This article discusses the effect of an autocorrelated error structure on the interpretation of traditional significance tests, especially the t-test and R2 measure It emphasizes first-order serial correlation, a common and often serious problem that researchers using time series data may...