You are on page 1of 2

Achieving a High Confidence Level

Unlike with quantitative methods, the problem of assessing the confidence level of qualitative research is
complex and quite debatable. However, the acid test of the confidence level of the findings of ethnography
rests on its presentation of evidence. In assessing evidence we look for credibility. If the evidence is not
credible, we have to continue the search for more evidence until the findings can pass the acid test
(Hammerley 2001, 1992). The credibility of evidence rests largely on the quality of the methodology
employed to collect data. In qualitative research, methodology relies heavily on the integration of research
tools and triangulation to make a claim of validation.

No single technique can claim methodological monopoly over others. Each has a unique value, a set of
advantages in collecting a specific kind of data in specific settings. The methods used in any study must
integrated in such a manner that each complemented the other, thus increasing reliability and validity. To
illustrate, participant observation, makes a claim to validity, it does not “provide us with data that are
demonstrably reliable either across investigators or over time” (Goldenberg, 1992:339). Therefore, it
depends on comparison with statistical data and documentation to strengthen its claim to some degree of
reliability. This is especially important, as no representative sample is not used in participant observation to
make a claim to generalizability.

As a case study most qualitative research are limited to the context they describe. Each community
therefore represents only those that are typical in terms of the set of factors which account for its level of
homicide. As Geertz (1973) suggested, most case studies produce local knowledge. Nonetheless the
information gained in one situation often has bearing on other situations, as humans do operate within
frames that are larger than local.

Triangulation is for the purpose of increasing the confidence level of research findings. Four types of
triangulation exist, all of which were relevant to this study. The first is methodological. This is the use of
more than two methods to investigate a particular phenomenon and is central to the approach of this
research. Obviously the issue of methodological triangulation is very critical for qualitative research. All
attempts have been made in this study to employ a series of methods that would ensure the highest degree
of validation. The aim was therefore not only to have different kinds of data produced, or to have one
method collect data in one setting that would be impossible for another method; equally important was the
matter of corroboration and refinement. Hence, in this research overlapping of data was not only
welcomed, but intended.

The second, data triangulation, is the search for a variety of data sources – different persons, time, and
settings. Whilst young men are usually the target of any study on homicide, all other age groups were
included in this study. Special attention was in fact given to the views of women. Quite often females are
more expressive than males. They are also quite often the witness of violence rather than the performer.
Interviewing women therefore aided in clarifying many of the issues discussed with young men. In the PLA
sessions and afterwards I spent much time with children and have made special mention of the times spent
with them in the community chapters. I have found their perspectives on community violence to be
untainted and honest. Quite often researchers are unable to understand a problem because their observation
was limited to a single setting and time. Living in the communities gave me a chance to observe problems
develop, as well as to assess the mechanisms for solving these problems.

The third, investigator triangulation, requires the use of more than one trained researcher to examine data.
Here PLA theoretically has an advantage in the sense that it utilizes a team and hence can reduce
subjectivity or researcher bias. The PLA team in Jamaica were trained and experienced in the method. The
team in London was trained only at the start of the fieldwork, as they had no experience in using the
method.

Finally, theoretical triangulation is based on the notion that the more theories that relate to a body of data
collected, the more credible a claim of confidence. While feuding relations is the main theory employed in
this thesis, I have had to draw on other theoretical frameworks to adapt the theory to post-independent
Jamaica. In summary several theories have been directly and indirectly tested against the findings of the
study.

You might also like