I used minitab to compare two means from samples of size 400 each. Firstly minitab gave a degree of Freedom of 358 rather than n1+n2-2=798. Moreover its computed mean difference fell into the 95% confidence interval of the mean difference. Yet it gave a verdict of significant difference at 5% level of significance? Why the contradictions?
Provide a copy of your Minitab output. The df=358 seems to be saying that Minitab does not recognize some of your data. I assume that you are trying to run an independent samples t-test and not a dependent samples t-test (aka paired t-text).
I tried to reproduce your results using an online t-test calculator. (See the attached JPEG file.) With the exception of the degrees of freedom, the online calculator results are essentially the same assuming some round-off error. I don't know where the "df=547" number is coming from.
I think some problem of steps programming in Minitab.Althogh it is not expected.Can't it be reported to Minitab for possible reason.?
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size