Taking a Call: Considerations for Informed Consent for Phone Surveys during COVID-19
Written by Maya Seetharaman, Nitya Kuthiala and Rashi Sabherwal from the CORE Net Secretariat
Gaining a respondent’s informed consent for surveys can be challenging even under normal circumstances, however gaining informed consent over a phone call in the midst of a pandemic is even more challenging — respondents can’t see the surveyors or establish rapport as easily, privacy is often difficult to ensure since most family members are home, and respondents may only be interested in taking calls related to access to entitlements or information. Gaining consent telephonically can also take a lot of time, with some organizations reporting that it takes them anywhere from 10 to 25 minutes, depending on how many questions the respondent may have. Further, standard consent scripts are often not designed to put the respondent at ease, stop what they’re doing and give a surveyor their time.
What goes into a consent script?
Institutional Review Boards (IRBs) mandate that surveyors ask respondents for their informed consent before starting the survey. This usually entails informing respondents of the purpose of study, providing organizational contact information, explaining the risks and benefits of participating, making sure the respondents understand how the data will be used, and specifying that their participation is entirely voluntary throughout the course of the conversation¹. Next, surveyors ask respondents if they have any questions for them — about the information shared, consent protocols, the survey process, or anything else that requires clarification — and answer those as best as they can. Finally, surveyors ask if the respondent is willing to consent to participating in the survey and possibly follow up surveys. If the survey is being recorded, then surveyors may also need to separately ask for permission to record audio.
CORE Net member consent strategies
In order to better understand how researchers and their organisations are navigating informed consents for surveys during the COVID-19 crisis, we reached out to CORE Net’s member organizations and their research partners for their unique insights and challenges. We wanted to hear about some of the innovative adaptations and solutions they may have engaged with, created or explored along the way, and here’s what we learned.
Organizations were thoughtful and deliberate in designing the consent scripts to gain trust from the respondent telephonically. For instance, IFPRI starts their consent script by letting respondents know where they obtained their contact information from and why they’re calling, so that respondents immediately understand how and why they’re receiving the phone call. This may seem like a small point, but makes a big difference for the respondent to understand why they were chosen for the survey and what the consequences of responding might be. This also goes a long way to show that the respondent’s contact information has been obtained from a credible source. EPoD India paid careful attention to keeping the language of the consent script as simple as possible, and worked closely with the translator to keep a conversational tone in the local language (Hindi) translation of their script as well. They also specify that the respondent will not receive any payments for participating in the survey, nor will their access to government entitlements be affected, which are key points to highlight for economically vulnerable households during such a crisis. While surveyors may already translate the consent script to make it sound conversational, it helps them when they have a conversational tone embedded within the script itself.
ICRW and Dalberg went the extra mile to explain who else from the research team was on the phone call so that respondents didn’t feel uncomfortable if they heard another voice during the course of the call. Dalberg also mentions upfront in the consent script that there is no right or wrong answer to any of the survey questions, which is aimed to relieve a respondent from the pressure of needing to answer questions in a certain way. OPM works closely with its partner organization to recruit respondents and let them know ahead of time that they will receive a call from the research team. The OPM research team doesn’t only ask the respondent if they have any questions right after the consent script is read out, but also after each section of the survey, as well as at the end of the survey. PHFI mentions how many other respondents will be interviewed (5000), and explains that all respondent data will be aggregated before sharing. This is helpful as the respondent understands the scale of data collection efforts, and again reinforces that they are not the only ones being targeted for the survey.
Many organizations reported that respondents were confused about if the call was to give them information about entitlements or if their answers would affect their receipt of entitlements. IDinsight deals with this problem upfront in their consent script — they clarify that their purpose is to understand the effects of COVID-19, not to influence their entitlement receipts and/or not to record the rate of COVID-19 infections in their locality.
Separately, each of these details may seem small, however put together, such efforts go a long way in having the respondents feel that their honest answers are valid, that their privacy is being respected and the surveyor at the other end of the line can be trusted.
Rescheduling calls and consent for follow-up surveys
An additional challenge of calling people on the phone is that it is often not clear if they’re free or wish to talk at that given point, leading to the need to make appointments or schedule a time to speak. This means that surveyors need to keep track of when respondents are available, either in their own notes or tracking sheets. NCAER and OPM prompt surveyors to ask respondents in the consent script itself if they are free or would like to receive a call at a different time. This works as a good reminder for surveyors to check in on this before starting to ask any additional questions. NCAER went the extra step of instituting a system for all surveyors. They were collecting quantitative data from 3,466 households (for DCVTS Round 3), which meant a lot of scheduling of calls to keep track of. They trained their surveyors to maintain detailed sheets in Excel with the respondent’s unique ID, the number of attempts made, the status of the interview, and the preferred time that the respondent wanted to speak. Supervisors checked these sheets regularly, which helped them keep track of the team’s daily productivity versus targets. IDInsight also gives their respondents the option of being put on a “Do Not Call” list, which ensures that they will not be inconvenienced with future requests for survey participation.
In addition to finding a good time to talk for that survey round, some organizations like OPM and IFPRI also let respondents know that they might call again for a follow up survey². This is a great practise as it helps the respondent understand the time commitment required from them in the present and possibly in the future. It also means that respondents are not inconvenienced when researchers call them again for follow up questions, which seems to be an issue researchers are beginning to face. Respondents are not only getting calls from multiple research organizations, but also from local service providers and scammers who are trying to take advantage of people’s economic precarity. Keeping in mind the overall environment that respondents are finding themselves in has been helpful to member organizations in fine-tuning their consent process.
Audio recording phone surveys
Another adaptation we saw was that more organizations are considering audio recording their surveys. Not all organizations have chosen to do this, and the choice depends on a range of factors, from the target population to the data quality processes at the organization’s end. In all the consent scripts we read, if organizations were recording the survey, they explicitly asked for the respondent’s consent to do this. ICRW very clearly asks respondents for consent on both these processes separately. This is a great practise as it gives the respondents the flexibility to agree to both, agree to participate in the survey but disagree to audio recording it, or disagree to both, thus giving respondents the power to decide how they want their data collected. If organizations do collect audio recordings, they have had to keep in mind that recordings have to be treated with as much sensitivity as the data itself, and IRBs often encourage researchers to encrypt the audio files just as they would data files. In case encrypting audio isn’t possible, there are workarounds like stripping the audio recording of personally identifiable information. ICRW followed such a practise where they requested the respondent not to reveal their personal information once the recording began.
Maintaining respondent privacy
In addition to the many aforementioned challenges that researchers continue to face and manage along their consent seeking process, a notable one seems to be the challenge of ensuring the privacy and safety of their housebound respondents. Given that many households in India tend to be communal living in nature, guaranteeing or even expecting that a respondent will be able to attend a call completely alone, is quite a difficult task and ask. Women in particular may struggle with ensuring complete privacy, and often answer calls using speaker phone to avoid any unnecessary suspicion on the part of family members. One way that organisations such as SEWA Bharat and Kantar, who both have experienced enumerators, are handling such situations is by encouraging the respondents to maximize their auditory privacy. By doing so, enumerators reduce the risks of other members listening in on a call, thereby ensuring some privacy. Other organisations such as OPM and NCAER have also encountered difficulties when asking respondents, especially women, to answer direct questions about domestic violence and financial information, which tends to often make them feel vulnerable. OPM has attempted to resolve this by offering their respondents proxy questions, which helps the team gather the information they need in an indirect and less triggering way to the respondents³. Lastly many organisations including SEWA Bharat and NCAER have found that gender matching enumerators with their respondents, such as appointing female enumerators to handle their female respondents — which may appear a small consideration — has come to serve as a swift and impactful way of ensuring the safety and comfort of their respondents over the phone.
Providing respondents with extra resources and support
However, some member organizations noted that ensuring respondent’s privacy — which proves to be a tricky task in itself — is often not enough, and on gauging the difficulty of such circumstances, have even taken the extra step of trying to provide similar levels of support that they ordinarily would during in-person sessions. During our chat with OPM and Kantar, they mentioned that it’s challenging to offer emotional support to respondents over the phone, and this has affected how they design their surveys. For instance, OPM has taken out questions related to domestic violence that might be triggering for respondents, and Kantar shortened their survey to only include necessary questions. Dalberg mentioned⁴ that their enumerators share helpline information at the beginning of the call, so in case the respondent drops off, they still have the relevant information to seek help. ICRW mentioned that they make sure to stop after different sections to ask if the respondent is okay to continue, and also alert the respondent before sensitive questions are asked. IDInsight, in their consent script, explicitly let respondents know that if they feel distressed at any point during the survey, they could stop the survey, and that withdrawing consent is their right, and does not affect them in any way. Surveyors are also instructed to take note if the respondent seemed distressed during the call.
Our biggest takeaway from this exercise was there isn’t a one size fits all model to designing the informed consent protocol. While IRBs do mandate standard concepts to be covered, there is a lot of flexibility in how language is used or what points are emphasized that can help put the respondent at ease and trust the research team. These are challenging times for respondents, which is exactly why we’re interviewing them in the first place. Given this, it’s all the more important to approach them with kindness and respect, which are the guiding principles for informed consent.
If you’re looking for help with ethical approval for your study, Monk Prayogshala, a CORE Net member, has an FWA-compliant IRB that can review proposals. We also found this National Data Quality Forum FAQs on Phone Surveys to be a helpful resource in thinking about ethics in phone surveys more broadly.
We would like to thank all the organisations who took out the time to speak with us and helped us better understand their consent processes and the challenges that arose with it.
¹ These are the typical components of a consent script but we are by no means suggesting that this is exhaustive or that there aren’t exceptions to this.
² Follow up calls can take the form of clarifications on the original survey, various survey rounds over time, or even back-checks where a supervisor might call the respondent to make sure that an enumerator had called as reported in the data.
³ The example shared with us was — instead of asking respondents directly about domestic violence, enumerators ask about stress
⁴ CORE Net’s Ethics Dialogue 01 — https://www.youtube.com/watch?v=tPgye2IZQ1g&t=3s & CORE Net Medium Blogpost — https://medium.com/corenet/4-rules-of-the-road-for-conducting-ethical-research-during-covid-19-2f1689c26162
