Multilevel Research

Fundamentals

Multilevel investigations simultaneously consider the relationships between predictors and criteria variables at two or more levels of analysis. Although most applications consider two levels of analysis, other than the complexity of doing so, there is nothing to preclude one from embracing three or more levels of analysis. The variables included within each level may be different or similar to one another across levels. Situations where variables are conceptually comparable across levels (e.g., efficacy, cooperation, competitiveness) are referred to as isomorphism. Instances where the relationships linking variables within levels are comparable to similar ones across different levels are referred to as homologous. Cross-level relationships describe instances, whereas predictor variables from one or more higher levels exert influence on lower level processes or outcomes.

Given the inherent nesting arrangement of multilevel models, the degree of linkage across levels is referred to as bond strength. The general rule is that the relative strength of bonds across levels of phenomena increases with proximity and inclusion, and decreases with distance and independence. For example, the notion of proximity suggests that individuals are most likely to be influenced by their personal attributes, followed by team-level variables, and then by variables from more distant levels, such as industry characteristics. At issue is that, all else being equal, variables residing within a given level are likely to have the strongest bonds, followed by forces from adjacent levels, and to a lesser extent influences from more distant or removed levels from the focal level. Whereas this general pattern is likely to be widely applicable, it does not preclude the possibility of a distant variable exerting a more direct or immediate effect should a theory warrant. For example, individuals may be directly susceptible to events occurring in the far-removed performance environment.

The notion of inclusion refers to how neatly the level entities are hierarchically arranged. To the extent that lower levels are wholly contained in higher level units, bond strength increases. To the extent that lower level entities bridge higher level collectives (e.g., team memberships that span organizations or organizations that are members of multiple strategic groups), the bond strength across levels weakens. Embeddedness describes how lower level phenomena are aligned with higher level factors and processes, such that greater alignment generates stronger bonds or inclusion across levels. The idea here is that higher level variables serve as a context or constraint within which lower level phenomena operate. Finally, entrainment refers to the rhythm, cycles, synchronicity, and pacing of organizational phenomena. As a general rule, the rate at which higher level phenomena (e.g., team cohesion) evolve and change is slower than those of lower level phenomena (e.g., individuals’ motivations and attitudes).

...