Question 1086348
It's not clear what one wishes to know, so strictly speaking,none of the above.  If one wishes a confidence interval of the population, one would classically use a z-test, because the sample is normally distributed and the size is over the "magic" 30, which I don't agree with (but many do).  When I construct a confidence interval, for example, I will use a t-test if I am estimating the standard deviation from the sample.  With large sample sizes, the t will approach the z, and it makes no difference, but it forces one to realize what is being done rather than just using z because the sample size is large.  The distribution is a t-distribution.  

The answer is likely normal z distribution (z-needs to be stated) because of sample size, but the actual one is a t which at these degrees of freedom (199) is almost the same.