EPTSo, I have 3 animals that I am testing a certain "characteristic" in. I have 3 trials for each of the "characteristic", which results in a number. I am also testing the animal's change in the "characteristic" under 4 different "conditions".
I tested my experiment, 36 different numbers, all organized in a chart. Now, I want to prove that the means of the numbers are statistically significant, and hopefully also mix in some stuff about the variance of the data.
Comments are appreciated. Answers or suggestions to the question are greatly appreciated. Detailed answers or step-by-step methods of how to do it are like "Ok,, how much money do you want me to mail to you" style appreciation.
kkthx.
Yeah, I planned to use an ANOVA test, but the only problem is, the variances of the population is different. It's been a while since I took stats, but if my sample population variance is different, can I still assume the population variance is the same? fyi, I have the same sample size for each different condition. Another question I have is in terms of the number of samples (n). Since I'm doing 3 trials per animal, should I average the three trials and count my n = 3? Or should I count the all the trials as separate "n"s. The first one doesn't let me assume normality of the sample (do I need to do this for normality?) because n < 30 so the central limit theorem doesn't apply. On the other hand, the second way lets me get closer to asuming normality, but seems to break one of the assumptions in an ANOVA test, the independence in the selection of data.
Thanks, Micronesia.
