Abstract
This paper examines the importance of time aggregation in causality testing. We find that temporal aggregates are between two and ten times more unlikely to detect a true causal relationship than are systematic sampled aggregates over short aggregation spans.