I will present a modeling methodology for the interpretation of the variability properties of AGN populations detected in extragalactic survey fields. The model is built upon empirical relations and is designed to account for the selection effects and biases that affect observational measurements. I will demonstrate the predictive power of the model by comparing in a forward manner with observational measurements of the ensemble excess variance of AGN in the Chandra Deep Field South. I will show that the proposed modeling approach has the potential to jointly constrain both models of the stochastic flux variations of AGN and the underlying Black-hole mass vs stellar mass relation of the population. Future prospects will also be discussed.