I would think that we wnat to test a proportion here.

I'm not sure if they want to disprove the claim or prove it's too high.

P is the true proportion of trains that arrive on time.

Based on n=64 and what seems to be an sample proportion of successes

There seems to be 53 of the 64 trains on time, in this sample.

I would go with =.85" alt="H_0=.85" /> vs. \ne .85" alt="H_a\ne .85" />

But one can argue (mainly after looking at the data, which is cheating), that < .85" alt="H_a< .85" />

The test stat would be

Another weird thing is the statement that the st deviation is 5 percent.

I did calculate the approximate st deviation as