we are starting a survey with 2 stage cluster sampling approach. At first stage we select a fixed number of clusters from total population and at second stage, 15 households per cluster. We calculate the design weight for households by taking into account the probability of selection for each stage. For e.g. If we select 30 clusters out of a total of 300 clusters in population and then select 15 households in one cluster with 250 households, that would make the probability of selection for each household in that cluster to be (30/300)*(15/250). Can someone please tell me if this approach is valid?

many thanks