To examine the inconsistency and ceiling effect when using the visual analogue scale (VAS) in measuring labor pain, data from four repeated measures of the sensation of labor pain in the early active phase from a clinical trial were re-analyzed. That study examined the effect of music on the labor pain experienced by 55 primiparas, while 55 control cases underwent labor under normal conditions. Both groups were balanced for demographic and obstetric data and other pain confounding factors. The re-analysis shows a 5 to 13% inconsistency in both groups in reported pain between the perception of whether pain was changing or not compared to their previous rating on VAS. The control group had more cases of ceiling effect in their reports of labor pain than the music group but analysis with and without the 13 cases of ceiling effect showed similar results that the music group experienced less pain than the control group. In the analysis without ceiling effect cases, the homogeneity of variance improved. This study supports the premise that a strong design and large sample size helps to make the statistics robust to the violation assumption of homogeneity of variance.