When rhizosphere complexity is too important for constraining into a single causality pattern: a causal inference methodology

Structural Equation Modeling is used in ecological studies to confirm pre-assumed multivariate causal relationships. However, the rhizosphere is a complex environment, and knowledge is not sufficiently consistent to propose unambiguous causal relationships to be tested. Using a Latent Variable Structural Equation Modeling framework, that aims to build and explore different causality patterns in rhizosphere environments, we designed an exploratory approach to detect causality patterns that are worth being investigated a posteriori and contribute to rhizosphere knowledge and applications. Grounded in statistical methods, exploration of the “causal space” is applicable to prioritize rhizosphere causality patterns that worth to be tested. Application of our framework to field studies is discussed. The term causal space is debated as a pioneer concept for causal inference in the rhizosphere. • Exploratory approach of structural equation modeling is different from confirmatory • Latent Variable SEM explores potential causality in rhizosphere environments. • "Causal space" concept aids in prioritizing rhizosphere causal relationships. • Latent Variable SEM narrowed down causality patterns from 38 to 5 validated models.