This paper examines the stationarity of beta coefficients, especially in regard to recent, major stock market trends. In addition to the usual correlation tests for stationarity, this paper describes a more direct method for testing the stationarity of portfolio betas. The method involves the use of paired t-tests which show separately the degree of stationarity for each portfolio beta. In the process of testing for stationarity, the portfolio betas also are adjusted for measurement error using a formulation suggested by Blume [3].