The article addresses the challenges in detecting and attributing biodiversity change due to the rapid global changes and the complex, non-independence structures in biodiversity datasets. It introduces a new statistical framework, the correlated effect model, which accounts for hierarchical and correlative non-independence across time, space, and phylogeny. This model is applied to ten high-profile biodiversity datasets, revealing that existing approaches often underestimate trend uncertainty and misestimate trend direction. The correlated effect model shows no significant trends at 95% credible intervals, highlighting the highly uncertain status of biodiversity change. Despite this uncertainty, the model improves local-scale prediction accuracy, offering hope for estimating biodiversity change at policy-relevant scales. The study emphasizes the need for more rigorous methods to address spatial, temporal, and phylogenetic structures in biodiversity data, which could enhance the reliability of biodiversity trend inference and guide adaptive conservation responses.The article addresses the challenges in detecting and attributing biodiversity change due to the rapid global changes and the complex, non-independence structures in biodiversity datasets. It introduces a new statistical framework, the correlated effect model, which accounts for hierarchical and correlative non-independence across time, space, and phylogeny. This model is applied to ten high-profile biodiversity datasets, revealing that existing approaches often underestimate trend uncertainty and misestimate trend direction. The correlated effect model shows no significant trends at 95% credible intervals, highlighting the highly uncertain status of biodiversity change. Despite this uncertainty, the model improves local-scale prediction accuracy, offering hope for estimating biodiversity change at policy-relevant scales. The study emphasizes the need for more rigorous methods to address spatial, temporal, and phylogenetic structures in biodiversity data, which could enhance the reliability of biodiversity trend inference and guide adaptive conservation responses.