Characteristics of Research on Leadership in Education. An Occasional Paper.

Any attempt to derive conclusions from the literature regarding leadership in education is limited because much of the literature consists of untested opinion and most empirical studies are based on information from one school or one school district and are thus very limited. Most studies utilize variables that are unique to the particular study and/or are nontheoretical, thus providing little in the way of a framework for integrating findings from different studies. Generally, studies fail in specifying the conditions under which a relationship between variables is tested, thereby running the risk that an actual relationship is undetected; and investigators tend to use perceptions of involved persons rather than observations by a third party as sources of information regarding variables, thus making findings subject to the attitudes or the memory of participants in the organization. In many cases, information regarding both variables is obtained from the same respondents, thus inflating the obtained measure of relationship because of a response set. Finally, the models used for differentiating among behavior patterns provide for limited alternatives and frequently involve one specified behavior (with the only other option being the absence of the one specified), thus running the risk of not differentiating what are actually meaningfully different behaviors. (Author)