Evaluating Text Messaging as a Contact Strategy and Predictors of Text Message Consent: Evidence from the Census Household Panel

As data collection efforts face concerns about declining response rates and rising costs associated with participant recruitment and follow-up, further research is needed to understand how different contact strategies may benefit survey researchers confronting these challenges. Text message invitations present an opportunity for data collection efforts – especially longitudinal efforts with repeated contact – to adopt a potentially faster and less expensive contact strategy compared to other commonly-used contact strategies (e.g., mail, in-person visits, telephone).

This study extends the literature by using nationally representative data from a unique data source: the Census Household Panel (CHP). The CHP was a probability-based nationwide survey panel that regularly fielded surveys to panel members as monthly “topicals”. Once enrolled, panelists were invited to the topical surveys via email and text messages. Panelists must provide a mobile phone number and consent to receive text message invitations. This research uses paradata and survey responses from the May 2024 topical (i.e. Topical 7) with approximately 10,000 respondents to explore the impacts of text messaging. More specifically, we examine four central research questions: (1) How do text message reminders contribute to response in the topical 7 survey?; (2) Are there demographic differences in who clicks on the survey through text message invites versus email invites?; (3) What are the consent rates for receiving text messages?; and (4) Are there demographic differences in who consents to receive text messages? Insights from this research help extend the field’s knowledge and understanding of the impacts of sending text message invitations.

Findings from this research show that: (1) text message invitations appear to contribute to survey response by bolstering initial response rates in the hours immediately following receipt of these text messages; (2) there are demographic differences by age, race/ethnicity, education level, family income, and respondent type in the propensity of panelists to access a survey through text message invitation; (3) text message consent rates for Topical 7 of the CHP are comparable to the existing literature (i.e. ~50%); and (4) there are demographic differences by race/ethnicity, age, family income, educational attainment, marital status, region, and respondent type in panelists’ consent behaviors. In sum, this paper’s findings suggest that the use of text messages as a contact strategy is advantageous for reaching certain demographic groups and is a promising tool for increasing response rates to national data collection efforts.