The probability that a given result would be obtained, assuming that the null hypothesis were true. The misuses of statistical significance often outweigh its benefits (as shown for economics by McCloskey and Ziliak 1996, and for psychology by Cohen 1994 and Smith et al. 2000). Fisher (1925) claimed that significance testing is best used in conjunction with replication. However, statistical significance is useful in some aspects of forecasting, such as in determining whether to use a trend factor or whether to use seasonal factors, particularly when these involve small samples and high variation. When using statistical significance to test multiple hypotheses, such as a comparison of three or more forecasting methods, one should adjust the levels of significance (see For Researchers).