Forest management aimed at promoting climate change resilience hinges on accurately quantifying the relationship between tree growth and climate. Aggregation is commonly used to upscale individual tree response (e.g., ring-width time series) to broader scales of inference, prediction, and decision-making. This approach assumes non-climatic drivers of tree growth vary randomly across a population such that their effects cancel out with replication and climate emerges as a strong predictor of aggregate tree growth. Aggregation, however, can affect statistical relationships, especially if there are spatial dependencies in the covariance of response and predictor variables. It can also lead to a loss of fine-scale information important for understanding underlying processes. In these cases, ascribing individual or fine-scale behavior from statistical relationships derived at a larger aggregate level can lead to flawed interpretation. This problem, known as an ecological fallacy, is well recognized in statistical and social sciences but remains largely unaddressed in dendroecology. We examine the effect of aggregation scale on climate-growth relationships quantified from Douglas-fir tree-ring data collected across the southwest United States. We discuss ways in which scale and aggregation can affect our understanding of forest resilience to climate change, and subsequently, management decision-making.