Origin seems to calculate the two sample t-test assuming equal variances for the populations, and it does not seem to test if this assumption is supported by the samples. If this is true, I suggest it should at least be clearly stated somewhere, or better, t-test should be adjusted for this case (i.e. Welsh's correction.
Thanks for your comments. Origin 7 documentaion for the Two Sample t-Test starts as folows.
"Two Sample t-Tests A two sample t-Test can be employed to test whether or not two population means are equal (i.e. whether or not their difference is 0). The two sample t-Test is performed on two sample data sets that are assumed to have been drawn from populations that follow a normal distribution with constant variance."
You are correct that the Origin Two Sample t-Test also assumes that the populations variances are equal and I will update our documention to include that fact.
While the Two Sample t-Test does not offer tests for equal variance we do include this feature in the Origin 7 One-Way ANOVA tool. We support both the Levene and the Brown-Forsythe tests for equal variance. It might also be useful to have a separate tool for testing equal variance or maybe just add it to the Two Sample t-Test and I will request that we add this feature.
Finally, I will also request that we also add Welch's correction for non-equal variances.
t-Test
Posted - 04/09/2003 : 04:35:11 AM
