Sample size estimation for cross sectional studies- i am trying to explore where exactly we use an absolute precision and where we use relative precision.
It makes no difference. For example, if you calc sample size for a proportion with expected value =50%, with an absolute precision of 5% (absolute = in original units) or a relative precision of 10% (relative to expected, 10% of 50% = 5% absolute) you will get the same result (as long as you do not change confidence and N)