This paper uses a survey data set that includes a large cross-section of currencies to test for the rationality of expectations in the foreign exchange market. Two standard tests are applied to the data. First, a test to determine if investors make systematic forecast errors. Second, a test to determine if investors use all available information efficiently when they make their forecasts. This paper's data set involves a more recent time period which saw more quiescent exchange rate changes than the 1980s. This paper finds that participants in the foreign exchange market do make systematic forecast errors and do not always use all available information when forecasting. This conclusion is similar to past research but is now extended to a broader range of currencies and more recent time period.