The bootstrap never increases the effective sample size.
Having only 300 controls is limiting precision- and power-wise and may make the selection process of controls and their comparability fo cases a bit more problematic.
@f2harrell I don’t know if i expressed the problem clearly. I have 5 subgroups so i will make 5 comparisons each 300 cases vs 300 controls. Is it a must that the 300 controls used must be different at each comparison ? or I should list it as a limitation in case of proceeding with sub-group analysis ?
It is common to re-use controls in such settings. The real challenge is showing that the controls are otherwise comparable for any of the case groups. Note that 300 is on the low side. Best to have 3x or at least 2x as many controls as in your largest comparison group. After 3x there are diminishing returns precision and power-wise.