ABSTRACT
Ideally, of course, different methods within a particular research programme would be differentially weighted in terms of their centrality to the hypothesis or research objective at stake; and the degree of extraneous variation that each is known to possess. The general problem of ‘objectivity’ that is encountered in any form of social observation, and particularly in the case of participant observation, concerns the degree to which the process of observation itself can be considered to be free of bias and distortion. Both participant and non-participant forms of observation are amenable to more structured research designs than many of those which are currently employed. In the foregoing discussion the authors have conceptualized observation, in its participant and non-participant forms, in terms of the surveillance by individuals of overt behaviour. When social observation involves the use of formal settings and the careful manipulation of independent variables the method being employed is better described as experimentation.
