The term “educational effectiveness” designates causal models of educational outcomes that may or may not contain school-level variables. The term “school effectiveness models” is used in the more restricted sense of outcome-oriented models that explicitly contain school-level variables. These models are categorized to a context-input-process-output structure, are multi-level, recognize causal chains, and sometimes include feedback loops. In this chapter the advantages of comprehensive vs. partial school effectiveness models are discussed. Furthermore alternative interpretations of the idea that higher-level variables some way facilitate lower-level variables in the production of educational outcomes are given. The debate concerns: (i) additive vs. interactive models, (ii) direct vs. indirect models, (iii) contextual vs. genuine effects, (iv) additive vs. synergetic interpretations, and (v) recursive vs. non-recursive models. In this chapter the results of empirical research are presented to test the first four completing interpretations. The results indicate that the indirect model is the least productive. The synergetic model looks promising, but more simple interactive models account for almost the same percentage of variance in students outcomes. The actual problem might be that the most likely school effectiveness models are indirect, non-recursive, contextual, synergetic, and interactive at the same time. To avoid complexities in future research decompositions are advocated. And therefore it is contended that organizational variables of interest first be examined on their possible alternative interpretation of cross-level facilitation, before formulating the model.