Assume that in a research paper you make one confidence interval for a mean. What is the objective probability that the true mean is within this interval?

Can you a valid t-test on data be done that is not Normally distributed? Explain, giving conditions and reasons.

In what sense is an ANOVA linear regression model related to a t-test that compares means from two treatment groups. Under what conditions would ANOVA be a more powerful option? Why - give a technical reason and the intuition behind it?