A medical researcher gets a P-value of 0.056 for testing H0: μ = 0 against Ha: μ ≠ 0. Since he believes that the true mean is positive and is worried that his favorite journal will not publish the results because they are not “significant at the 0.05 level,” he instead reports in his article the P-value of 0.028 for Ha: μ > 0. Explain what is wrong with
a. Reporting the one-sided P-value after seeing the data.
b. The journal’s policy to publish results only if they are statistically significant.