Discussion

Nancy Bates has pioneered an important subfield in survey methods: hard-to-survey populations. The topic of the Bates article in this issue is the U.S. Census’ experience using social marketing to improve mail and web response rates for these populations. Populations can be hard to survey in many ways. The article is in line with a whole body of work Bates has developed on methods for asking about race, Hispanic origin, sexual orientation, and other characteristics, with a major focus on survey and census nonresponse. I was privileged to work with her on what became a major milestone in this field, an international conference in 2012 on hard-to-reach populations, and on the book that followed (Tourangeau et al. 2014). The conference was also the springboard for a special issue of the Journal of Official Statistics on the hard-to-reach (Willis et al. 2014). The U.S. Decennial Census is a unique vehicle for studying the role social marketing can play in improving response from hard-to-survey groups because of its very large scale and its mandate to count every single person living in the United States. Other large scale vehicles, such as the U.S. presidential elections or national campaigns to change health behaviors, do not have such an exacting goal or such broad scope. Sample surveys are much smaller in scope and lack the resources to explore social marketing. The capability to target small areas like census tracts or even individuals in a sample at very low cost has only just emerged. I work on surveys, not censuses, and for a contracting organization, not the U.S. Census. Most of my work has been on face-to-face surveys, not mail or mixed-mode like the Decennial Census. Nonetheless, there are many parallels in our work, and the program at the Census may become a model that brings social marketing into the forefront of survey methods.