This study investigates the relationship between uncertainty and investment using U.K. data at different levels of aggregation. Motivated by a comparative econometric analysis using a firm-level panel and aggregate time-series data, we analyze the implications of aggregating nonlinear microeconomic processes. Replicating firm-level evidence that uncertainty influences investment dynamics proves to be challenging. Even using perfectly consistent data sources, this requires both exact aggregation of the underlying micro equations, and controlling for the unobserved influences on investment that are commonly subsumed into time dummies in panel studies. These conditions are unlikely to be satisfied in most aggregate econometric studies.