Appropriate tests for comparing 2 normally distributed datasets (ttest?)


It's been a while since my A-level statistics course so I just want to make sure I'm doing the appropriate test for my data. I'm a PhD student working in X-ray crystallography.

My situation is that I have 2 populations of data, of varying sizes, in one population the samples have been subjected to a specific treatment and in the other they have not (control group). I want to be able to show that the mean values of various statistics produced from my data processing software are as good in the test group as they are in the control.

I've performed a two tailed unpaired student’s t test, assuming normal distribution and equal variance. The t-values I got were lower than the critical t-value from the tables at 0.05 so I accepted the null hypothesis that the means are the same. From my understanding this means I can accept with a 95% confidence interval that the means between the 2 populations are statistically similar.

Where I'm getting a little unsure is with the confidence interval because obviously the critical t-value gets higher when going from 0.05 to 0.0001 so obviously my t-value is still going to be lower. Does this mean I can assume a 99.99% confidence interval that my means are the same?

Any help and advice would be greatly appreciated.


The first thing that should be known that in the study the first thing to be determined is the level of trust. we can not simply change the level of trust in the research.
second, the confidence level also depends on the research conducted and how much researchers willing to take risks.


Less is more. Stay pure. Stay poor.
You probably don't have to assume equal variances, you can establish this with a test.

Typically before testing, you establish a level of significance for your test and/or confidence interval. You then stick with that level regardless of how significant your results are revealed to be.