Current Research

What social media may reveal that focus groups may not

Project leads: Rebecca Dolgin

Anyone, anywhere, anytime can join a conversation, broadcast an opinion, or express their feelings, fears, and beliefs on social media. The fact that social media is not constrained by geography, status, or often even invitation makes it a potentially rich source for examining public opinion. Not only is it a less costly way to measure public opinion, but it can also provide more up-to-date information and tap the thinking of members of the public who would not necessarily agree to participate in sample surveys or formally constituted focus groups, potentially revealing new and noteworthy insights. For our study, we examined the extent to which social media posts reveal the same insights that come up in focus groups; how much the two sources overlap; and whether new or more nuanced insights be learned through social media posts.

We assembled a corpus of 17,497 tweets about the US Census collected before and during administration of the 2020 Decennial Census. A set of MTurk workers (n=19) was then asked to read samples of tweets from that corpus for 20 minutes and to characterize what they learned about reasons that members of the public might not participate or were against others participating in the Census. This method confirmed themes about barriers to participation in the 2020 Census–e.g., distrust in government–uncovered by Census Bureau focus groups that were themselves constituted with members of the public unlikely to participate in Census surveys. It also identified themes and nuances that expand on what was learned from those focus groups–for example, concerns about how inaccurate data or counting unfairly might affect US congressional representation, as well as timely concerns (e.g. about COVID transmission) that could not have been identified in focus groups carried out in an earlier moment. The findings and analyses of tweet-readers’ strategies point to new practical ways in which samples of social media posts read closely by analysts may uncover thinking from members of the public who might not participate in formal quantitative or qualitative data collection methods.

Interested in learning more? Contact us here.

What is it about live video interaction that makes some people uncomfortable?

This research examined reasons people gave for feeling discomfort answering a sensitive question in live video relative to other modes (in person, phone, web, prerecorded video) in a hypothetical standardized survey interview. Coding of open-ended explanations revealed reasons that, to our knowledge, have not been discussed in previous literature on survey respondents’ concerns in answering sensitive questions in other survey modes, and which are likely to be video-specific. For example, respondents reported finding the same features of live video–e.g.,social copresence with the interviewer, seeing and being seen, a sense of privacy and anonymity–as leading to greater discomfort or greater comfort.