separately, and that specialist and/or different amounts of
Goa 1 s Aims and Objectives !
5.2 Th© Research Methodology
The adoption of either a positivist or an interpretative
perspective opens up various possibilities for insight but also excludes others; as Mangham (1979, p.3) remarks, 'any way of seeing is simultaneously a way of not seeing*. However, although there is a definite disjuncture between the positivist and
interpretative methodologies, the approaches used in one may overlap to some extent those employed within the other. Indeed, the approaches as outlined above represent two extremes between which there is substantial middle ground containing elements of both perspectives in varying degrees. Rather than seeing the methods as mutually exclusive opposites, therefore, they can be usefully combined to support and supplement one another. Thus as Silverman (1985) argues:
The question remains whether any bridging position is possible between these two (positivist and interpretative) apparently incompatible perpectives ... I shall argue that
... the internalist (interpretative) concern with form and universality and the externalist (positivist) commitment to content and variability are complementary rather than
contradictory (p. 170) .... Neither kind of data are
intrinsically better than the other; everything depends on the method of analysis (Silverman, 1985, p. 156).
Partly in agreement with this analysis by Silverman and partly to make the best use of the strengths of both methodologies relative to particular problems under investigation, more researches have, in recent years, drawn from and utilised the methods of both
paradigms (eg Lancaster, 1988). This is the position adopted in this study. In making that attempt it is recognised that:
... the kind of research which seeks to bridge the gap between the two standpoints comes under attack from both sides. Research in this area is regarded by the objective purists as woolly and unscientific, and by those adopting a subjective stance as too concerned with essentially
artificial theoretical concepts far removed from the
day-to-day realities of the actors in the situation. To hold the middle ground means accepting that both subjective and objective viewpoints are important in interpreting a social situation (Hewton, 1986, p. 172).
In attempting to draw from both traditions here, it is recognised that there is a danger that neither will be as fully developed as an analysis rooted within one theoretical tradition only.
According to Shipman (1985b), research methods, the data
collected and the theories which are verified or generated are inseparable. The choice of an appropriate methodology therefore depends not only upon the nature of the problem under
investigation, in this case the management of resources in
education, but more importantly on the objectives of the research and relevant research criteria. Based on this consideration, a combination of positive and interpretative approaches is adopted here, although weighted towards the interpretative. For example, no hypotheses are formulated for testing and no 'interventions’
are made, or variables 'manipulated* either experimentally or statistically. For instance, hypotheses contending that
allocative choices in schools reflect a rational more than a political perspective, or vice versa, are not formulated.
Rather, the various implementations of TVEI in selected LEAs and schools are examined in their natural settings using the case study method. At the same time, analysis is based on both quantitative and qualitative data.
In developing the methodology of the research, the importance of the criteria of reliability and validity are recognised. Most text and handbooks on educational research discuss various kinds of reliability and validity as well as possible ways of
addressing them (eg Denzin, 1970; Tuchman, 1972). Essentially, reliability or internal validity relates to the extent to which the data used for analysis are not only representative of the actual state of affairs, but also relevant to the problem under
investigation. Similarly, external validity relates to the
extent to which the findings of a study are applicable to similar situations; that is, the extent to which findings are
generalisable. Of course, of the two criteria, internal validity is the sine qua non because external validity and therefore
generalisation is only possible when findings are based on
reliable data. Thus, most of the potential threats to internal validity are also relevant to external validity.
Methodologically, internal and external validity in this study are addressed at several levels of the research process. Thus account has been taken of both criteria in the development of the
theoretical model, the selection of the research method, the determination of the research design and the collection and analysis of data. The details of these will be given in the
relevant sections; for now, it will only be necessary to outline the general approach to the criteria.
With respect to reliability, in addition to possible confounding effects of history and maturation, other potential threats to
internal validity identified in literature on research methodology include the possible lack of instrumentation
reliability, reactivity (the effect of repeated data collection), bias in sample selection and sample attrition. The adoption of a case study method means that test-retest and split-half types of reliability are neither relevant nor necessary. Reliability in this study is therefore concerned with issues relating to
instrument and data collection techniques as well as with the interpretation of data. For example, it is necessary to take account of the possibility that a particular occurence can result from different causes or be subject to more than one
interpretation. This has been largely addressed, as discussed in the last chapter, by analysing resource management at different decision stages (objectives, processes and outcomes) and, as will be discussed shortly, by interviewing different groups of people
(at both LEA and school levels).
The internal validity of the research is further addressed by the time horizon of the research data. In relation to internal
validity, Lancaster (1988) notes that:
While cross-sectional studies using, for example, analysis of variance methods can, in principle, help to separate out the confounding effects of various independent variables,
longitudinal studies using interpretative methods offer more persuasive possibilities of addressing the problem of
simultaneity and of untangling causes and effects by using inter-tempora1 variations to facilitate causal inference (Lancaster, 1988, p. 115).
The relevant terms here are 'cross-sectional* and 'longitudinal* studies. Although this reserch is mainly cross-sectional since
data is focused mainly on one academic year, it also includes a longitudinal element because, as will become clearer in due
course, data relating to some activities were collected over the TVEI pilot (3-4 years) in the schools. In addition to
illuminating the evolution of management concerns over the
period, such data also help to explain processes by isolating the causes of some actions, thereby enhancing the validity of
f indings.
While the case study method is not particularly problematic in terms of internal validity, the same cannot be said of the method with respect to generalisation or external validity. While the
limitations of qualitative and case study methods used in this study are recognised, its claim to validity and genera 1isabi1ty rests on two methodological and theoretical foundations. First, as already noted, attempts have been made (conceptual 1y,
methodologically and analytically) to increase the reliability of the research which is fundamental to any claim to validity.
Secondly, and more importantly, it is also argued that the claim to validity of this research rests largely on the validity of the theoretical model on which it is based. This argument finds
support in the logic of case studies as demonstrated by Mitchell (1983) and elaborated by Silverman (1985).
According to Mitchell, the logic of case studies is theoretically rather than statistically defined; thus, the aim is not to select a 'typical* case, but a deviant or compelling case. Based on this argument, Silverman (1985) points out that:
... in a case study, the analyst selects cases only because he believes they exhibit some general theoretical principles. His account’s claim to validity depends entirely on
demonstrating that the features he portrays in the case are representative not of a population but of this general principle .... inference from case studies cannot be
situations by logical inference based on the demonstrated power of our theoretical model to account for initially negative instances. The claim, therefore, is not to
representativeness but to faultless logic (Silverman, 1985, pp. 113/114).
This is the view held here. While recognising the limitations of case studies in general and the approaches used here in
particular, it is argued that the validity of the research is substantially addressed. Thus, the rational and political models have been selected largely because they represent management
strands which are most likely to be in evidence in organisations in general and schools in particular. In addition, as already noted, it is not hypothesised, for example, that the rational or political perspective is the dominant mode; instead, the focus is on explaining the actions of organisational participants based on the theoretical bases of both the rational and political models of management. In addition to limiting the tendency of the researcher to to look for 'only what he wants’ it is argued that this approach limits the particularity and enhances the
representativeness of findings.