Re: "Use of census-based aggregate variables to proxy for socioeconomic group: evidence from national samples"