Survey researchers are always looking for new ways to contact respondents to allow for accurate and fast data collection. In recent years, the use of text messaging to contact potential respondents has become more common.
SSRS routinely conducts text surveys using the SSRS Text Message Panel. The Text Message Panel is a part of our probability-based SSRS Opinion Panel, consisting of panelists who have pre-consented to receive text messages from us.
Our Panel Uses Texting in Two Ways:
- Most surveys use a push-to-Web contact protocol in which panelists are emailed a link to an online survey. For these surveys, Text Message Panel members also receive a text with a link to the online survey, in addition to the standard email invitations used for all panelists. This increases response rates by providing an additional way for text-consented panelists to learn about and access the survey.
- Sometimes, for very quick-turnaround surveys, we use a live-text contact protocol in which survey questions are sent directly by text. Text-consented panelists answer questions one-by-one by simply texting us back, rather than having to click a link to an online survey.
In March 2025, we conducted two surveys that covered very similar topics and were launched within a few days of each other. One used the standard push-to-Web protocol and one used a live-text protocol. This provided a “natural experiment” to allow us to directly compare these two approaches.
Our Methods, Analytics, and Data Science (MADS) team compared the live-text approach to the push-to-Web approach, analyzing response rates and the representativeness of the responding sample. In addition, we looked at the impact of each protocol on panelists’ willingness to complete future surveys. Since panelists who stop responding to surveys need to be replaced with new recruits, it is important to understand whether changing the data collection protocol for a given survey affects their response patterns to the next survey they receive.
Here is What we Found:
- The push-to-Web survey obtained about twice the response rate as the live-text survey, partly because it was in the field for considerably longer (89 hours vs. 16 hours).
- However, live-text responses came in much faster. Most live-text responses came in within 3 hours of launch. After 16 hours in field, the live-text response rate was still slightly higher than the push-to-Web response rate.
- Despite the differing response rates, respondents to both surveys were demographically very similar, with the live-text survey having slightly better representation of adults without a college degree.
- Live-text respondents were slightly less likely to complete one of their next few push-to-Web surveys.
Overall, we think these results are positive for the use of live-text surveys as a means of collecting very quick-turnaround data from a probability panel. When it is critical to complete a survey in less than a day—for example, when collecting point-in-time data on reactions to breaking news events—live-text surveys perform quite well. However, given the differences in response rates, and the small but noticeable impact on subsequent panelist engagement, we continue to recommend the push-to-Web protocol as the default when a longer field period is available.
