Face-to-Face interviewing interrupted
Back in May 2020, Patten Smith wrote a blog speculating about the impact of the lock-down cessation of face-to-face fieldwork on the future conduct of high-quality complex population surveys. The blog argued that, because we could not predict the future population prevalence of the virus, what future Government regulations would look like, or how the public would react to being approached by interviewers, the short-term impact of the pandemic on face-to-face fieldwork was uncertain in the extreme.
It then concluded that, in the longer term, the Covid-19 crisis would accelerate a pre-existing shift from face-to-face to web-led data collection methods, but that despite this, as our highest quality data collection method, face-to-face data collection would continue to have a long-term future.
It is now eight months since that blog piece, and, although the pandemic very much remains with us, we remain confident that, in the long-term, face-to-face fieldwork has a secure future in the administration of high-quality surveys in the UK. However, in the intervening months, much work has been done exploring how best to adapt the method to current circumstances, and it feels timely to share this with you.
When will face-to-face interviewing resume?
In May we pointed out that the survey environment was very fluid in respect of Covid-19 prevalence rates, Government regulations and public attitudes to home visits. It is all too clear at the time of writing that nothing has changed here, and that making firm predictions about when, and in what form, normal interviewing will restart in earnest remains a fools’ game.
We say ‘normal’ interviewing, because one type of face to face fieldwork has resumed across the industry, albeit on a smaller scale. Much of the UK interviewer workforce has been redeployed onto the ONS COVID Infection Study, and hundreds of face to face interviewers are now delivering substantial fieldwork volumes. The huge upside of this is that the experienced interviewing workforce is retained until the time it can be refocussed on normal face to face fieldwork when demand returns.
In the short term, we have to accept that we live in uncertain times, and that the environment in which we will conduct face to face fieldwork will be continually evolving. Therefore, the primary focus of those of us who are involved in face to face fieldwork should be on responding quickly and effectively to the ever-changing restrictions generated by the pandemic.
How do we retain core features of face-to-face interviewing?
This is easier said than done! There are many challenges to be overcome, of which two are central: (i) how do we recreate the rapport built by an interviewer with a participant during a conventional face-to-face interview?; and (ii) how can we safely and effectively share and show materials, whilst maintaining social distancing?
And even if we successfully manage to address these primary challenges, there will of course be many others whilst Covid-19 remains with us. Will we have to reduce the length of our interviews? How should we best ask sensitive questions? Can we still administer cognitive tests to respondents? Can we still collect bio-measures, and if so, how?
In response to these various challenges Ipsos MORI and other agencies have tested and developed a suite of options that can be used in place of traditional in-home face to face interviewing. However, none of these is a silver bullet that can neatly fix or replace all our face to face requirements. Instead our overall guiding principles have to be adaptability, fleetness of foot, and readiness to dial up or dial down alternative approaches as required.
We discuss some of these options below.
‘Push to video’?
Initially, many organisations felt that “Push-to-Video” – in essence, interviewing through a Zoom-style call – would be an obvious next step from in-home interviewing. It retained some of the valuable features of traditional face to face interviewing, such as maintaining visual contact (thereby helping build rapport) and enabling show materials to be shared (through screen sharing). Furthermore, the pandemic had removed some of the technological and cultural barriers we faced before: zoom calls, once an alien concept to many, have become a norm in a way that transcends generations.
So, it was the obvious replacement for conventional face-to-face interviewing. Except it wasn’t!
In pilots and mainstage fieldwork, conducted across audience measurement trackers, omnibus and social surveys and totalling several thousand interviews, we found that the general public, when given the option, preferred alternative, less intrusive approaches, such as socially distant doorstep interviews or appointments made to conduct interviews later by phone. Although our pilots helped us identify improvements to the approach that should help reassure and encourage participants to take up push-to-video, our findings also demonstrated that for most surveys, this is not the sole solution.
It may turn out that its greatest value will be in longitudinal and cohort studies where we have pre-existing relationships with respondents and engagement levels and trust have already been established.
‘Knock to nudge’
Another approach we have been actively exploring is the so-called ‘Knock-to-Nudge’ enhancement of the push to web method. Push-to-web is now the favoured method for administering self-completion questionnaires in the UK. It involves our writing to a random sample of addresses, and asking residents to go online and use a supplied URL to open and complete an online questionnaire.
In order to minimise bias arising from non-participation of those who are digitally less confident, late in the reminder process sample members are usually approached for offline data collection (often via a mail questionnaire). Typically, the method delivers response rates in the 20%-30% range.
Knock-to-nudge uses interviewers to drive up participation in push to web surveys. It involves interviewers calling on the previously written-to households and encouraging them to go online to participate. It can also be used to offer alternative ways of participating such as paper self-completion, web-based completion using a secondary device or an interviewer administered telephone interview, depending on the nature of the survey.
During our tests of the method during the pandemic on an audience measurement survey, we more than doubled the response rates to the initial online phase increasing them from 17% to 36%, and achieving just under 1,000 additional interviews. We expect only to improve upon this performance as we hone our approach, and start to understand how best to utilise knock to nudge as a way of optimising public participation.
Socially distant doorstep interviewing
Another approach that is possible for some projects is socially distant doorstep interviewing. The approach works well with surveys of around 20 minutes or less where show cards are not essential and “read outs” can be used as a substitute (although see discussion of show-materials below).
While interviewers cannot enter the homes of participants they are still able generate a rapport with the participant and administer questions that may be complex or sensitive. To date, when we have moved to doorstep interviewing, response rates have not dropped and public reaction has been positive. However, as we enter the cooler winter months we may encounter more challenges around doorstep interviewing!
Interviewer-led telephone interviews
Another approach we have explored on many of our trials for both Audience Measurement and Social Surveys is to offer participants the option of conducting the interview over the telephone, conducted either immediately (with the interviewer administering the interview from their car) or later at an agreed appointment-time. Our interviewers enjoyed the flexibility this gave them, and anecdotally reported feeling that they secured interviews they would otherwise not have got.
At first glance, this method may not appear to offer benefits distinct from traditional telephone interviewing, but this is to ignore the considerable recruitment advantages conferred by initial face-to-face contact. Initial contact face-to-face is considerably better at contacting and recruiting more reluctant and harder to reach members of the public.
Innovative ways of sharing show materials
A seemingly trivial problem that arises when interviewers are no longer able to enter people’s homes is that it becomes difficult to use show cards (and other show materials).
Seemingly trivial, because surely all the interviewer has to do is to read out the relevant lists of response options? But actually less trivial than it seems because the two approaches often fail to produce identical data - both because lists that are read out have to be substantially shorter than lists that are read by respondents, and because the approaches typically lead to different kinds of response error (recency v. primacy effects for example).There are therefore good reasons for continuing to use show cards if one has done so in the past.
One possible approach is to host show cards on a secure server and provide participants with a QR code or short URL, thereby allowing them to be viewed with no physical contact. In pilot work, carried out for a central Government Department, achieving 140 interviews, we found that participants were willing to use this method, and that those with hand-held devices could then take part in a socially distanced interview on the doorstep with the interviewer.
Others chose to view the materials from a laptop or desktop whilst our interviewers conducted the interviews via telephone. But again, this approach is not one that will work for everyone: those participants who are uncomfortable with technology, or who don’t have internet access, would remain reliant on interviewers reading out show materials. And as just discussed, doing this can change people’s responses, and may also impact on response, engagement and interview length.
A second approach is to provide respondents with single-use show cards. Although, this is a fail-safe method for sharing materials with all participants, its environmental impact is less than ideal.
A third possibility is to hand to the respondent a wipe-down secondary handheld device. This is a more environmentally friendly way of ensuring that all respondents, irrespective of their technical literacy, see the same set of show materials. It also has the advantage that such a device can be used to ask participants to complete an instant self-completion questionnaire.
We intend to carry out further trials to assess the impact and success of sharing materials. However, at present, the low-fi option of single use, recyclable show cards is faring well: it is a neat failsafe method and is unlikely to alienate many.
Where are we now?
It’s not all doom and gloom. First, as we have already said, face-to-face fieldwork has a secure future simply because it is irreplaceable in high quality surveys. If we need data of the highest possible quality, one element of its collection has to be face-to-face interviewing.
Second, the pandemic has meant that data trends, once the dictator of the pace and type of change, are already broken for face-to-face surveys, and this has allowed us to rapidly roll out new techniques that have been long in the planning, and which, pre-COVID, may have taken decades to achieve. We have seen developments in hybrid methods involving push-to-web and face-to-face mixed mode data collection in a matter of weeks and months rather than years.
And third, we have been taught important lessons on the importance of adaptability. Whereas in the past our face-to-face fieldwork method relied heavily on expensive multiple face to face call backs, our current approach is analogous to asking respondents to choose from a menu. How would you like to respond? Push-to-web? Push-to-video? Telephone?
Anecdotally interviewers have reported that flexibility and discretion over mode has meant they have been able to secure interviews with participants who would otherwise have refused. The ability to offer a phone or video call later, and crucially at the convenience of the participant, has given interviewers another weapon in their arsenal to secure an interview.
Once fieldwork returns to a new normal, we expect to hold on to this flexibility with all the benefits that brings, but with one difference: sitting in the living room doing an interview the traditional way will also be on the menu!
And next?
Once life starts returning to normal in 2021, we anticipate that there will be strong client-demand for face-to-face fieldwork, because no other method delivers data of equivalent quality.
We predict that, overall, the public will be receptive to this. During COVID we have seen a bounce in participant willingness to help with surveys (human contact is seen as prized rather than a nuisance!); and, anecdotally, interviewers have commented on the willingness of participants to take part in surveys with hostile refusals at an all-time low. We expect this to continue in the short term once face-to-face fieldwork resumes.
This probably won’t last as the years go by, but face-to-face interviewing will remain an essential ingredient of the emerging standard, mixed-mode, method for administering surveys of the highest quality. The future will be mixed-mode, and flexible face-to-face interviewing will be an essential part of that mix.
Adele Bearfield is the Client Service Director for Face-to-Face Field at Ipsos MORI, she has worked in Market Research for 25 years, operationalising and delivering many high quality random probability surveys.
Patten Smith is Director of Research Methods at Ipsos MORI and has worked on high quality social surveys for around 40 years. Patten is visiting professor at Surrey University and was Chair of the SRA from 2011 to 2017 .