Quantitatively oriented researchers differ in how they report the results of tests of statistical significance. Some authors report statistically significant results as less than an alpha level (e.g., p <.05), while others report exact p‐values (p =.04) from statistical software readouts. This paper argues that three reporting errors are possible when p‐values are reported based on computer readouts. First, the p‐value of obtained data can never be equal to, or less than, zero. Thus, reporting that a finding is p =.000 is an error. Second, rounding down errors can lead to factual inaccuracies. For example, it is inaccurate to report a p =.052 as p <.05. Finally, p‐values for wrong‐direction effects with one‐tailed tests must always be p>.50. Awareness of these issues should facilitate accurate statistical reporting. © 2004, Taylor & Francis Group, LLC.