviews are undertaken. For example, R.

insights that economic geographers only be obtained direct from people being examined. obtaining the A number information;

are seeking may in the businesses

of options remain for namely, a selfadministered questionnaire, a telephone interview, or a face-to-face interview. Self-administered questionnaires are most appropriate where there are a few simple factual questions, many of which may be answered by the respondents ticking the relevant boxes. HAGUE (1987a) suggests that postal questionnaires are most suitable for surveying businesses run by one key decision maker. One of the advantages of the technique is that it ensures consistency in the questions asked and avoids variability between interviewers. The written questionnaire also means that respondents can answer the questions at their own pace, which allows for potentially more considered replies. However, in the absence of a skilled interviewer, there is no control over the order in which the questions are completed, some questions may be missed by the respondent and there is no opportunity to probe vague answers (DE CHERNATONY, 1988; SWAIN, 1978). Interviews are generally essential when there are a large number of questions, or the subject of the survey involves an investigation into the reasons for decisions or the perception of the owner or manager. The main advantages of an interview are that the interviewer not only has more control over who answers the questions than can be achieved with a postal survey, but can also clarify any ambiguous questions, probe answers that are too brief, and query discrepancies in the replies. Standardized interviews are often used as an alternative to selfadministered questionnaires sent through the post [e.g. HEALEY (1984)], because, though more expensive in terms of time and cost, they usually

342 achieve a higher response and regional rate, One review of 43 local activity in the U.K. and quality

Geoforum/Volume of the survey methods

24 Number

3/1993

used and often

surveys of economic

found that, although response rates varied widely, the average response rate for postal surveys was 51%. while for face-to-face (HEALEY, 1991a). Telephone economic interviewing research. interview surveys it was 75%

reflect the underlying research design adopted. Interview techniques may be categorized into standardized and non-standardized. The main characteristics of these two types of interview are summarized in Table 1. With standardized interviews, each business owner or manager in the survey is asked an identical set of questions in a fixed order. suitable for collecting information, They are most which is largely

is growing

in popularity cheap method,

in

It is a relatively

especially if restricted to a local area, and in many cases an immediate answer to the survey questions can be obtained. However, it is harder trusting relationship over the telephone face-to-face interview, to build up a than it is in a for the

factual or non-emotive, from businesses with several features in common. They are particularly appropriate when the aim is to quantify the relative importance of different responses to questions about a set of well-defined topics. In contrast, a non-standardized interview is much less structured and the questions asked and the phrasing of the questions vary from interview to interview. At its most informal the nonstandardized interview can be a non-directive convcrsation. but in economic geography interviews of this kind are almost exclusively focused or guided to cover a predetermined set of topics. Non-standardized interviews are most helpful when exploring new topics, sensitive or emotive issues, and when the businesses are highly variable in their characteristics. They are particularly suitable for detailed examination of topics and when the nature of the experience of respondents is likely to vary widely, such as when examining the sometimes conflicting logic that can underlie corporate decisions (SCHOENBERGER, 1991), or where the issues are ill-defined, illunderstood or conceptually complex (HEDGES and RITCHIE. 1987). Examples of standardized interview schedules and non-standardized interview topics used in economic geography are reproduced in HEALEY (19Yla). The predominant type of interview used depends in part on the research design adopted. Standardized interviews arc most common in extensive research projects concerned to identify general patterns and outcomes. Such projects are typically associated with surveys of representative samples or populations and quantitative analyses in which taxonomic (i.e. classificatory) groups, such as rural areas or closures, arc used [e.g. FOTHERGILL and GUDGIN (1982) and STAFFORD and WATTS (199(l)]. Extensive research is the dominant research design used in positivist approaches to social enquiry. In contrast, nonstandardized interviews are more associated with intensive research designs in which the main ques-

and it is also easier

respondent to terminate the interview (WILSON, 1968). Telephone interviewing is sometimes used to increase the overall response rate in postal surveys [e.g. ELIAS and HEALEY (1991) and MARSHALL (1983)]. Although telephone interviews have been used successfully for up to half an hour [e.g. NORTH et al. (1983)], they would seem most suitable for obtaining answers to a small number of simple questions where it is reasonable to suppose that the person being spoken to knows the answers without having to consult any records. HAGUE (1987a) suggests that there is a limit of about 10 min for the length of time that a respondents attention can be held on the telephone for a structured interview. In many instances an interview-survey of businesses is

the only way to obtain the answers to the questions that researchers pose. For example, FOTHERGILL and GUY (1990, p. 44), in examining the closure of factories state that: The interview-based approach means that we can find out directly about the reasons for managers decisions rather than merely infer causation from statistics. In our view, this is by far the most fruitful and illuminating way to understand branch closures. Direct quotations from interviews can give useful insights into the way businesses respond to various situations and can provide anecdotes to illustrate other material [e.g. MARKUSEN (1991)].
Interview Techniques and Research Design

Standardized

und non-standardized

interviews

The reliability and usefulness of research findings and their interpretation are closely linked with the nature

GeoforumNolume

24 Number

3/1993
of standardized and non-standardized interviews*

343

Table 1. Outline Characteristics Research Theoretical Sample design approach

of the main characteristics Standardized

Non-standardized Usually intensive and qualitative, causally related groups Commonly used in realist and phenomenological approaches Selected to cover a range of issues and phenomena of interest. May be chosen as research progresses List of topics. Flexible, form and wording of questions vary with knowledge of respondents and interviewer, and direction of the interview error common Interactive, interview following issues raised in the open-ended examines

Usually extensive and quantitative, examines taxonomic groups Commonly used in positivist approaches

Representative

or whole population

Interview

schedule

Identical

questions

in a fixed order

Interview Questions Suitability

style

Minimization

of interviewer-related questions

Factual and pre-coded

Nearly all questions

For summarizing answers for sample, comparing responses to the same questions, generalizing, testing hypotheses, inferring causality skills Ability to interview consistently non-directively and

In-depth studies investigating causally related mechanisms, exploring new research areas seeking explanation and understanding Thorough understanding of research topic, ability to converse intelligently and with sympathetic understanding and MORGAN (1985). (see the text for a more

Interviewer

*Based in part on ideas drawn from Fig. 13 in SAYER (1992) and Table 6.1 in SAYER Summaries of this kind tend to emphasize the differences between the two types of interview balanced review). tions involve how the observed behaviour of a business is related to its own history and circumstances, or how some causal processes work out in particular cases. Intensive research typically uses mainly qualitative forms of analysis. It is the dominant research design associated with realist and phenomenological approaches in the social sciences. Intensive research commonly examines groups whose members are causally related (i.e. are connected in a relationship) to one another, such as firms which are related vertically through linkages [e.g. RAWLINSON (1991)] or horizontally through competition (i.e. operate in the same market sector). The businesses selected may not be typical and may be selected one by one as the research proceeds and as an understanding of the membership of a causal group is being developed. Often businesses are chosen to encompass the variability in the situations they face [e.g. CREWE (1989)]. A small number of cases examined in detail is common practice. The stress placed on identifying the context in which businesses operate in case study research means that sometimes lengthy interviews are required. R. Hayter (personal

communication, 1991), refers to interviews with four large Japanese companies undertaken by a doctoral student extending in one case to between 16 and 18 hr in total. Although associated with particular research designs, standardized and non-standardized interviews are not exclusive to extensive and intensive research designs, respectively. For example, non-standardized interviews may be used in the early stages of designing a standardized interview schedule. Moreover, the boundary between standardized and nonstandardized interviews is indistinct (BURTON and CHERRY, 1970). In some interviews a list of the required information is given to the interviewer and it is left to her or him to collect the data from the respondents. In such a semi-standardized interview, information on the same topics is obtained from each business owner or manager, but the way in which the questions are phrased and their sequence is left to the interviewer. This technique gives the interviewer the flexibility to tailor the interview to fit the particular circumstances faced by the respondent while ensuring

344 that comparable information is collected later be quantified [e.g. FOTHERGILL (1990)]. A mixture of interview the same investigation. Thus, one section which may and GUY of an interbuild upon

Geoforum/Volume knowledge gained

24 Number beforehand

3/1993 the

about

styles is often used in

specific characteristics of the business, instead of having to affect ignorance . in order to ensure uniformity or controlled conditions and avoid what bias (SAYER, might be taken as observer-induced 1984, p. 223). In non-standardized interviews the questions dependent posed on of the of the Conscfor

view may ask a common set of factual questions of all respondents, while in another section a semistructured qualitative approach may be used to explore an aspect of the behaviour RAWLINSON (1990)]. of the business [e.g.

vary from one interview

to the next,

One

of the

aims

of standardized errors

interviews (FOWLER

is to and where

how the interview develops, the knowledge respondent and the level of understanding interviewer quently quantitative VERMAN at the time of the interview. a different form of analysis

minimise

interviewer-related

MANGIONE,

1990). This applies particularly are involved

is required

several interviewers

in the same project,

though consistency of approach by a single interviewer is also important. The desirability of trying to create socially sterile conditions in the interview situation is, however, a normative ideal. It implies that the meaning of a given question is apparent to all respondents and the language used is interpreted by all of them in the way intended by the researcher (BRIGGS, 1986). Moreover, the interview involves an interpersonal encounter and it is unrealistic, as BRENNER (1978, p. 138) argues, to try to make the interviewer act without acting. as if the interviewer could ever not influence the situation of action in the interview by means of his own performance.

and qualitative information [c.g. SIL(1985)]. This is not the place for a dis-

cussion of techniques of analysing the data collected in interviews, although it is. of course, important that the research worker decides how he or she is going to analyse the information before it is collected. Many research workers think that it is inappropriate, for example, to attempt to reduce qualitative data using a mechanical quantitative style of analysis. As GORDEN (1975, p. 61) notes, with respect to material collected in non-standardized interviews. there is no way that the information can be statistically summarized to reflect the aggregate response of the group or individuals response to compare one with anothers. SCHOENBERGER (1992, p. 198) develops this argument a stage further when she notes that with in-depth interviews although statistical generalizations cannot be made, the method does permit analytical generalizations relevant to theoretical positions. This does not, however, mean that statistical information cannot be obtained in nonstandardized interviews, as long as some similar questions are asked of all respondents. MILNE ( 199 1), for instance, employed an unstructured discussion technique in his examination of small firms in the U.K. hifi industry, yet, because he asked a set of similar questions to his sample, he was able to produce nine tables which help to give an in-depth understanding of recent competitive pressures and reorganization in this sector. Validity and reliability The difficulty of generalizing about the responses to non-standardized interviews raises the debate about the validity and reliability of the findings of intensive qualitative research. Much has been written on this

With non-standardized interviewing the interaction between interviewer and interviewee is emphasized rather than minimized (HEALEY, 1991a). Sayer (personal communication, 1986) suggests that, as non-standardized interviewing is a social process, which involves both the interviewer and the interviewee, it is important to give the interviewee an active role and try to be as flexible as possible regarding the form and order of questions so as to accommodate for this. This view is confirmed by DEXTER (1970, p. 5) who states that the investigator is willing, and often eager to let the interviewee teach him what the problem, the question, the situation, is, within the limits of the research topic. It is important that the interviewer should also be actively involved in the process and the interview format should capitalise on the strengths of open-endedness. . . In it should allow for discussion and particular, dialogue+ven debate over controversial points (SCHOENBERGER, 1991, p. 187). With this method, the research worker is able to refer to and

GeoforumNolume important issue

24 Number [e.g. EYLES

30993 (1988), KIRK and

345 p. 11) summarizes the problem well when he observes that: Interpretative geography does not stand outside its subject-matter: and of the discourse similarly linear argues process, that it is part of the investigation itself. SYKES (1991, p. 4) qualitative proceeds research is not a of in a series

MILLER (1986), SAYER (1992) and SYKES (1990, 1991)]. Space permits room for only a few observations. The first point confused by the different are used. Here validity ing and liability results. Validity may be assessed in terms of both the inferand the meaningfulness is taken to refer is that discussion is often ways in which the concepts is taken to refer to the meanof the data, whereas reto the consistency of the

it rather

iterations with modifications of understanding occurring throughout the interviewing stage as well as during the formal analysis and writing up stages. goes on to suggest that given the problems She of assess-

ences which may be made from the findings kind and accuracy individual sample search workers

ing reliability it is important that the entire process of the research is made clear and the logic of the process of discovery however, compatible is communicated. reliability (BRIGGS, goals It is questionable, and validity 1986). are entirely SCHOENwhether

of the information obtained from units (SYKES, 1991). Some rethe theoretical inferences

emphasize

which may be made from qualitative research. For example, CREWE (1989) argues that the goal of intensive research is to expand and promote theoretical inferences, not to enumerate frequencies. This view is also used to justify case studies (MITCHELL, 1983), which like experiments, are generalizable to theoretical propositions and not to populations or universes (YIN, 1984, p. 21). Others suggest that qualitative findings may be used to extrapolate beyond the data and to make modest speculations about the likely applicability of the findings to other situations under similar, but not identical conditions. Extrapolations are logical, thoughtful and problemoriented rather than purely empirical, statistical and probabilistic (QUINN PATTON, 1986) [cited by SYKES (1991, p. 7)]. The second meaning of validity refers to the goodness of the data. Here the strengths of qualitative research are clearest. As SYKES (1991, p. 8) notes: The main reason for the potential superiority of qualitative approaches for obtaining information is that the flexible and responsive interaction which is possible between interviewer and respondent(s) allows meanings to be probed, topics to be covered from a variety of angles and questions made clear to respondents. Even if qualitative research is accepted as capable of producing valid results, there may still be doubts about whether they are reliable. These doubts are often expressed in terms of a question along the lines: would the same study carried out by two different research workers using the same techniques produce the same findings? The difficulty of assessing the reliability of qualitative research revolves around the nature of the research process it uses. EYLES (1988,

BERGER (1991, p. 11) suggests that the standardized interview is undoubtedly more reliable than the non-standardized interview. But the latter, when carefully administered, may offer greater accuracy and validity because it allows a more comprehensive and detailed elucidation of the interplay among strategy, history, and circumstances. By contrast, the standardized survey instrument must necessarily standardize and simplify a complex reality. Research workers need to be familiar with the strengths and weaknesses of both standardized and non-standardized interview techniques, and to know in what circumstances each is most appropriate. There are many examples where the design of standardized interview schedules is inappropriate, because insufficient thought is given beforehand as to why the information is being collected and the use to which it is to be put. Consequently, although some standardized surveys are technically very efficient, the findings may be of limited value (BLOOM, 1988). On the other hand, the findings from some nonstandardized interviews may be vague and anecdotal because of lack of prior preparation and/or insufficient skill in obtaining information during the interview. Despite the debate about the validity and reliability of qualitative research, it is apparent that, as economic geographers become more concerned with examining processes, relationships and interactions, rather than simply identifying patterns and outcomes, intensive research methods and nonstandardized interview techniques are becoming more frequently used [e.g. COOKE and WELLS (1990), D. Gibbs (personal communication, 1991) and Le Heron (personal communication, 1991)].

346 Identifying views Whichever Respondents and Obtaining Interalthough

Geoforum/Volume Clark (personal

24 Number

30993 1991)

communication,

suggests that it is now more important than ever to arrange interviews in advance. Prior permission to interviewing technique is used potential interview usually needs to be obtained when undertaking surveys in private spaces, such as a managed shopping centre. For example, LOWE (1991) found that centre management at Merry Hill Centre, Dudley placed restrictions on the amount and type of questions she was allowed to ask. A prior appointment is particularly important for large businesses. A request for an interview may be made either by letter or telephone. The latter often obtains an immediate reply, but may not achieve a higher response (FORSYTHE, 1977). Telephoning and fishing for a named person who can best handle the enquiry bcfore posting an introductory letter can work well (F. Peck, personal communication, 1991), as can obtaining an introduction by an appropriate intermediary (G. Linge. personal communication, 1991). Another possibility, when needing to contact people with broad overviews of their companies, is to write to the chief executives or managing directors requesting an interview with them or the person they think most appropriate [e.g. CREWE (1989)]. It can be helpful to enclose with the letter a short outline of the nature and purpose of the research project, and how the findings may be useful to the respondent (HEALEY. 1979). A letter from the sponsor of the survey (where relevant) or another influential party may also be helpful in persuading business managers and owners to participate in the research. It is often useful to follow up an introductory letter with a telephone call a few days later [e.g. ELIAS and HEALEY (1991) and WATTS (1991)]. It is more difficult for owners and managers to refuse an interview when speaking to the researcher over the tclephone. It also provides an opportunity for the researcher to deal with any queries that may exist (D. North, personal communication. 1991). Polite persistence is important in obtaining an interview. A series of rejections can be dispiriting for the research worker; however, it is almost always worth querying an initial refusal to see what is the reason. For example, a company may consider itself to be too small to be of importance to the researcher (HEALEY, 1979). Given the number of requests that busy executives receive, any opening in the correspondence request-

respondents have to be identified first. Talking to informed individuals and examining the business press are particularly useful when studies and when trying to maximise experiences sources are listings among the also useful used seeking case the range of

businesses chosen. These in updating the directory for constructing sampling

commonly

frames (HEALEY, 1991a). Identifying the best person in a large company to approach for an interview can itself be a major problem, especially in a general survey. In a specific study, for example of purchasing linkages, the section of the firm to approach can be much clearer. For many enquiries the most suitable person to contact may be someone at the head office or divisional office rather than at a branch. Sometimes the range of information required may make it desirable to interview more than one individual from a respondent company. Generally, when one good senior contact has been made it is easier to make others within the same organization. Introductions to managers in other tirms in the same industry may also be generated as a survey progresses. For example, in a study analysing buyer-supplier relationships in the motor industry. many of the names of buyers were obtained from the supplying firm respondents (WELLS and RAWLINSON, 1992). Identifying the business owner or manager can be problematic in some situations. For example. the concept of business owners and managers is anathema in the case of worker co-operatives (LOWE, 1989). In farming also. the hierarchy of power and management may be less visible and definitive than in most other industries and it may not be immediately clear where the locus of decision making lies between farm managers, agents and land owners. Where father and son(s) are working in parallel it may be necessary to talk to several farmers (G. Clark, personal communication, 199 1). The best strategy for obtaining an interview seems to vary with the size of business and the nature of the investigation. A short interview at a small firm can often be obtained by simply knocking on the door [e.g. HEALEY (1984)]. Similar tactics are also often used in studies of farmers [e.g. ILBERY (3985b)],

GeoforumNolume ing assistance

24 Number that gives them Although

3/1993 an excuse for refusal to outline ent interpretations. Should, for instance,

347 part-timers, about For

may be exploited.

it is important

sales staff and drivers issues facing a business

be included?

Questions

clearly the kind of information being asked for, refusal to cooperate is likely if the amount and level of detail of information requested is unreasonable. KINCAID and BRIGHT (1957a) advise against multiple requests in the same letter. They report that, when they asked not only for an interview but also for any literature pertinent to the study, several companies refused cooperation on the ground that since they had no pertinent literature there would be no point in conducting an interview. The evidence is equivocal, marring shoddy as few researchers their work, are prepared to risk success by deliberately sending out but a well-designed and presented

may also be problematic.

example, there is no unambiguous definition of skill shortages for employer-based research (MEAGER, 1986). One employer may include all unfilled vacancies, while another may only refer to vacancies which cause loss of output. These examples point to the desirability of probing the answers given to many questions Obtaining motives in order to clarify what respondents mean. accurate answers to questions on values, and perceptions is even more difficult.

DEAN and WHYTE (1958, p. 38) suggest that there is no point in asking whether the informant is telling the truth. Instead the researcher should ask What do the informants statements reveal about his feelings and perceptions and what inferences can be made from them about the actual environment or events he has experienced? On the other hand, MARSH (1979, p. 304) warns that: We must not confuse an impossible attempt to achieve absolute truth through unbiased questions, with the aim of being objective in our quest for truth, through trying to be as rigorous as possible in the way we draw conclusions from observations we make about the world, what people say and how they behave. The answers given to questions about motives, values and perceptions are influenced by a variety of factors. For example, in a study of the goals and values of farmers, GASSON (1973) notes the danger of relying on verbal indicators of values in that the answers given are influenced by the relationship established with the interviewer and who else is present at the interview. At the technical level open questions tend to elicit fewer responses than closed questions (BELSON and DUNCAN, 1962; HEALEY, 1991a). WATTS (1991), for example, found that business rates were mentioned as a reason affecting the choice of plant to close only when business executives were presented with a list of possible factors. There is also some evidence that the internal corporate context in which location decisions are made is more emphasized by respondents in in-depth analyses using openended and non-standardized interview techniques than in short analyses using standardized interview techniques (STAFFORD, 1974). To use pre-coded responses for attitudinal and motivational questions, a knowledge of the range of probable responses is necessary. They would therefore seem suitable only

letter, typed on headed notepaper, which is personally addressed with a handwritten signature, would seem to be a sensible way of trying to persuade the owners and managers of businesses to cooperate. Other examples of strategies for obtaining interviews are given in HEALEY (1991a).

Preparation for Interviews rate Answers

and Obtaining

Accu-

Before contacting potential respondents considerable effort needs to be put into preparing and planning interviews so that the information collected is useful in meeting the aims of the research. The nature of the responses obtained in standardized surveys depends on the quality of the interview schedule design and the way the questions are phrased. In nonstandardized surveys the quality of the answers is affected as much by the ability of the interviewer to engage the respondents in relevant discourse as by the phrasing of the questions. It is especially important in non-standardized interviews that the interviewer has a thorough understanding of the research subject so as to be able to make sense of the interview and to know when necessary. A pilot dardized interview for developing the searcher undertaking clarification or further probing is survey is essential to test a stanschedule and is highly desirable skills and knowledge of the rea non-standardized interview.

The way in which questions are phrased can have a critical effect on the responses given. Even apparently simple questions, such as How many people are employed in this establishment? are open to differ-

348 when a topic has already been well researched.

GeoforumNolume successes may become exaggerated minimized. The material gleaned needs, of course, to be interpreted

24 Number

3/l 993

Alternatively an open version of the questions may be used at the pilot stage before designing pre-coded responses for the full survey.

and the failures from interviews in the context of the interviews 1985). This is communi-

the conditions apparent at the time took place (MASSEY and MEEGAN, well illustrated by Le Heron

Respondents, in attempting to be helpful, may give answers which they think will please the questioner, or they may try to justify their actions (the problem of post-facto rationalization). Moreover, the interviewer cannot safely assume that the particular words used in a question interviewee problem are in fact the stimulus (DEXTER, the reasons to which the choice responds 1970). A related for locational for the choice of

(personal

cation, 1991). He argues that in New Zealand there have been two distinctive recent eras. In the era before deregulation in the mid-1980s the key words in the business community sification, and market cost reduction, included exporting, diverdevelopment; since dereguto phrases such as and competitiveness,

lation the key words have changed international back to basics. He suggests

in asking about reasons

is the need to distinguish

that if similar questions

region and settlement from those influencing site choice. Considerable confusions arises in many surveys where the scale element is not made clear in the phrasing of the questions and/or the interpretation of the results [e.g. DTI (1973) and MORIARTY (1983)]. This is because different factors tend to be important at different scales. For example, market accessibility, where significant, tends to distinguish one region from another, while the availability of industrial land may make one part of a region more attractive than another. Separate questions may be asked to identify the scale element [e.g. SCHMENNER (1982)] or different scales may be identified at the analysis stage [e.g. STAFFORD (1974)]. Further difficulties arise when questions are asked about events which occurred several years previously. or when the person answering the question had not taken part in the decision. There is also the problem of inferring the motives of a company from the responses given by a single representative of an organization. These problems, concerning questions about motives, values and perceptions can be reduced to some extent by asking only questions that it is reasonable to expect respondents to be able to answer, by careful interview design, and by exploring the topic in detail. Much of value may be learnt where such surveys are carefully executed. However. the results should be interpreted cautiously and are, perhaps, best taken as simply the views of a group of managers representing their businesses (HEALEY, 1991a). North (personal communication. 1991) also advises that interview data are treated with some scepticism, because for some owners/managers the interview provides an opportunity for an ego trip so that the

had been put a decade earlier the interviews would have veered off in entirely different directions. Several research workers counsel against accepting the view of one interviewee. For instance. at the conclusion of each interview STAFFORD (1974) made a subjective attempt to corroborate the testimony of the respondent through a brief discussion with another senior manager involved in the location decision. SAYER and MORGAN (1985) recommend interviewing both sides of industry in order to learn about their different interests, perceptions and responses, so as to reveal the structural positions of capital and labour. FOTHERGILL and GUY (1990, p. 106) also emphasize the dangers of accepting only one viewpoint in their study of the reasons for closurcs. They state that: Where management blamed the local trade union WC therefore did not take their view as the gospel truth. But equally, where the union blamed the management we did not necessarily accept their view either. Perceptions may also vary between establishment and higher-level managers. In managing reductions in the workforce and the introduction of new technology EDWARDS and MARGINSON (1988), for example, found that the local managers perceived less involvement from above than their higher-level counterparts. As much, if not more. skill and thought is needed in undertaking the non-standardized interviews used in intensive research as is required in constructing the standardized surveys used in extensive research. Not only does the interviewer need a thorough understanding of the research topic, but he or she should also have a sound knowledge of the industrial sector of the business and background information on the

GeoforumNolume

24 Number 3/1993 interpretation (1989)].

349 of the findings [e.g. MARSHALL

business itself, particularly where it is a large concern. Where the interviewer is well informed about the firm and the business it is engaged in, the respondent is likely to be more open and more detailed (SCHOENBERGER, 1991). A well-informed interviewer has a basis for assessing the accuracy of some of the information offered. This may be illustrated by the experience of two researchers undertaking a study of strategic alliances (COOKE and WELLS, 1990). In one interview they were given the impression that all the joint ventures the company was involved in were going smoothly. However, when one of the researchers mentioned that he knew of one which had run into problems, the respondent opened up and a frank discussion of the advantages and disadvantages of joint ventures followed. The knowledge level required by the interviewer to carry out nonstandardized interviews means that they should be undertaken by the principal researchers and should not be delegated to a team of interviewers. Unfortunately, the quality of the information about businesses available to research workers is limited. Much published information tends towards either superficiality, as in company reports, or towards selfcongratulatory rhetoric, as with glossy promotional publications (CREWE, 1989, p. 70). However, when combined with other sources, such as local and national newspaper reports, trade journals and inhouse company journals and newsletters, a useful overview can be built up (HEALEY, 1991b). For example, in a study of the introduction of new technology among motor vehicle assemblers, reported in RAWLINSON (1990), a helpful source of information was The Engineer, because it featured many of the developments along with photographs and explanations suitable for non-engineers to understand. Before a survey, it is useful to discuss the proposed interview topics with a few individuals familiar with the sectors being studied, because there are important differences in the terminologies used in different industries. Formulating questions in the respondents own language is preferable to assuming that the respondent can be taught to think within the researchers frame of reference (SCHOENBERGER, 1991, p. 187). It can also be very useful to talk to industry experts and trade associations after completion of the interviews to check the validity and

Conducting

and Recording Interviews

The most important factors contributing to a successful interview are trust and rapport (MOYSER and WAGSTAFFE, 1985). DEXTER (1970, p. 25) recommends that sympathetic understanding is the attitude most likely to promote such an atmosphere and hence yield the best response. The social skills of the interviewer are thus a key factor. The wider social science literature also suggests that interviewer characteristics, such as age, race and sex, may also influence the attitude of the respondents and the kind of answers they give [see also discussion in McDOWELL (1992) and S~HOENBERGER (1992)j. These characteristics are not malleable and thought may need to be given where appropriate to the composition of the research team in relation to the composition of the interviewees. This could, for example, be an issue when interviewing ethnic-minority business owners and managers [e.g. HEALEY et al. (1992)]. Power relationships can be important when interviewing top management. particularly in very large organizations. Interviewees may act in a patronizing manner to junior research workers. This problem may be reduced if interviewers can show a sound knowledge and understanding of the topic under discussion and that of the company being investigated. In most geographical research investigations it is appropriate to emphasize that no commercial secrets are being sought. H. Stafford (personal communication, 1991) recommends starting interviews by telling the respondents that no confidential information is being sought and that they should feel free to ignore any questions they wish. This, he suggests, allows the respondents to feel that they control the interview and as a result they are more relaxed and usually more forthcoming. Starting an interview on the right note is important. After explaining who you are, the sponsorship for the study, and to the necessary degree what the project is about, DEXTER (1970, p. 55) recommends beginning with comments or questions where the key words are quite vague or ambiguous, so the interviewer can interpret them in his own terms and out of his own experience. Le Heron (personal communi-

350 cation, 1991) and STAFFORD (1974) also recom-

Geoforum/Volume

24 Number

3/l 993

mend commencing an interview tion before the other content response. For example,

with an open quescan influence the

interviews were used by the respondents to assess the interviewers and how much they could safely be told (WELLS and RAWLINSON, 1992). Once this early period of probing by the respondents had taken place and they were satisfied that they had confidence in the researchers the interviews were more relaxed and the respondents were more prepared to answer the sensitive questions. This view is supported by Peck (personal communication, a degree prepared sensitive relations, 1991) who has found that once of rapport is established many managers are to be surprisingly open about potentially issues, closures such as attitudes and redundancies. to be prepared to unions, labour He has found to say quite a lot

in his study of the restructur-

ing of farm businesses in New Zealand, Le Heron began by asking about the history of the farm business so as to obtain a working knowledge advises of the responthat a question dents circumstances. Dexter

which sharply defines a particular area for discussion is far more likely than a general question to result in omission of some vital data of which the interviewer would not have thought. Some favour research workers [e.g. HAGUE using pre-coded responses because (1987b)] they save

that managers

appear

respondents having to think of possible replies and avoid having to make difficult coding judgements at the analysis stage. Other investigators choose to use open questions so as not to restrict the range of responses. MOYSER and WAGSTAFFE (1985, p. 18) express surprise at the extent to which some elite ligures will try to fit themselves and their views into boxes for the sake of social scientific methods. However. many elite interviewees prefer to give their own interpretations rather than be forced to choose between categories of responses which often do not seem to give an adequate summary of the situation as they perceive it. SCHOENBERGER (1991, p. 183) makes a similar point when she notes that: The respondent may also be frustrated by questions or by a range of possible answers . that do not apply precisely to his or her own experience. Respondents are likely to feel less frustrated if they are able to explain exactly what they mean in their own terms rather than trying to fit themselves into the terms of reference proposed by the researcher.

of things which might be considered self-critical, or at least critical of the methods used by their firms. To maintain the flow of the interview and to obtain useful insights into business behaviour, SCHOENBERGER (1991) recommends building problems into the interview for the respondents to solve. This may involve asking interviewees to explain. where appropriate, why their accounts vary from that predicted by a particular theory or why the behaviour of their firms differs from that of other firms in the same industry. This tactic provides a useful check on the consistency of the views expressed earlier in the interview by the respondents. Another strategy, advocated by Schoenberger, is for the researcher to summarize her or his interpretation of what the respondent has said about particular topics that have already been discussed and invite the interviewee to evaluate its adequacy or appropriateness. This tactic also helps to avoid the researcher inadvertently misinterpreting the interview to produce a given answer. In a standardized interview a useful check on the completeness of the information provided is to end the interview with an open-ended discussion. This technique was used in a case study of defencc-based industries in Chicago where the researchers closed each interview by eliciting from their informants their views on why Chicago has not fared better as a centre for defence production and research (MARKUSEN and McCURDY, 1989). In many interview situations the interviewer needs to be flexible as to where and when interviews take place and be prepared for frequent interruptions as the respondent continues to run the business. This is particularly true for small businesses where inter-

Asking questions about sensitive issues can be a problem. For example, in farming these may include land tenure. pollution, livestock movements off the farm, family contributions to the farm, other jobs off the farm, and other sources of income such as investments (Clark, personal communication, 1991). It is usually best to leave sensitive questions until near the end of an interview, because this allows a greater time for the respondent to build up trust and confidence in the researchers. In a six-country study of the automotive industry, which dealt with many commercially sensitive issues, such as future corporate strategies, the researchers noticed that the first part of the

GeoforumNolume views may take

24 Number place on the

3/1993 shop floor. Clark that hand, the act of taping may inhibit the responses

351 and

(personal

communication,

1991) recommends

transcribing

the interviewer is suitably dressed for the likely location of the interview so that he or she blends in. Sensible attire for the farm is clearly different from that for the head office of a large corporation. Inappropriate clothes may have a quite disproportionately negative effect (BULMER, are, however, insurmountable 1988). Some problems as, for example, when

and analysing the interviews can take a Several commentators [e.g. considerable time. ave noted that some of the most BULMER (1988)] h interesting insights are obtained when the tape recorder A number is turned off. issues can arise in the way

of ethical

Clark turned up at a farm seeking an interview only to find that it was run by Trappist monks whose vow of silence made the quest rather Obtaining important an accurate skill. There difficult!

interviews are conducted the information. Apart confidentiality geographical BULMERs research applicable surveys:

and the use that is made of from some comments on upon in the interviews. seem

these are hardly touched literature on business (1988, p. 156) advice would,

on undertaking however, faced in business

record of an interview is an is a wide variety of ways of

in organizations

to many of the situations

taking notes. However, one thing all research workers are agreed on is the importance of writing up the notes as soon as possible after the completion of the interview. Some researchers favour interviewing in pairs as a way of eliminating interviewer bias and providing a check on what is said during the interview [e.g. KINCAID and BRIGHT (1957b) and NORTH peret al. (1983)]. I n one study of the competitive formance of small firms one interviewer used a schedule to structure the interviews, while a second engaged the respondents at appropriate points in a more discursive dialogue about their knowledge base, competitive strategy and flexibility (HORNE et al., 1987). An alternative to taking notes is to tape-record the interviews. This is common practice in North America [e.g. GORDEN (1975) and STAFFORD (1974, 1985)] and has been used successfully in Britain [e.g. CREWE (1989) and PRATT (1989)]. It provides an accurate record of the interview and enables the finished report to be enlivened with extended direct quotes. However, in one study of subcontracting, whereas the managers of large manufacturing companies raised no objection to having their interviews recorded, an attempt to tape-record interviews with the owners of small precision engineering companies had to be abandoned after the first eight owners declined (RAWLINSON, 1990). The acceptability of taping an interview may, however, depend on the approach the interviewer takes. GORDEN (1975) advises explaining why it is used rather than asking for permission. By taping interviews it enables the interviewer to concentrate on the phrasing and order of questions rather than on note taking. On the other

The researcher gaining entry to a setting has the responsibility to explain fairly and openly to the organization being studied what is the purpose of the enquiry. There is also a responsibility to show members of the organization draft material in order to allow the correction of factual inaccuracies. There is the further obligation in any material that is published to safeguard the organization from revelations that could be harmful to its commercial activities or its reputation. The commonest device for achieving this end is the use of anonymity in naming the organization and possibly not too closely identifying the locality in which it is situated. On the other hand, the researcher should retain to themselves (sic) the right to publish material gathered in the course of the study.

A promise of confidentiality is reassuring to respondents and is likely to make them more cooperative and open. However, it is not always possible to hide the identity of the businesses being examined, particularly when examining industries or local economies dominated by a few large enterprises. Where confidentiality has been promised, the problem becomes one of how to guarantee it. This may require a degree of judicious vagueness. Some detail may have to be suppressed in order not to provide a set of clues that, taken together, would lead someone else to be able to identify the firm. Similarly, if direct quotations from interviews are to be used in the final report these should be non-attributable, unless permission has been given from the respondent.

Conclusion Interviewing business owners important method of obtaining and managers is an information in econ-

352 omit geography. However, surprisingly little has

Geoforum/Volume

24 Number

30993

been written about this method. Although the issues involved in business surveys are not unique, several points about research are highlighted views, including niques and obtaining methods concerning interviews intertechand when undertaking the link between design, access, preparing business interview

deposited there. Collaboration in the analysis and dissemination of information possibility, business Second, business different including particularly surveys (FOLEY, where there 1990).

collection, is another in

is overlap

and research

identifying

respondents

there is no one best way of interviewing owners and managers. Methods vary for situations, the research depending on a range of factors, of design, the kind and amount

for interviews

obtaining accurate answers, and ways of conducting and recording interviews. This paper has drawn upon both identify published sources and a survey methods of Angloto of interviewAmerican economic geographers conclusions in an attempt

information required, the size, organizational

the resources available, and structure, sector and location Although the may provide a interview (as illus-

some of the successful methods

of the business to be approached. philosophy of the research worker guide to the methods have methods associated

ing. Three general of interview

arise from this review

used and different characteristics

and techniques.

First, the cut backs in official surveys in some countries, such as the United Kingdom, have increased the need for research workers to collect their own data, yet the associated anti-bureaucratic/ form-filling climate of cutting the red tape can make this task more difficult (HEALEY, 1991~). Done properly business surveys take considerable time and effort and should only be contemplated after all other sources of information have been exhausted. It is also desirable to examine the interviewing process from the point of view of the respondents. To encourage their continued cooperation in survey work it is important to maintain good relations with the business community, not least because of the value of resurveying the same businesses at a later date [e.g. MUNTON et al. (1988) and NORTH et al. (1983)]. Too often research workers neglect to send a short thank you letter following an interview or forget to send interviewees the promised summary of the results of their projects. Surveys which attempt, within the limits of the research topic, to make the experience as interesting and stimulating as possible for the participants are also likely to make business owners and managers more favourably disposed to participate in future surveys (HEALEY, 1991a). This may be easier with the increased use of qualitative nonstandardized interviews which give a more active role to respondents. There is, however, a danger of oversaturation, particularly of large businesses, if too many surveys are attempted. Consequently, there is a case for making the data from surveys more widely available. In Britain the ESRC Data Archive acts as a repository for surveys, though relatively few economic geography business surveys have so far been

trated in Table l), the reality of undertaking research shows that the fit between theory and method is not nearly so neat as is sometimes supposed. As BRYMAN (1984, p. 89) notes: there is no necessary 1: 1 relationship between methodology and technique. Different philosophical assumptions do not simply imply different approaches but also imply the different usage of similar approaches. The link between methods and philosophical approach needs further investigation. Last, the lack of clear accessible guidance in the literature means that most researchers have little choice in designing their surveys than to rely on a combination of common sense and previous experience. Little progress will be made with improving the quality of information obtained from interviewing business owners and managers until research workers are prepared to experiment with different methods to try to increase response rates, hasten replies. obtain clearer answers, improve the flow of interviews and so on. Equally it is important that researchers and journal editors give greater priority to reporting these findings so that a reasonable map may be produced to guide economic geographers through what is currently still largely terra incognita.
Acknowledgements-We gratefully acknowledge the help given to us by colleagues who told us about some of their experiences of interviewing business owners and managers. many of which have been incorporated into this paper. Thanks are due to Gordon Clark, Louise Crewe, David Gibbs, Roger Hayter, Rob Imrie, Richard Le Heron, Gordon Linge, Anna Markusen, Michelle Lowe, Kevin Morgan, David North, Frank Peck, Andy Pratt, John Rees, Allen Scott, Howard Stafford, Doug Watts and Peter Wells. Erica Schoenberger helpfully provided us with a pre-

G~~forum~olum~

24 Number 311993

353
ences In perception between establishment and higher level managers, In: Beyond the Wtlrkplace: Managing

publication copy of her paper on interviewing. Andy Pratt kind$ commented on a draft of this paper. Respo~s~biIity for what is expressed &es with the authors alone,

~nd~tr~a~ Related in the Marts-esta~l~hment Enterprise, pp. 227-257, P. Marginson (Ed.). Basif Blackwell,

References
BACHTLER, J. (1984f Wherefore ant thou, questionnaire?, Frontier (the magazine of the ~epa~meot of Geography, University of ~i~in8ham~, 3, 105-113. BELSON, W. and DUNCAN, J. A. (1982) A comparisan of the check-list and the open r~sponse/questi~~~og systems, Appl. Statist., 11, 120-132. BLOOM, N. (1988) The limitations of unqualitative research, frad. Mark~f~ng Dig. ) 13,49--S& BRENNER, M. (1978) Inte~i~wing: the social phenomenobgy of a research instrument, In: The Social Cantexts of Method, pp. 122-139, M. Brenner, P. Marsh and M. Brenner (Eds), Croom Helm, London. BRIGGS, C. L. (X988) Learning How to Ask: a Sclciolinguistic Appraisur of the Role of the Interview in Social Science Research. Cambridge University Press, Cam-

Oxford. ELIAS, P. and HEALEY, M. J. (1991) Pe0~~eund Work in Coventry: a Survey of ~~~~~ayers. Coventry City Council, Coventry. EYLES, J. (1988) Interpreting the geographical world: qualitative approaches in geographical research, In: Q~~itat~ve Methods in Human Geugraphy, pp. I-16 .I. EyJes and D. M. Smith (Eds). Polity Press, Cambridge. EYLES, J. and SMITH, D. M. (Eds) (1988) Qualitative Methods irz Human Geography. Polity Press, Cambridge. FOLEY, P. (1990) Growth in the folk&on of labour market ~nformation~ Reg. Stud., 24,367-371. FORSYTHE, J, B. (1377) Obtaining cooperation in a survey uf business executives, J. Marke?~~~~ lies., 14, S. and GUDGTN, G, (1982) Unequal Growth: Urban and Regional ~~n~~~~~e~~ Change in the UK Heinemann Educational I London. FOTHERCILL, S. and GUY, N. (1990) Retreatfrom the Regions: Corporate Change and the Closure of Factories, FOTHERGILL, Regional Policy and envelopment Series 1. Jessica Kingsley and Regional Studies Association, London. FOWLER, F. J. and MANGIONE, T. W. (1990) S&zzdardized Survey ~nte~~ew~ng: ~~~njrniz~g fnterviewrelated Error, Applied Social Research Methods Series 370-373.

bridge. BRYMAN, A. (1984) The debate about quantitative and qualitative research: a question of method or epistemology?, Br. J. Social., 35,75-92. LAYMAN, A. (1988) Doing Research in ~rganizat~~~~s~ Routledge, London. BRYMAN, A. (1989) Research retards and ~rgan~zat~o~ Studies. Unwin Hyman, London, BULMER, M. (1988) Some reflections upon research in organizations, In: L?oing Research in Organizations, pp. 151-161, A. Bryman (Ed.). Rout~~~ge, London. BURTON, T. L. and CHERRY, G. E. (1970) Social Re~~earc~ Techniques for Pianners. George Allen &z Unwin, London. COOKE, P. and WELLS, P. (1990) Strategic A~~~an~es in XT: Learning by Interaction. Regional Industrial Research Report 4, University College Wales, Cardiff. CREWE, L. (1989) Industrial restrwet~ring in West Yorkshire: some emp&ical and theoretical developments with particular reference to the textiles-clothing-retail distribution system, ~npub~~sh~d Ph.D thesis, University of Leeds. DEAN, J. P. and WI-IYTE, W. F. (1958) How do you know if the informant is telling the truth?, Hum. Org., 17,3438. DE ~HERNATONY, L. (1988) iZ&zttirzg the Most from Postal Research. Occasional Paper No. 3, Industrial Marketing Research Association, Lichfield. DEXTER, IL. A. (4970) E&e and Specialized ~nter~?~ew~ng. Northwestern University Press. Evanston, IL. DIXON, C. .I. and LEACH, B. (1978) Questionnaires and ~~~terviewsin ~e~gruphica~ R~.~e~rc~, Concepts and Techniques in Modern Geography 18. Geo Abstracts, Norwich. DTI (DEPARTMENT OF TRADE AND INDUSTRY) (1973) Memorandum on the inquiry into location attitudes and experience, In: Regional ~e~efo~~e~~ &cerltives, Session 1933-74, Minutes of Evideme (from July 1973), Appe?zdi~es and Index, pp. X5-668. Expenditure Committee (Trade and Industry Sub-committee), HMSO, London. EDWARDS, P. K. and ~~G~NSGN~ P, (19%) Differ-

18, Sage, London. GARDNER, G. (1978) So&al Surveysfor Social Planners. Open University, Milton Keynes. GASSON, R. (1973) Goals and values of farmers, 1. ugric. Econ., 24521-532. GORDEN, R. L. (1975) ~~~e~~~e~~~~~ Strategy, ~e~h~~~4aes and Tactics. Dorsey Press, Homewood~ IL. GUDGIN, G. (1978) hdustriaf Locafion Processes and Regional Employment Growth. Saxon House, Farnborough. HAGUE, P. (1987a) The I~d~triu~ Market Research Handbook. Kogan Page, London. HAGUE, P. (X987b) Good and bad in questionnaire design, Ind. Marketing Dig., 12, lfiI--f7#. HART, S. (3987) The use of the survey in industrial market research, J. Marketing Mgt, 3(l), 25-38. HEALEY, M. J. (1979) Changes in the location of production in muIt~-plant enterprises, with particular reference to the United Kingdom textile and clothing industries, 1%7-72, Unpublished Ph.D thesis, University of Sheffield. HEALEY, M. J. (1983) The changing data base: an overview, In: Urban and Regional Industrial Research: the Cha~g~~~ UK Data Base, pp. l-29, M. J. Healey (Ed.). Geo Books, Norwich. HEALEY, M. J. (1984) The Covenfry Region ~nd~tria~
Esta~~~s~mentdatabank 1:A Guide to Sources, Methods and renditions ~sedfor the City ~fcaven~ry. Industriaf

Location Working Paper 2, Department of Geography, Coventry Polytechnic. HEALEY 1M. J. (1991a) Outlining information from businesses, In: Economic Activity and Land Use: the Chang ing information Base for tocai and Regional: Studies, pp_ 193-251, M. .I. Heaky (Ed.). Longman, Harfow,

354
HEALEY, M. J. (Ed.) (1991b) Economic Activity and Land Use: the Changing Information Base for Local and Regional Studies. Longman, Harlow. HEALEY. M. J. (1991~) Improving the local and regional economy information base, Reg. Stud., 25, 363-369. HEALEY, M. J., NOON, D., BERKELEY, N.. DUNHAM, P. and HASLUCK. C. (1992) Industrial Strategies, Economic Development and Job Patterns in Leicesters Retail Sector. Leicester City Council, Leicester. HEALEY. M. J. and WATTS, H. D. (1987) The multiplant enterprise, In: Industrial Change in the United Kingdom, pp. 149-166, W. F. Lever (Ed.). Longman. Harlow. HEDGES, B. and RITCHIE, J. (1987) Research and Policy: the Choice of Appropriate Research Methods, SCPR, London. HOINVILLE, G., JEWELL, R. and ASSOCIATES (1978) Survey Research Practice. Heinemann Educational Books, London. HORNE, M. R., LLOYD, P. E.. PAY, J. L. and ROE, P. (1987) Structuring Knowledge of the Small Firm: a Framework for Informing Local Authority Intervention within the Private Sector. Working Paper Series 20, School of Geography, North West Industry Research Unit, University of Manchester. ILBERY, B. W. (1985a) Agricultural Geography. Oxford University Press, Oxford. ILBERY, B. W. (1985b) Factors affecting the structure of horticulture in the Vale of Evesham: a behavioural approach, /. rur. Stud., 1, 121-133. ILBERY. B. W. (1987) The development of farm diversification in the UK: evidence from Birminghams urban fringe, Jl r. agric. Sot., 148,21-35. JOBBER, D. and BLEASDALE, M. J. R. (1987) Interviewing in industrial marketing research: the state-ofthe-art, e. Rev. Marketing, 12(2), 7-11. KEEBLE, D. and WEVER, E. (Ed.) (1986) New Firms and Regional Development in Europe. Croom Helm, London. KINCAID, H. W. and BRIGHT, M. (1957a) Interviewing the business elite, Am. J. Social., 63, 304-311. KINCAID, H. V. and BRIGHT, M. (1957b) The tandem interviews: a trial of the two-interviews team, Publ. Opinion Q., 21, 304-312. KIRK, J. and MILLER, M. (1986) Reliability and Validity in Qualitative Research. Sage. London. KRUMME, G. (1969) Towards a geography of enterprise. Econ. Geogr., 45,30-40. LOWE. M. (1989) Worker co-operatives-panacea or radical enterprise?. Envir. Plann. D Sot. Space, 7. 3149. LOWE, M. (1991) Trading places: retailing and local economic development at Merry Hill. West Midlands, E. Midld Geogr., 14(June), 31-48. MACFARLANE-SMITH, J. (1972) Interviewing in Market and Social Research. Routledge & Kegan Paul, London. MARKUSEN, A. (1991) The Rise of the Gunbelt. Oxford University Press, Oxford. MARKUSEN, A. and McCURDY, K (1989) Chicagos defense based high technology: a case study of the seedbeds of innovation hypothesis, Econ. Dev. Q. 3, 15-31.

Geoforum/Volume

24 Number

3/1993

MARSH. C. (1979) Problems with surveys: method or epistemology?, Sociology, 13,293-305. MARSHALL, J. N. (1983) Business-service activities in British Provincial conurbations, Envir. Plann. A. 15, 1343-1359. MARSHALL, J. N. (198Y) Corporate reorganization and the geography of services: evidence from the motor vehicle aftermarket in the West Midlands region of the UK, Reg. Stud.. 23, 139-150. MASON. C. M. (1989) Explaining recent trends in new firm formation in the UK: some evidence from South Hampshire, Reg. Stud., 23. 331-346. MASSEY, D. and MEEGAN, R. (Eds) (1985) Politics and Method: Contrusting Studies in lndustriul Geography. Methuen, London. MCDOWELL, L. (1992) Valid games? A response to Erica Schoenberger. Prof. Geogr., 44,212-215. McKINNON. A. C. (1985) Physical Distribution $ystem.s. Routledge. London. McNEE, R. B. (1960) Towards a more humanistic economic geography, Tijdschr. econ. sot. Geogr., 51. 201206. MEAGER, N. (1986) Skill shortages again and the UK economy, Ind. Relat. .I., 17. 236248. MILNE, S. (1991) Small firms, industrial reorganization, and space: the case of the UK high-fidelity audio sector, Envir. Plann. A, 23, 833-852. MITCHELL, J. C. (1983) Case and situational analysis, Social. Rev., 31. 187-21 I, MORIARTY, B. M. (1983) Hierarchies of cities and the spatial filtering of industrial development, Pup. reg. Sci. Ass., 53, 59-82. MOYSER, G. and WAGSTAFFE, M. (1985) The Methodology of Elite Interviewing. Report to the Economic and Social Research Council, Grant H00250003, London. MOYSER, G. and WAGSTAFFE, M. (Eds) (1987) Research Methods for Elite Studies. Allen & Unwin, London. MUNTON, R., WHATMORE, S. and MARSDEN, T. (1988) Reconsidering urban-fringe agriculture: a longitudinal analysis of capital restructuring on farms in the Metropolitan Green Belt, Trans. Inst. Br. Geogr. New Ser., 13, 324-336. NORTH, D. J., LEIGH, R. and GOUGH, J. (1983) Monitoring industrial change at the local level: some comments on methods and data sources. In: Urban and Regional Industrial Research: the Changing UK Dutu Base, pp. 111-129, M. J. Healey (Ed.). Geo Books. Norwich. PRATT, A. (1989) Towards an explanation of the form and location of industrial estates in Cornwall 1984: a critical realist approach, Unpublished Ph.D thesis, University of Exeter. QUINN PATTON, M. (1986) Utilization-focused Evaluation. Sage, London. RAWLINSON, M. B. (1990) Subcontracting relationship between small engineering firms and large motor vehicle firms in the Coventry area, Unpublished Ph.D thesis. Coventry Polytechnic. RAWLINSON, M. B. (1991) Subcontracting in the motor industry: a case study in Coventry, In: Restructuring the Global Automotive Industry, pp. 215-230. C. Law (Ed.). Routledge, London.

GeoforumNolume

24 Number

3/1993

355
Industrial Marketing Research, pp. 33-42, A. Rawnsley (Ed.). John Wiley, Chichester. SYKES, W. (1990) Validity and reliability in qualitative market research: a review of the literature, J. MarketRes. Sot., 32,289-328. SYKES, W. (1991) Taking stock: issues from the literature on validity and reliability in qualitative research, J. Market Res. Sot., 33, 3-12. TAYLOR, M. J. and THRIFT, N. J. (Eds) (1982) The Geography of Multinationals. Croom Helm, London. TAYLOR, M. J. and THRIFT, N. J. (1983) Business organization, segmentation and location, Reg. Stud., 17, 445465. THRIFT, N., DANIELS, P. and LEYSHON, A. (1988) Growth and Location of Professional Producer Service Firms in Britain. Final Report to Economic and Social Research Council, Grant D00232194, London. WATTS, H. D. (1980) The Large Industrial Enterprise. Croom Helm, London. WATTS, H. D. (1991) Plant closures, multi-locational firms and the urban economy: Sheffield, UK, Envir. Plann. A, 23,37-58. WELLS, P. and RAWLINSON, M. (1992) Technology procurement practices, and structural change in the automotive presswork sector, In: Sheet Metal, pp. 9% 112, B. Shirvani and D. Briggs (Eds). IOP, Bristol. WILSON, A. (1968) The Assessment of Industrial Markets. Hutchinson, London. YIN, R. K. (1984) Case Study Research: Design and Methods, Applied Social Research Methods Series 5. Sage, Beverly Hills, CA.

RAWNSLEY, A. (Ed.) (1978) Manual of Industrial Marketing Research. John Wiley, Chichester. SAYER, A. (1984) Method in Social Science. Hutchinson, London. SAYER, A. and MORGAN, K. (1985) A modern industry in a declining region: links between method, theory and policy, In: Politics and Method: Contrasting Studies in Industrial Geography, pp. 144-168, D. Massey and R. Meegan (Eds). Methuen, London. SCHMENNER, R. W. (1982) Making Business Location Decisions. Prentice-Hall, Englewood Cliffs, NJ. SCHOENBERGER, E. (1991) The corporate interview as a research method in economic geography, Prof. Geogr., 43, 18G189. SCHOENBERGER, E. (1992) Self-criticism and selfawareness in research: a reply to Linda McDowell, Prof. Geogr., 44,215-218. SILVERMAN, D. (1985) Qualitative Methodology and Sociology. Gower, Aldershot. STAFFORD, H. A. (1974) The anatomy of the location decision: content analysis of case studies, In: Spatial Perspectives on Industrial Organization and Decisionmaking, pp. 169-187, F. E. I. Hamilton (Ed.). John Wiley, London. STAFFORD, H. A. (1985) Environmental protection and industrial location, Ann. Ass. Am. Geogr., 75,227-240. STAFFORD, H. A. and WATTS, H. D. (1990) Abandoned products, abandoned places: plant closures by multi-product multi-locational firms in urban areas, Tijdschr. econ. sot. Geogr., 10,206-227. SWAIN, W. (1978) The postal questionnaire, In: Manualof