Conference Agenda

Relevance & Research Question
Declining response rates (Groves, 2006) is a potential threat for survey validity, e.g., non-response bias (Groves & Peytcheva, 2008). While reminders have been reported to increase the response-rate in web surveys (Sammut et al., 2021), few studies have investigated the effects of number of reminders, and even fewer have examined specific content (Pedersen & Nielsen, 2016). This study applies psychological theories of social identity (Tajfel & Turner, 1979) and social influence through descriptive norms (Goldstein et al., 2008), to test if text interventions in reminders can increase response rates and decrease non-response bias.
Methods & Data

This is a 2x2 factorial experiment using the Swedish Citizen Panel, a non-commercial non-incentive web panel. Each group received a different intervention: (1) control group/no intervention (a general reminder), (2) social identity intervention (eliciting a sense of social identity by using phrasing such as “We need more responses from people like you”), (3) descriptive norm intervention (inducing a norm activation with a text like “Most people have already responded but we still need your answer”), and (4) a combined social identity intervention (2) and descriptive norm intervention (3).

Results

Results indicate a positive pattern for all three experimental conditions. The response rate after the first reminder was significantly higher in the combined social identity and descriptive norm intervention (18,8 %) compared with the control group (15,6 %). Sub-group analyses indicate that certain social groups are more sensitive to the experimental manipulations.

Added Value
This study adds to the literature of reminders as a tool to increase response rates, while also not adding any cost. Further it adds to the research of theories of social identity and social persuasion by exploring if these psychological mechanisms can be utilized in the context of survey reminders.

Relevance & Research Question

Text messages have often been used as a cost-effective way to remind sample persons to complete mailed paper-and-pencil and online questionnaires. However, in some countries, the likelihood of identifying phone numbers of sample persons have decreased rapidly. Fortuitously, digital mailboxes have been implemented in Sweden where government authorities and companies can send mail digitally instead of physically. According to the Agency of Digital Government (2024), more than 6 million Swedes used these digital mailboxes in 2023. The present study assessed the effect that reminding potential respondents by digital mailboxes instead of through text messages.

Methods & Data

The preregistered experiment was implemented in a self-administered mixed-mode survey (paper-and-pencil and web questionnaire) administered to a sample of 26,250 randomly selected individuals. Prior to being invited to complete the questionnaire, all sample persons were randomly assigned to one of two groups. One group of sample persons received reminders via four regular mailings and four texts, whereas the other respondents received reminders through four regular mailings, two through texts, and two through their digital mailbox. Per the preregistration, the study will assess the two reminder strategies in terms of response rates, nonresponse bias, and data quality.

Results

The data collection period started in August 2024 and was finalized at the beginning of January 2025. The results indicated that sending reminders through digital mailbox had a significantly positive effect on response rate.

Added Value

Relevance & Research Question

Nonresponse is a particularly salient issue in the context of business web surveys, as response rates for such surveys are on average lower than for personal web-surveys. Previous research in the field of business surveys has mainly focused on testing the utility of different contact and survey modes to increase response rates. The following study uses an experimental approach to test the implementation of an adapted and expanded design feature to increase motivation to complete the survey, originally proposed by Lewis et al. (2019). This is done by implementing a more interactive and targeted communication with late respondents, i.e. potential nonrespondents.

Methods & Data

Therefore, a randomly selected sample of late respondents (N=3'000 in total) to a well-established Swiss business web survey was selected and asked to indicate their primary reason for nonresponse at various stages in the survey design. Subsequently, they were randomly presented with one of three appeals: a targeted appeal addressing their previously stated reason for nonresponse, a generic general appeal addressing some of the most prevalent reasons for nonresponse, or a main appeal providing arguments for the nonresponse reason of "not time". After reading the appeal respondents had to indicate whether they wanted to proceed immediately to the main survey or postpone the decision to participate.

Results

As hypothesized, a targeted appeal demonstrated a more pronounced motivational effect compared to the two alternative appeal types in persuading respondents to proceed to the main survey. However, the motivational effect was limited insofar as it did not extend to a higher motivation to provide more complete data. Furthermore, more detailed analyses revealed that the conversion rates are neither independent of the type of nonresponse reason nor timing factors.

Added Value

The study demonstrates that the implementation of such an interactive feature is feasible and potentially worthwhile, even in the context of a challenging business survey. It shows the potential to motivate target respondents to participate, particularly at a later stage in the fieldwork period.