then Simple random sampling is okay since weighting of different scientific disciplines is proportional to the number of studies in your sample, SRS guarantees getting an unbiased estimate ...), and if we (iii)