T vs Z Test

Anyone have some tips/tricks for remembering this?

When Sample is… Small N<30 Large 30>N

Normal w/ knows Var Z Z

Normal w/ Unk Var T T or Z (T better)

Non normal w/ knows Var N/A Z

Non normal w/ Unk Var N/A T or Z (T better)

  1. known var -> Z 2) unknown var -> T

EXCEPT

  1. for large samples can always use Z (but T is more conservative and therefore preferred)

  2. Nothing works for small non-normal samples