When I found a tool developed by Dr. Daanish Mustafa to quantify vulnerability (see Mustafa et al., 2011), I realised this was an opportunity to apply it to a context I knew. An index that enabled policymakers to identify the areas and regions most vulnerable to climatic hazards could be very useful in targeting the limited resources available more effectively, in order to better prepare and adapt to such hazards. In order to tailor the index to the local context, my intention was to conduct in-person interviews with village chiefs and other local representatives in the region. I then aimed to lead a series of surveys to collect data to feed into the index to obtain a representative picture of the variation of vulnerability across southern Madagascar.
A change of plans due to COVID-19
Unfortunately, by March of 2020, it had become clear that the COVID-19 pandemic would force me to adopt a different strategy. Since I could no longer travel to collect data in person, I would need to move my research online. I had already counted on using secondary research to inform the adaptation of the vulnerability index to Madagascar, but now this was going to have a much larger weight.
I also started looking for tools to build online questionnaires. Because I wanted to present the results in map form, the survey I shared needed to be equipped with geotagging features. I found the answer in a geospatial software, ArcGIS, which had a survey add-on. This enabled me to create an online questionnaire, easy to navigate for respondents from any device. A key part of the process was translating the survey materials to the local language, Malagasy, to enhance the local’s understanding. I shared the questionnaire through local Facebook groups, since a large majority of Madagascar’s "connected" population is present in that social media platform.
The challenges of online questionnaires
These surveys, conducted via the Internet, were compelling because they reached potential respondents simultaneously, reducing the time requirements significantly. In recent years, the widespread penetration of mobile devices, fast network speeds and social media in some remote regions has enhanced the feasibility of online approaches. However, online research is not devoid of limitations. Because my questionnaire now required access to a mobile phone or computer , I was forced to restrict the scope of my study. It was no longer possible to collect data from rural areas —since many of them lack electricity access— so I decided to focus on the city of Toliara exclusively. The high illiteracy rates in the region represented an additional barrier: the questionnaires were less accessible for many respondents than the in-person, oral surveys that I had originally sought to carry out. Some potential participants even mentioned their fear of the online survey being a sort of scam.
I also realised there were some more profound and subtle effects taking place within my sample. Inequalities in phone ownership or access to the Internet had likely skewed my target sample away from the most marginalised and vulnerable groups. For instance, most of the questionnaire respondents reported having access to running water and adequate sanitation, despite local research pointing out that these conditions are not widespread in the region. Clearly, the online approach had missed the 'poorest of the poorest', so it was important to acknowledge this flaw. I decided not to treat my results as 'final', but rather use them as part of a 'proof of concept' approach to test the vulnerability index.
Outcomes of my research
My objective was to produce a tool that could be easily adopted by policymakers and other relevant organisations operating in the region. For this, I created a policy briefing outlining the index's main elements and detailing its functionality. Through this document, I also shared other findings of my research — such as the fact that some governmental recommendations to reduce the risk of climatic hazards had not been successfully taken up in Toliara. Currently, I am working on developing a program to calculate vulnerability scores automatically from the responses to the questionnaire, which will make the index more practical with large datasets.
Even if remote research can feel somewhat disconnected from the targeted study region, online methodologies may well lead —just like traditional research methods— to having real, tangible impacts. The pandemic has highlighted some of the obstacles that researchers may encounter when mobility is constrained, and online approaches have proved to be effective tools in instances like these. Like those who went before us, we should seize this moment to rethink how we approach our work and start considering integrating remote research methods alongside more traditional ones.