observation, ticking off events at regular intervals and according to a set list of
categories, was not deemed appropriate as it was thought that such an
approach would do less than justice in capturing the 'totality' of the primary
school classroom situation. Lesson observations were to be naturalistic and the
observer did not start with any specific preconceived notions as to the variables
to be studied or with 'a hypothesis to test' (Long, 1980: 21). No specific
schedule such as discussed in Long (1980), Mitchell, Parkinson & Johnstone
(1981), Mitchell (1985) or Chaudron (1988) were to be used and to begin with it
was assumed that everything might be relevant for later analysis.
Although informal observation does not'in various respects satisfy the canons
of scientific demonstration' a holistic approach was chosen as it enabled the
researcher to look at a set of complex interactions and allowed for the
recognition Offeatures that formal instruments might not detect as they could
not have been anticipated during the design stage (Barrow, 1984: 244).
Highly structured observation, which might have yielded more quantifiable data,
inevitably would have led to non-participation and would not have allowed the
observer to form relationships with the children. Forming relationships with the
children was deemed to be crucial if they were to behave 'in their normal way'.
Hammersley & Atkinson (1991) state that researchers and methodologies
inevitably affect the setting, the participants, the data and the interpretation of
the data. Lesson observations were to follow the normal pattern of procedures
as much as this was possible and no attempt was to be made to interfere or
259