Hi! I have a population size of 55 and a sample size of 48. we originally planned to use paired t-test to compare the scores of the same sample within two different timeframe. Should we use z-test because n>30? is there a paired z-test of one population? What should we use?

plus,

we have 25 survey questions using a 6-point likert scale. how should i take the standard dev? standard dev for every survey question and get the average of all 25 items? Or do i just get the standard dev of all raw answers? Im confused.