I want to calculate the sample size a for a clinical trial, I want to compare control and treatment results with a t-test alpha=0.0.5 power=0.9.

I dont have historical data= mean control, mean treatment or standard deviation because its a new trial. The only requirement for sample size beetwen this 2 groups its that the mean of treatment group its greater 20% from control group.

Acording whith this post:

Syntax interpretation for pwrss package in R (power analysis on proportions data)

Using pwr library in R I could get sample size for relative difference beetwen 2 groups.

library(pwr)

pwr.t.test(d= , sig.level = 0.05, power = .9, alternative = “greater”)

where:
d=effect size= (mu control- mu experiment)/sd

I understand that effect size can be expressed as the absolute or relative difference. For example, relative difference/reduction with drug intervention is 10/20 or 50% 0.5.

On the other hand, some tables show the relation on effect size and percentage from control vs experimental group. Effect size its different to relative difference:

My question is: I could input percent o increase of 20% bettween control vs experiment. as effect size?

pwr.t.test(d= , sig.level = 0.05, power = .9, alternative = “greater”)