Remote data collection for large-scale longitudinal research: reflections from the pre- and lessons for the post-pandemic era

Dr Bonamy Oliver, Associate Professor in Developmental Psychology, recently joined the UCL Institute of Education and has worked with longitudinal studies for many years. Here, she writes about the challenge of family psychology in 2020, and what pre-pandemic efforts to conduct detailed family assessments suitable for large-scale studies have to offer.

2020 will not be quickly forgotten. As I write, the pandemic continues to alter lives across the globe and the long-term effects are yet to be known. Importantly, COVID-19 is not only a physical health problem: the virus and its associated measures will have adverse psychological consequences for many adults and children, in ways unequal across societies, and potentially for a long time to come (World Health Organisation, 2020). For children, the impacts of their direct experiences (e.g. loss, school closure, social isolation) are confounded by those of the adults around them. In particular, the wellbeing of parents and other primary carers (hereon ‘parents’) is crucial for all family relationships, and in turn for children’s home-schooling, development and mental health. In times of stress, hardship and close proximity, these family processes can change dramatically.

At this critical time for children and families, improving our understanding of the psychological impacts of the crisis, as well as how to best support families in need, is essential. To do so, robust methodologies for data-collection are crucial. Yet, effectively characterising family life is a challenge, and, for reasons of resource, not a new one for large-scale, longitudinal research. A common solution is the use of questionnaires, providing invaluable perspectives from individuals, but nevertheless a compromise of detail. Other traditional methods can offer such nuance, but involve home or lab visits, always expensive on a large or longitudinal scale, and currently out of the question. How, then, to best understand ‘family’ in current times?

In fact, there has been strong rationale for addressing the methodological challenge of remotely assessing family relationships for some time: to increase sample size and diversity, to bring further detailed understanding of family to the power of large samples, and to reduce the carbon footprint of research and practice. Now the challenge could not be more topical, but the solutions needn’t be different. Here, some existing solutions are presented, as well as some questions and suggestions for moving forward.

Personal perspectives are important

Easily adapted to remote use and standardised across studies, questionnaire reliability and validity is assumed to be not much shaken by being put online, on mobiles, or asked over the phone. Notably, there is variance in the extent to which this premise holds true across psychological domains, and in any case there are undeniable biases in questionnaire reports that should not be ignored. However, evidence over many years and across numerous domains has demonstrated that individual perspectives offer important and unique insights to advance our understanding of family and children’s psychological development. Perspectives collected from online questionnaires are likely to be no different. If evidence from other quarters tells us anything, it is that different data-collection approaches may not tell us the same thing, but they do each tell us something important. As such, online questionnaires are one robust remote method of data collection available to us, and surely our quickest win.

Don’t throw the baby out with the bathwater

Whilst I am a strong proponent of questionnaires, families are complex, and multiple methodologies are essential to capture a full picture of them. Family interviews and observations are commonly used methods, but the traditional home- or lab-visits needed to collect them are currently out of the question. Should we then scramble to innovate and generate new methods that go beyond questionnaires? It’s true that innovations of methodology may add importantly to our toolkit, however, we need to take care to be rigorous and evidence-based in our approaches, not simply rush to make the new. Rather than rejecting existing robust methods, small modifications to existing procedures can produce valid and reliable remote methods fruitful for family research.

One example is the well-documented Five-Minute-Speech Sample (FMSS; Magana et al., 1986). FMSS involves eliciting a five-minute monologue as an economical approach to collecting so-called expressed emotion (EE). Parent EE refers to the emotional attitudes of parents towards their offspring (e.g., critical, hostile, warm, sensitive), and, following its history in adult psychiatry, is now recognised as a principal mechanism for children’s psychopathology, adding important information over and above parent questionnaires (e.g., Mark et al., 2017). FMSS data collection needs little training and, most relevantly, can be quickly and reliably achieved over the phone, meaning that wide and diverse reach is possible. Coding is more time consuming and specialised but is nothing compared to in-home or lab data collection resource need: trained coders can code FMSS in around 20 minutes. Note that FMSS can be used with general populations and specific clinical groups, with teachers and children as well as parents, and coded for many pertinent constructs other than EE.

Another widely used approach in family psychology is direct observation of parents and children. Before COVID-19, in work that now seems prescient, a novel approach for observing parents with their children was developed, known as ESO (Oliver & Pike, 2019). ESO is a freely available, online direct-observation tool based on a well-established and widely used traditional methodology. Using video calls, ESO allows researchers to observe and record parents and children interacting while they complete a joint task using a game-like interface. As such, ESO provides the means to collect standard observations of parent-child interactions in the family home without the need to travel or have direct, in-person contact. Like the FMSS, ESO can be delivered with little training, is time-efficient, and provides detailed information about parent-child interactions. Moreover, it is flexible in terms of data collected and coding, and can predict children’s outcomes over and above parent questionnaires (Oliver & Pike, 2019).

Budget for making a difference

COVID-19 has revealed our extraordinary digital divide: critical for understanding children’s current and long-term outcomes. For our data collection aims, ensuring methods are suitable for tablet or smart-phone use will help considerably, and approaches that involve phone calls will also go a long way, but there will still be families missed. Here, rather than jumping to resource efficient methods, we can aim to have impact beyond our research. Offering our families IT/Internet provision will broaden our research, while serendipitously reducing the technology gap in ways crucial for children’s wellbeing and education. We can make a difference to people as well as to our research by encouraging funders to meet these costs.

Change should be long-term

With widening achievement and mental-health gaps for children, it is more important than ever to improve our understanding of family processes in the contemporary context – only by doing so can we best support families. Because of the current crisis, the use of remote methodologies is essential. However, we should not forget that there have long been important reasons to use and develop these methods. For reasons of improving diversity, sample size, detailed understanding and our green credentials, we must aim beyond the crisis, to change the landscape of family research and practice in the long-term.

Oftentimes, we hold cards close to our chest in our field, happier to stay in our silos than share our ideas. Now is the time for innovation and collaboration. It is a time to think about novel approaches to family assessment for both research and practice, and it is a time to share. I would like to start a library of validated remote methods for family psychology that would put all protocols, stimuli and so on out there and available. I aim to include my methods and to hope that people will respect our academic currency: credit. Are you with me?

Further information

A developmental psychologist and behavioural geneticist, Bonamy Oliver has worked with large and small-scale longitudinal samples, and is well-versed in the strengths and pit-falls of both. As a consequence, she is passionate about bringing detail to our understanding of family in large-scale studies, as well as about bringing scale and diversity to clinical research. Funding permitting, and with the help of key collaborators, she is currently working on reducing the coding burden of the FMSS, broadening intervention reach, and extending the ESO methodology to be appropriate for a range of study designs and objectives, as well as for different relationships.

If you are interested in using the methods outlined, or adding your own family methodology to the library mentioned in this article, please contact b.oliver@ucl.ac.uk

To read more blogs in the COVID-19: Perspectives series, visit our COVID-19 Longitudinal Research Hub.

Dr Bonamy Oliver is an Associate Professor of Development Psychology at University College London. Follow Bonamy on Twitter: @BonamyOliver

Suggested citation:

Oliver, B. (2020). ‘Remote data collection for large-scale longitudinal research: reflections from the pre- and lessons for the post-pandemic era’. CLOSER. 7 August 2020. Available at: https://www.closer.ac.uk/news-opinion/blog/remote-data-collection-for-large-scale-longitudinal-research