## One tailed t test or CI calculations.....

Hello all,
I am working on an assignment and having a bit of an issue determining the actual use of the term "confidence interval". I am asked to use the given data, and a 95% confidence interval to evaluate whether a null or alternative hypothesis is correct. The null hypothesis is H0=mu>0.2, Ha= <=0.2 and I have a small sample size so I am using the t table, with df n-1

So, does this mean that I should be calculating the confidence interval using the equation x-bar +/- [ t * s/sqrt(n) ], or is this simply the use of tcalc and t crit to determine whether the mean falls below the level of the critical value, and that is the actual answer to the question.

To my understanding, the particular null hypothesis calls for a one-tailed test, yet the use of the term 'confidence interval' leads me to think I should be calculating the CI, which implies a two-tailed test.....which is most of the source of my confusion. I do understand there is a way to change the value of alpha to account for it being a one tailed equation, however I'm not sure that we are supposed to be going quite that far with this assignment.

Thanks in advance, I hope I'm clear on the issue I am having......I'm really not terribly good at this!