Relevance & Research Question

To ensure continued survey participation and data quality among respondents, the panel survey landscape must adapt to the changing societal reality, especially in terms of mobility and digitalization. Computer-Assisted Web-Interviewing (CAWI) has been shown to be a useful self-administered and cost-effective survey mode. Nevertheless, in surveys with complex instruments, interviewer assistance may still be necessary. Given a growing prevalence of videoconferencing in many people’s lives, Computer-Assisted Live Video Interviewing (CALVI) has potential to be a solution.

Methods & Data

To develop a high-usability CALVI methodology and examine fitness-for-purpose of this mode across societal subgroups, we have gathered data on panel respondents’ willingness to try CALVI in the Socio-Economic Panel (SOEP), developed and pretested an implementation of CALVI together with infas Institute for Applied Social Science, and started fieldwork in an experiment comparing CALVI to online self-administered and offline face-to-face modes in the context of the SOEP Innovation Sample (SOEP-IS).

Results

To assess the potential of CALVI in household panel surveys in Germany, we introduced a hypothetical inquiry regarding respondents' willingness to be surveyed via videocalls in the SOEP 2022 data collection wave. Of the 22,549 respondents who provided a valid answer, 39% expressed their willingness to try CALVI. Based on the results, we designed a CALVI implementation trial experiment for the SOEP Innovation Sample. In initial pretests, 600 participants of an infas panel study were invited to try video interviews. In total, 73 of those made an appointment for the interview and 46 participated in the interview. On average, the interviews were 106 minutes long and were conducted by experienced and CALVI-trained infas interviewers. Our presentation will provide more in-depth results from the hypothetical scenario and the pretest as well as lessons learned from the experimental fieldwork.

Added Value

The results of our feasibility project will inform future implementations of CALVI, especially with regard to usability for both respondents and interviewers as well as fitness-for purpose of the new mode for different population subgroups. Furthermore, our results will be used to develop targeted multi-mode strategies for data collection in SOEP-IS.

Relevance & Research Question

The Internet of Things (IoT) is a rapidly evolving technology, bridging the physical and virtual worlds (Zhang, 2021). By integrating Internet-connected devices supported by hardware or software, IoT enables the emergence of smart devices with communication capabilities. These capabilities create a new form of interaction between humans and computers, enabling new possibilities in survey research. Smart speakers could conduct interviews more cost-efficient than face-to-face interviews and potentially improve data quality compared to web surveys. However, it is unclear whether people are willing to particpate in interviews administered by smart speakers. This study examines socio-demographic factors associated with the willingness to participate in such interviews.

Methods & Data

Data was collected from two non-probability online access panels in the U.S. in Spring 2024 and a similar online panel in Germany in November 2024 with samples representing the general English-speaking U.S. population aged 18 and older and the respective German counterpart. Each country had around 2,000 respondents. The analysis included socio-demographic characteristics like gender, age, ethnicity, education, income, and metropolitan status. Additionally, participants’ attitude on the importance of surveys were considered to explore factors influencing willingness to participate.

Results

In the U.S., 25% of repondents are “very likely” to participate in a survey conducted by a smart speaker. Willingness was significantly influenced by factors such as age, ethnicity, gender and attitudes towards survey importance. Younger respondents, males, African Americans and people who placed higher importance on surveys were significantly more likely to report willingness to participate. Analyses of the German data is ongoing to compare the willingness and the factors influencing the willingness between the two countries.

Added Value

This study underscores the selective appeal of smart speaker surveys, suggesting that while not universally preferred, they may reach traditionally underrepresented groups. Smart speaker surveys can extend survey methodologies for technology-affine participants in mixed-mode designs. Although current acceptance varies across demographics, investment in this approach is advisable, as its applicability is expected to broaden with the advancing IoT technology.

Relevance & Research Question

Many surveys in the UK are transitioning to online data collection. However, long surveys or those involving complex elements, such as data linkage consent, cognitive assessments, and sensitive questions can be difficult to move to online self-administered data collection. As a result, in the UK several surveys explored online live video interviewing (VI), which represents a visual video-mediated, remotely interviewer-administered, and computerised survey mode.

This presentation investigates the use of VI, focusing on opportunities and barriers, and collating evidence from seven major social surveys in the UK, with an emphasis on longitudinal surveys and the collection of complex data. The research questions are:

1. What is the uptake of VI (response rates)?
2. What are the characteristics of those that respond via VI (representativeness)?
3. Are there advantages in using VI when collecting complex elements, in comparison to web only?

Methods & Data

This study uses data from seven surveys conducted in the UK 2020-2023: National Child Development Study, British Cohort Study, Next Steps, Children of the 2020s, English Longitudinal Study of Ageing, European Social Survey and Health Survey for England pilot. We examine response rates, composition of VI samples and response to complex questionnaire items.

Results

One of the main findings is that VI in the UK is used either as the primary survey mode, or as a complementary mode in mixed-mode designs. VI as the primary data collection mode can lead to lower response rates and potentially also to an increase in representation bias. On the other hand, there are encouraging findings, including that this mode proves to be a very suitable approach for collecting complex elements. This is a key finding since previous research has identified limitations of other remote methods for collecting this kind of data, which is an important component of many studies, especially longitudinal studies.

Added Value

This paper is the first to investigate the use of online VI in UK surveys. Particular benefits exist for longitudinal surveys. The collection of complex data is common practice in longitudinal studies and evidence presented here suggests VI performs well.