of commodity flow, passenger trips, traffic
movements and so on. The volume of move-
ment between an origin and a destination is
proportional to the number of opportunities at
that destination, and inversely proportional to
the number of opportunities between the ori-
gin and the destination. Stouffer argued that
distance of itself has no effect oninteraction
patterns and that any observed decline in the
number of movements with distance (seedis-
tance decay) is due to the increase in the
number of intervening opportunities with dis-
tance. A variant of this approach was devel-
oped in Fotheringham’s (1983) ‘theory of
competing destinations’. rj
interviews and interviewing Widely used
methods for learning about the experiences,
attitudes and demographic characteristics of
individuals, households or groups. Interviews
can be structured or unstructured, and they
can be administered face-to-face, over the
telephone or via email. An interview using a
surveyquestionnairefollows a set order of
pre-established questions. Survey data under-
pins large scale quantitative social science, and
can be effective for establishing attitudinal,
demographic and socio-economic patterns
across large samples representative of vast
populations (see quantitative methods;
sampling;survey analysis). Acensusis a
national survey of the entire population.
Unstructured interviews are more conversa-
tional than a questionnaire, and allow inter-
viewees to express the details and meanings of
their experiences in their own terms and at
their own pace. Unstructured interviews can
be conducted with individuals, households or
as focus groups, and are an appropriate
qualitative methodfor understanding com-
plex and contradictory social processes and
experiences, and when respondents need the
opportunity to explain and qualify their
accounts. But even the least structured ethno-
graphic interview is very different from an
ordinary conversation. It is important for an
interviewer to recognize this, so as not to be
caught within implicit rules of social conversa-
tion (Anderson and Jack, 1991). In contrast to
ordinary conversations, the norm in ethno-
graphic interviews is to repeat questions, ask
for clarification of terms, and introduce a ser-
ies of ethnographic explanations and styles of
questions (Spradley, 1979: seeethnography).
Because they are unstructured, in-depth inter-
views are time-consuming (they typically last
for between one and two hours), and the sam-
ple is usually much smaller (and likely less
representative of the population) than is the
case for questionnaire surveys.
There has been considerable discussion of
the influence of the interviewer on what is told
and heard, and of thepowerdynamics be-
tween interviewers and interviewees. Although
it is advisable for interviewers to dress and
comport themselves so as to ‘fit in’, it is widely
assumed to be impossible and undesirable
for them to neutralize their presence. The re-
searcher is integral to the interview process,
and his or hergender, age,sexuality,class
andrace(and many other characteristics) will
affect access and what they are told. In survey
research, this is known as the ‘interviewer
effect’. There are two issues here. First, what
we are told is situational, depending on the
perceived social characteristics of the inter-
viewer, the location of the interview, and
many other contextual factors. Second,
because most interviews are structured by ‘a
division of labour in which one talks and one
listens’, relations of oppression and domin-
ation may be unwittingly reproduced within
them (Bondi, 2003, p. 70; but see England
(2002) and McDowell (1998) for a different
set of dynamics when interviewing elites). Both
concerns have led to recommendations that
interviewers reflect upon theirpositionality
in order to assess how they may be affecting
(and affected by) the interview situation (see
situated knowledge). Rose (1997b) cautions,
however, that there are limits to suchreflex-
ivity, because we are not and cannot be fully
conscious or transparent to ourselves; and
Bondi (2003) notes that much communica-
tion within an interview is non-verbal and
non-cognitive (see alsonon-representational
theory). Power relations, and points of com-
monality and difference between interviewers
and interviewees are also mobile and complex
(McDowell, 1998; Kobayashi, 2001; Crang,
2002; England, 2002). Concerns about power
relations have led to experiments training and
working with community-based interviewers
(Gibson-Graham, 1994; Pratt, 2004), and to
calls for more activist research (Kobayashi,
2001; see alsoactivism,action research).
Key objectives for any unstructured inter-
view are to create an intersubjective space in
which the interviewee can express him or
herself fully, and to strive to understand
what is communicated as fully as possible,
whilst causing no harm to the interviewee.
Bondi (2003) makes a distinction between
empathy and identification: the former (the
recommended stance) involves the capacity
to understand the interviewee’s feelingsGregory / The Dictionary of Human Geography 9781405132879_4_I Final Proof page 393 31.3.2009 7:05pmINTERVIEWS AND INTERVIEWING