Findings, implications and lessons learned
Preliminary or ´pilot´ studies
Another lesson to be taken from this case study is that
the research process is often an iterative one built on a series of studies, one
following another. Preliminary or ´pilot´ studies can often be undertaken to
test an idea or investigate some fact before proceeding with the next step. In
order to minimise costs the researcher will wish to use as few experimental or
sampling units as possible. The danger is that if studies are too small the data
will not render themselves suitable for statistical analysis and hence the
results will be difficult to interpret.
The design of preliminary investigations is as important
as the design of main studies and it is necessary to ensure that sample sizes
can allow conclusions to be made that can justify decisions taken for the next
phase of the research. The biometrician often finds himself/herself advising on
the design of pilot studies and needs to ensure that he understands the goals
that the researcher has in mind and how the results from a current study will
lead to the next phase of the research. Sometimes it is possible to plan a study
which in itself may be too small to merit analysis on its own but, if the
results look promising, the study can be replicated and the two
studies analysed and reported together.
|