One of the most widely applicable quantitative methods in research synthesis is the analysis of p values. The purpose of this study is to examine the properties of functions of the significance value to develop a better understanding of several quantitative analyses based on p values in meta-analysis. I examine the summaries, called diffuse test(proposed and promoted by Rosenthal and Rubin). That test is based on inverse-normal-transformed significance values from the sample studies. I described the hypotheses comparing the results of studies tested by the diffuse test, and derived asymptotic sampling distributions of the test. Since the small sample behavior of the test depends on the accuracy of asymptotic distributions of the test, I did a simulation study to see how the test behaves in finite samples, comparing its derived theoretical moments and distributions to empirical(simulated) moments and distributions. A Pearson chi-square statistic was applied to test the goodness of fit between the asymptotic distributions and the simulated distributions of the diffuse test statistic. The distribution of the diffuse test statistic was well approximated by a modified chi-square distribution (the theoretical distribution) when one-sample t statistics led to the diffuse test statistics. Therefore the power values obtained from this asymptotic distribution should be useful for the diffuse test.