Information Technology Reference
In-Depth Information
and politeness, and rapport and defined two meta-concepts, namely “information”
and “strategy”, to account for the interpretations that remained. To determine
whether these theories matched the factors, we investigated whether these factors
could be explained by the concepts underlying the theories. We found that our
initial factor interpretation and the match between factors and concepts overlapped
broadly. We also found that many factors were matched to more concepts than
initially were associated with the factors. We used this finding to create a collection
of interrelated concepts that gives insight into how the different theories relate to
each other. With this collection, we are able to (at least partly) describe the behaviour
of both police officers and suspects in an interview setting.
Our combination of holistic and theory-driven methodology does, however, have
its limitations. As is the case in most observational studies, our annotations of
the police interview corpus were based on our interpretations of the behaviour of
the interacting parties and thus subjective. For future work, our methodology may
be repeated to include more observers (and more independent observers) which
may lead to a broader semantic frame, possibly alleviating problems inherent with
interpretation of behaviour.
A second limitation of our approach is that it currently focuses on describing
short
fragments from the corpus. In the previous section however, we illustrated
how our findings may be extended to explain changes in the behaviour of interacting
parties over longer periods of time. We based these examples on how temporal
aspects are explained by the theories from which we drew our concepts. We wish to
continue this line of research by investigating how the interplay of these concepts
influences the dynamics in police interviews. This may, for instance, be done
by locating moments in our corpus in which a person's behaviour changes. For
example, there may be moments when a person changes his or her stance or becomes
less polite. Someone may also consciously change his behaviour to evoke desired
behaviour of the other party. This may, for example, be the case when a police officer
adopts a “together” stance to build rapport with a suspect. Thus, the communicative
contexts before and after this type of changes in behaviour should be compared, to
discover what may have caused the change in behaviour. This causality is of vital
importance for the creation of a virtual suspect agent, as such an agent needs to be
capable of taking logical (and explainable) actions. Thus, this calls for an extended
empirical study of the corpus. Such a study may also validate the links we found
between concepts, as our current work only investigated a number of fragments
from our corpus.
Lastly, we also wish to investigate how the methodology we present in this
paper translates to other domains. Whether our approach can be used to analyse
communicative behaviour in other domains depends on the availability of a corpus
and theories on interaction that explain (parts of) the behaviour. A related domain
in which we are also involved is that of street interventions by police officers with
loitering juveniles. This domain features a different setting and the environment
imposes other restrictions on the interaction, such as an easier “way out” for the
juveniles because they are not kept in a room like the interviewees. Still, this domain
does not differ strongly from the domain discussed in this paper, as they are both
Search WWH ::
Custom Search