Opportunities for general, synthetic understanding that emerge from collaborative scientific networks

Background/Question/Methods

Collaborative ecological research networks generate data across many locations that allow novel insights into the generality and site- or region-scale contingencies of ecological phenomena and processes. This approach has led to significant new understanding of marine, terrestrial, and freshwater ecosystems, and provides a way to overcome the limitations of other approaches such as local-scale studies and meta-analyses for understanding general ecological rules. Despite the great potential for achieving consensus opinions from network efforts, generalizing across the sample remains a major challenge. Ecological systems are individually unique in many ways. This means the challenge is to match general theoretical notions with heterogeneous ensembles of study sites in ways that avoid oversimplification. This is essentially a “measurement problem” and needs explicit treatment.

Results/Conclusions

Addressing general hypotheses across a heterogeneous sample can be accomplished through the rigorous application of meta-model specification. Meta-modeling represents a general form of model construction that hypothesizes relationships among high-level categories of system attributes instead of considering all the specific attributes that might be considered. Meta-models based on general theoretical constructs/concepts provide a natural connection between ecological theory and hypothesis evaluation. Various approaches will be discussed in this presentation, but particular focus will be placed on the potential value of “composites”.  In simple terms, composite variables represent the combined influence of collections of specific variables in a theoretical construct. Contrasts between composites, latent variables, and indices will be a point of emphasis in the examples presented.