An old colleague used to refer to the biannual conference of the European Survey Research Association (ESRA) as the Glastonbury of social researchers and survey methodologists. I think it describes this four-day festival of research methods perfectly - well, if you’re as inspired by research methods as we are here at Ipsos.
ESRA gathers every two years in a different European capital, this year in beautiful Zagreb, Croatia. The conference brings together hundreds of social researchers and methodologists from across Europe and beyond to discuss a variety of topics around survey methodology. Like at Glastonbury, there are multiple ‘stages’ (i.e. endearing old university lecture rooms) where different ‘acts’ (researchers, methodologists, academics) deliver their research papers at four days of simultaneous sessions on a variety of topics. Old ‘stars’ in the survey research world are guaranteed to attract dedicated audiences year on year alongside plenty of new faces.
A total of six papers from Ipsos MORI Public Affairs were presented, a summary of which will be available soon. In between all of this, much fun was had on breaks and evenings filled with networking and opportunities to explore the host city.
The key themes this year were as follows:
How to best make use of online data collection opportunities. The survey research sector has undergone some rapid changes in the data collection environment in recent years. First and foremost, these have entailed a shift from ‘old’, prestigious single mode designs, such as probability-based face-to-face interviewing, to mixed mode designs that incorporate online data collection in the mix, combining for example traditional questionnaires with passive data collection, social media or admin data. The ensuing mood of change was woven throughout the conference from its start. In the keynote speech, Barry Schouten of CBS challenged whether old assumptions still hold at the time when alternatives to traditional questionnaire surveys are increasingly available and offer new, potentially enhanced opportunities for measurement. He encouraged us to rethink the underlying concepts; survey measures are often proxies for underlying concepts that we may now be able to measure directly and more accurately through hybrid data collection methods, integrating data from traditional surveys with sensors or app data.
A substantial part of sessions that followed were indeed dedicated to this: exploring how to move data collection online and how to integrate the old and new without compromising quality. Much of this had been discussed in Lisbon two years ago, but some impressive new developments were showcased, such as Westat’s food logger app, which combines passive data collection, geolocating of food establishments and supermarkets as well as API food recognition features. Completely new this year was also a session on machine learning in survey research. While the results of application of machine learning and predictive modelling still seem varied in this context, it’s positive to see that many colleagues are experimenting with what is possible.
New developments also bring with them a set of new challenges and questions that sparked extensive discussion; respondents’ viewpoints on ethics, privacy and control over their data may never have been as important as today, nor the question of how representation of all groups can be ensured through online modes, and what impact the devices respondents use to complete a survey may have on their responses.
Linked to this shift to mixed mode research, there has been a move away from response rates and non-response bias as a topic in its own right, compared with previous years. At this year’s ESRA, the greater focus was on response rates’ relationship to measurement issues and data quality, particularly looking at online surveys. In fact, there was only one dedicated session on non-response bias, the very last of the conference, comprising of just two papers. It is a welcome development that researchers now appear less fixated by response rates, and are taking a broader view of survey error and what constitutes high quality in 2019.
In our increasingly transactional society, respondents also expect to get something in return for participation. Indeed, sophisticated communication is required to motivate participation in the first place. This was reflected in the number of sessions dedicated to how to engage respondents, particularly in the absence of face-to-face interviewers. For example, the ONS’s approaches on the Labour Force Survey rooted in iterative testing processes and successful use of a pretty tote bag as an incentive were inspiring to hear about.
There were also many more sessions on surveying migrants and refugees, most specifically on surveying Syrians in Germany. This clearly aligns with what is currently topical in Europe. We found these sessions particularly helpful for similar work we are conducting on researching hard-to-reach and sensitive groups. All in all, ESRA 2019 was a refreshing one. We’ll be considering how we can use these findings to finetune our approaches at Ipsos over the coming months.