With the rise of interest in endocrine disruption by the scientific community and the general public, a clear and scientifically sound approach to evaluating potential endocrine disrupting properties is critical. Advances in in vitro and high throughput technologies have provided a wealth of data which provide insight into the EATS (endocrine, androgen, thyroid, steroidogenesis) modalities. A current regulatory challenge is the utilization of multiple data streams to arrive at a conclusion regarding endocrine activity. An additional challenge is that many chemicals have not been tested using existing in vitro approaches, thus there is a need for additional tools to provide information regarding endocrine disrupting properties. In silico modeling offers a robust, widely-applicable, cost- and time-effective approach to fill these data gaps. This work set out to place the development of an in silico predictive model into the relevant biological context to ensure the appropriate interpretation of model outputs and builds off an accompanying presentation which highlights the technical robustness of 18 models developed to predict endocrine bioactivity. Of the 18 ToxCast assays for estrogenicity, this work focuses on the five assays that describe transcription and protein production and weighs less heavily those assays that describe earlier aspects of the adverse outcome pathway, such as receptor binding. This weighting approach was validated by a selection of reference chemicals designated by the EPA and derived from in vivo and in vitro studies. Compared to the EPA in vivo reference chemicals, the weighted scheme of in silico model outputs predicts with 93% concordance. An alternate aggregation approach has gained wide acceptance in the regulatory community, but lacks the ability to evaluate novel chemicals; the approach described herein is thus advantageous in that it can predict activity for chemicals that are not already associated with in vitro data. This research provides the incorporation of adverse outcome pathway methodologies into the interpretation/aggregation of in vitro and in silico data sources and can be used to predict estrogen bioactivity for novel chemicals. Such frameworks are useful for the condensation of many data points into a single output or conclusion for regulatory decision-making purposes.