2002ApJ...580...12M

Two groups recently deduced the positive value for the cosmological constant, concluding at a high (≥99%) confidence level that the universe should be accelerating. This conclusion followed from the statistical analysis of dozens of high-redshift supernovae. In this paper this conclusion is discussed. From the conservative frequentist's point of view, the validity of the null hypothesis of the zero cosmological constant is tested by the classical statistical χ² test for the 60 supernovae listed in Perlmutter et al. This sample contains 42 objects discovered in the frame of the Supernova Cosmology Project and 18 low-redshift objects detected earlier. Excluding the event SN 1997O, which is doubtlessly an outlier, one obtains the following result: the probability for seeing a worse χ²–if the null hypothesis is true–is in the 5%-8% range, a value that does not indicate significant evidence against the null. Furthermore, if one excludes five possible outliers, as proposed by Perlmutter et al., then the sample of 54 supernovae is in excellent accordance with the null hypothesis. It also seems that supernovae from the High-z Supernova Search Team do not change the acceptance of the null hypothesis. This means that the rejection of the Einstein equations with zero cosmological constant–based on the supernova data alone–is still premature.