Question 1040001
The confidence interval is larger with smaller samples. We are less confident of a result when the sample size is 2 compared to a sample size of 1000.  This is because the standard error, which is one part of the confidence interval, is s (or sigma)/sqrt(n).  The larger (n), the smaller the standard error and the narrower the confidence interval.
To narrow the confidence interval, one needs either a larger sample or less confidence.
If one wants to decrease the sample size, the result will be an answer with less confidence, all other things remaining the same.