Which test is appropriate?
Message posted by Christina on January 19, 2001 at 12:00 AM (ET)

I have a statistical question that I have not been able to answer definitely...I have two relatively small groups of normally-distributed data (A (n=6) and B (n=7)). I performed a t-test to compare them, and, as hypothesized A is greater than B. My problem is that I also expect the ratio of A/B to be equal to (or close to) a particular ratio X. The ratio (mean A/mean B) is, in fact, fairly close to X. However, I want to use statistical analysis to explore the robustness of this statement (i.e. how "fairly close" is A/B to X). I had thought to compute every possible ratio of A/B (giving me a total n=42) and try to answer the question of whether or not they could come from a population with mean=X using a t-test. I am not sure though that this is appropriate, since I don't think my (n=42) data is completely independent. How can I perform this analysis?

READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Which test is appropriate?
Message posted by Loreen on January 22, 2001 at 12:00 AM (ET)