OLS-based unit root tests (e.g., Dickey-Fuller, Augmented DickeyFuller, and PhiHips-Perron tests) typically fail to reject the unit root null at conventional significance levels when applied to macroeconomic time series data. This has stimulated a large amount of research regarding the applied and theoretical econometric implications of unit root processes. However, these tests are known to have low power against trend stationary and near-unit-root alternatives, the leading alternative data generating processes for these data. Recently, Pantula, GonzalezFarias, and Fuller (1994) proposed a unit root test based upon the weighted-symmetric estimator of an autoregressive model developed by Park and Fuller (1993). Their simulation studies suggest that this is a more, powerful test than the OLS-based tests. In this paper we consider the practical implications of the new test by applying the Pantula, GonzalezFarias, and Fuller (PGF) test and the augmented Dickey-Fuller (ADF) test to the extended Nelson-Plosser data set.