The success of longitudinal studies depends on maintaining survey cohorts and achieving low attrition rates over long periods of time. Investing in successive in-person surveys are one way to help achieve this, however the COVID-19 pandemic made these impossible. Young Lives therefore adapted to collect data remotely through our innovative COVID-19 Phone Survey in 2020. In 2021, we have further developed our approach to include collecting GPS data of our participants’ household location. This wasn’t without challenges, not least developing this cutting-edge approach whilst protecting the anonymity – and safety – of our participants. In this blog we set out the important advantages of collecting GPS data, our method, the challenges we faced and lessons learned for future research.
GPS data supplements the address information provided by survey respondents helping to clarify the precise location of households and reduces potential errors when entering the address manually. There are multiple advantages to this:
- GPS data provides significantly greater accuracy of participants’ location which facilitates subsequent in-person data collection, improving our ability to reach respondents (and ultimately reducing long term attrition rates).
- GPS data significantly increases the ability to match data with other georeferenced datasets; for example, when investigating migration, access to social protection programmes or to better understand the living context of the study respondents. In the case of Young Lives, it allows us to combine 20 years of unique longitudinal data with other “big data” sets, such as climate data, as illustrated in a recent article by Chang, Favara and Novella (2022)
- Collecting GPS data remotely could lead to better and more cost-effective tracking strategies, for Young Lives and other longitudinal surveys. For example, participants’ location could be updated more frequently enabling researchers to keep better track of the survey respondents between in-person survey rounds. The benefit for cohort maintenance and reduced attrition rates is clear.
As part of our 2021 phone survey, we designed a short online survey to collect participants’ GPS data remotely. To be eligible, participants needed access to a smartphone and internet at their home address (either through a mobile data plan or Wi-Fi connection), requirements that were checked through the phone survey. Eligibility of participants varied across our study countries, ranging from the lowest in Ethiopia (34%) to the highest in Vietnam (90%).
Taking part in the online GPS survey was voluntary, with consent being sought twice (both oral and written), initially at the end of one of the phone survey calls, when the enumerator introduced the participant to the idea, and again at the beginning of the online survey itself. Participants were reassured that all information collected would be treated in strict confidence and stored securely and anonymously, in the same way as all the information they have provided in the past. Collecting GPS data is not new to Young Lives: in past in-person survey rounds, GPS coordinates were collected by fieldworkers using satellite navigation devices with participants' consent, but this is the first time we have collected the data remotely, with participants actively providing the data.
The online survey was very simple to complete. Eligible participants accessed the online survey at home via a link provided to them by an enumerator through text/WhatsApp/Telegram and entered their name and birth date manually. To submit their GPS location, the participants simply had to click on the “record location” button within the survey. The GPS data would then be automatically collected using the device’s built-in GPS support. To increase the accuracy of the GPS data collected, the participants were encouraged to stand outside their home for an improved signal when possible.
We piloted this new collection technique with a sub-sample of participants in all four study countries during a getting in-touch call in August, and evaluated the method for compliance, usability and accuracy.
Results from the pilot showed that:
- Compliance varied significantly by country. The rates among participants who were invited to access the link and who completed the online survey were highest in Peru (81%) and Vietnam (64%), and much lower in India (44%) and Ethiopia (31%).
- The online survey was easy to use and had high success rates for data collection: of those who completed the survey, nearly 100% were able to follow the instructions and submit their GPS data across countries.
- GPS accuracy was good. The GPS coordinates collected in the latest in-person survey rounds (in 2016) allowed us to undertake accuracy checks by triangulating previous GPS data collected in person and the new GPS data collected remotely. More specifically, we calculated the distance between GPS coordinates collected in 2016 and GPS coordinates collected in the pilot, for those who had not moved location; the median distance ranges were between 60 meters (Vietnam) and 1030 meters (Ethiopia)
Following the pilot, we used the GPS online survey in Call 5 of our Phone Survey series (conducted between October and mid December 2021) with the eligible sample in all countries except Ethiopia. We did not proceed in Ethiopia because of the low eligibility rate (and compliance) and challenging circumstances due to ongoing conflict
The key headline is that it is possible to accurately collect GPS data remotely! Our innovative new method is a relatively easy (and cost-effective) exercise and where eligibility is high, GPS data can be collected for most respondents.
However, collecting GPS data through smartphones is not always feasible for a number of reasons. First, eligibility and compliance rates might be too low, as in the case of the Young Lives sample in Ethiopia. This, combined with a persistent divide on access and ownership of digital devices can make data collection unfeasible and may lead to biased results.
Second, it might lead to unnecessary anxiety and concern in the respondents, which needs to be understood and appropriately addressed. For example, participants expressed concerns about digital security in all our study countries. We were able to address these concerns by offering specific support and additional information – reassurance supported by the long term, trusting relationships between Young Lives and our study participants.
Lastly, risks and mitigation of potential selection bias needs to be considered in any related research analysis (both eligibility and compliance-related). To increase compliance, Call 5 included two reminders (by phone call or text message) to complete the online survey, a small financial incentive in selected countries, and the option to borrow a smartphone from a friend or neighbour in India.
The data collected has huge potential for data matching and better tracking of Young Lives participants. With ever increasing mobile phone access, this innovative way of data collection could offer interesting possibilities in the future for Young Lives as for other longitudinal studies.
The success of longitudinal studies depends on maintaining survey cohorts and achieving low attrition rates over long periods of time. Investing in successive in-person surveys are one way to help achieve this, however the COVID-19 pandemic made these impossible. Young Lives therefore adapted to collect data remotely through our innovative COVID-19 Phone Survey in 2020. In 2021, we have further developed our approach to include collecting GPS data of our participants’ household location. This wasn’t without challenges, not least developing this cutting-edge approach whilst protecting the anonymity – and safety – of our participants. In this blog we set out the important advantages of collecting GPS data, our method, the challenges we faced and lessons learned for future research.
GPS data supplements the address information provided by survey respondents helping to clarify the precise location of households and reduces potential errors when entering the address manually. There are multiple advantages to this:
- GPS data provides significantly greater accuracy of participants’ location which facilitates subsequent in-person data collection, improving our ability to reach respondents (and ultimately reducing long term attrition rates).
- GPS data significantly increases the ability to match data with other georeferenced datasets; for example, when investigating migration, access to social protection programmes or to better understand the living context of the study respondents. In the case of Young Lives, it allows us to combine 20 years of unique longitudinal data with other “big data” sets, such as climate data, as illustrated in a recent article by Chang, Favara and Novella (2022)
- Collecting GPS data remotely could lead to better and more cost-effective tracking strategies, for Young Lives and other longitudinal surveys. For example, participants’ location could be updated more frequently enabling researchers to keep better track of the survey respondents between in-person survey rounds. The benefit for cohort maintenance and reduced attrition rates is clear.
As part of our 2021 phone survey, we designed a short online survey to collect participants’ GPS data remotely. To be eligible, participants needed access to a smartphone and internet at their home address (either through a mobile data plan or Wi-Fi connection), requirements that were checked through the phone survey. Eligibility of participants varied across our study countries, ranging from the lowest in Ethiopia (34%) to the highest in Vietnam (90%).
Taking part in the online GPS survey was voluntary, with consent being sought twice (both oral and written), initially at the end of one of the phone survey calls, when the enumerator introduced the participant to the idea, and again at the beginning of the online survey itself. Participants were reassured that all information collected would be treated in strict confidence and stored securely and anonymously, in the same way as all the information they have provided in the past. Collecting GPS data is not new to Young Lives: in past in-person survey rounds, GPS coordinates were collected by fieldworkers using satellite navigation devices with participants' consent, but this is the first time we have collected the data remotely, with participants actively providing the data.
The online survey was very simple to complete. Eligible participants accessed the online survey at home via a link provided to them by an enumerator through text/WhatsApp/Telegram and entered their name and birth date manually. To submit their GPS location, the participants simply had to click on the “record location” button within the survey. The GPS data would then be automatically collected using the device’s built-in GPS support. To increase the accuracy of the GPS data collected, the participants were encouraged to stand outside their home for an improved signal when possible.
We piloted this new collection technique with a sub-sample of participants in all four study countries during a getting in-touch call in August, and evaluated the method for compliance, usability and accuracy.
Results from the pilot showed that:
- Compliance varied significantly by country. The rates among participants who were invited to access the link and who completed the online survey were highest in Peru (81%) and Vietnam (64%), and much lower in India (44%) and Ethiopia (31%).
- The online survey was easy to use and had high success rates for data collection: of those who completed the survey, nearly 100% were able to follow the instructions and submit their GPS data across countries.
- GPS accuracy was good. The GPS coordinates collected in the latest in-person survey rounds (in 2016) allowed us to undertake accuracy checks by triangulating previous GPS data collected in person and the new GPS data collected remotely. More specifically, we calculated the distance between GPS coordinates collected in 2016 and GPS coordinates collected in the pilot, for those who had not moved location; the median distance ranges were between 60 meters (Vietnam) and 1030 meters (Ethiopia)
Following the pilot, we used the GPS online survey in Call 5 of our Phone Survey series (conducted between October and mid December 2021) with the eligible sample in all countries except Ethiopia. We did not proceed in Ethiopia because of the low eligibility rate (and compliance) and challenging circumstances due to ongoing conflict
The key headline is that it is possible to accurately collect GPS data remotely! Our innovative new method is a relatively easy (and cost-effective) exercise and where eligibility is high, GPS data can be collected for most respondents.
However, collecting GPS data through smartphones is not always feasible for a number of reasons. First, eligibility and compliance rates might be too low, as in the case of the Young Lives sample in Ethiopia. This, combined with a persistent divide on access and ownership of digital devices can make data collection unfeasible and may lead to biased results.
Second, it might lead to unnecessary anxiety and concern in the respondents, which needs to be understood and appropriately addressed. For example, participants expressed concerns about digital security in all our study countries. We were able to address these concerns by offering specific support and additional information – reassurance supported by the long term, trusting relationships between Young Lives and our study participants.
Lastly, risks and mitigation of potential selection bias needs to be considered in any related research analysis (both eligibility and compliance-related). To increase compliance, Call 5 included two reminders (by phone call or text message) to complete the online survey, a small financial incentive in selected countries, and the option to borrow a smartphone from a friend or neighbour in India.
The data collected has huge potential for data matching and better tracking of Young Lives participants. With ever increasing mobile phone access, this innovative way of data collection could offer interesting possibilities in the future for Young Lives as for other longitudinal studies.