99. "The Levels of Levels in Organizational Phenomena"

ABSTRACT

Increasingly organizational researchers are recognizing and accepting the multi-level, multi-issue, and processual nature of organizational phenomena. The problem many have is that their research methods and available constructs are essentially designed to deal with single-level, single-issue, and static variables. These are useful for many research questions but they need to be augmented by new methods in order to deal properly with the more complex reality of live organizational phenomena.

Despite the growing interest in multi-level research, there is no accepted agenda, theory, or method for conducting such work. One problem is the variety of levels referred to in this literature. This article presents a general representation of level relationships. Such a representation provides a unifying framework for conducting and comparing research and can define levels of the rigor of such research.

The article defines nine different types of levels that the author has personally employed in his multi-level research. These include three inclusion levels: structural inclusion, levels of task aggregation, and levels of interdependence. There are also six causal types of levels. These include: simple functions, level of derivation, task process type level, implicit/explicit phenomena, process frameworks, and causal "drill-downs."

All nine of these types of levels have the same abstract analytical form: they are representable as process frameworks. This finding raises the hope that using process frameworks to create the models can present comparisons across studies and development of appropriate tools for studying multi-level organizational phenomena.

The article also looks into the inadequacies of the usual variable models in lieu of their failure to cope with multi-level research. This is based on defining rigor for multi-level research, the philosophical analysis of L. Mohr, and the properties of process frameworks.

There may be other types of levels for which a relationship among or between them cannot be defined as a process framework. Any representation has it limitations. But, for now, process frameworks are arguably the best choice extant.