Taking your account as fact (I have no data), this does not entirely refute nikatwork. It may be that in past generations lactose intolerance generally developed earlier than it does now. Today's later intolerance developers are the stubborn residue in the gene pool after the tolerance gene became dominant amongst the children of earlier intolerance developers. The assumption that intolerance was not switched off, but slid back to ever later years could be tested in families of multiple living generations whose intolerance is in flux.