Understanding the factors driving population dynamics is critical for conservation efforts, enabling e.g. the implementation of effective recovery strategies. However, selecting the most effective method for identifying drivers of population change can be challenging, given the wide range of models available for this purpose. In this study, we employed a virtual ecologist approach to compare three methods: time-series (TS), species distribution model (SDM), and process-oriented model (POM), in terms of their ability to correctly identify factors contributing to population changes. While the first two approaches are commonly used in ecological studies in the presented context, the process-oriented method is a novel approach that integrates mechanistic components with correlative approach to identify drivers of population changes. We show that POM outperforms the other methods on all performance measures used, as it has the highest accuracy (0.88), sensitivity (0.84), and specificity (0.93). SDM has the medium (0.68), while TS has the lowest accuracy (0.50). Moreover, our results suggest that variable selection slightly improve the performance of suboptimal models (TS), while it significantly reduces the accuracy of other approaches (SDM and POM). Policy implications: Our results highlight the importance of incorporating more mechanistic aspects into species distribution modelling to better identify the causes of population changes. These findings may be applicable to a range of ecological and conservation studies.