Values of the log-likelihood statistic |$\Pi _S$|⁠, defined in equation (10) based on the null distribution of the power entropy, |$S(l)$|⁠, are listed in second column for various data sets listed in first column. The statistic values mentioned against different data releases were evaluated using the distribution obtained from a corresponding simulation ensemble of 1000 mocks. The p-values of |$\Pi _S$| thus obtained are given in the third column. The likelihood-based p-values are smaller than those of Table 1, which are evaluated with a cut-off of |$p \le 0.05$|⁠. This is evidence that using a hard cut-off on assessing multiple p-values tends to create a bias against detecting a signal.