Thanks a lot Paul for your time and extended help!
I agree picking high values doesn't match with the word random in the page heading itself "Stratified Random Sampling".
So, I will try to understand your codes and try to make a fifth option for high values based on the proportionate random size. This infract will be duplicating the options, that's the reason I though to have checkbox option.
I think, instead of adding another option, I can keep the original 3 options and call the subs according to the checkbox checked or not.
Next step, I'm also thinking to pick samples only for the strata values selected from the list box.