given a sample of 12 values,

50 90 50 60

60 80 50 50

70 40 45 70

and a claimed mean on 48

How would you calculate the 99% confidence interval?

do you use [x - t(n-1)a/2 x (s/sqrt(n)] < u < [x - t(n-1)a/2 x (s/sqrt(n)]

where x is the mean, t(n-1)a/2 is from a table, s is the sample std dev, n=12

Also what hypothesis test would one use to determine if the real average is less than the claimed?