Using smartphone applications to enhance longitudinal survey methods

BACKGROUND
When researchers want to explore the development of social behaviour, such as delinquency or substance use, they usually rely on questionnaires to collect their data. In order to explore issues that questionnaire data can't answer, they use qualitative studies. Combining the data from large scale quantitative surveys with those of small scale qualitative ones is challenging, because the groups of people involved are often not the same. And asking people to complete a questionnaire is not a very good way to find out about their behaviour. People do not always recall their behaviour accurately; they may not wish to disclose some things, and people find it easier to remember some kinds of things than others. In short, our memories are notoriously inaccurate and they can be very biased. Finally, there is a limit on how many questions we can ask people to answer on a questionnaire.

TRANSFORMING DATA COLLECTION
Smartphone apps could transform data collection, providing social scientists with richer, more reliable data on which to build theory and knowledge. Although researchers already use GPS data, and data from Facebook and Twitter, these sources cannot match the potential of a bespoke smartphone, that will - simultaneously - be able to collect different kinds of data on the same individual or group of individuals. This can overcome many, if not all, of the limitations summarised above.

THE STUDY
In this 'proof of concept' study, we propose to examine factors influencing adolescent behaviour, particularly substance use, as a means of testing the feasibility and acceptability of this form of data collection.

METHOD
All consenting pupils aged 16-18 (n 200) attending two schools will complete a baseline survey providing basic demographic information, and information about their friendship networks, activities, and substance use. Parts of the survey will be repeated at 2 and 6 months, to capture self-reported changes in friendship networks and behaviour. They will be asked to install an app on their phone that will, without their further intervention, collect the following data across this 6 month period:
- GPS data on where, when and with whom adolescents spend their time
- Data on the frequency and duration of texting and phone use, providing information on these important forms of communication
We will also collect:
- Contextual data on young people's 'lived experience', by inviting participants to contribute photo-journal/visual ethnographic data about features of the local environment relevant to their lives.
- Data on substantive issues, such as drug and alcohol use, via periodic sampling of current behaviour.
In short, the app will do something that the traditional survey method cannot. It will provide data on young people's social connectedness, the extent to which this changes on a daily basis, how it changes and develops over time, and how this is influenced by the physical environment.

We have included 'mini-experiments' within the study to assess the effectiveness of different ways of retaining the participation of young people.

ANALYSES
We will be compare the information from paper surveys with the electronic data, shedding light on the 'value added' of the latter from of data collection, and any problems associated with it. The electronic data will enable us to investigate associations between proximity, peer nominations and features of the physical environment that can influence behaviour e.g. isolated spots, social venues, private versus public spaces. We will be able to use the data to examine their value added over reliance on imputational methods typically associated with Social Network Analyses.

ETHICAL ISSUES
Approval will be sought from a University Ethics Committee. We will explore the ethical issues with colleagues from a range of social science disciplines, and run focus groups with participating young people to explore their experiences, attitudes and concerns.

The primary impact of this pilot study will be for academic beneficiaries. However, because smartphone apps could transform data collection in the social sciences, providing researchers with richer, more reliable data on which to build theory and knowledge, it has the potential to impact on a range of users in the longer term.

For example, if the pilot study is successful, it could pave the way for a more substantive study on adolescent development, which could in turn provide policy makers from a wide range of government departments, and service providers from the statutory, voluntary and private sectors, with a better understanding of how spatio-temporal factors and peer networks influence the development of patterns of alcohol and substance use, delinquency, or healthy patterns of behaviour e.g. physical activity. This information could, longer-term, benefit adolescents via the development of policies or interventions designed to support optimal development, and minimise health problems or risky behaviour such as the misuse of alcohol, licit and illicit substances. This would also benefit society more widely, in terms of NHS costs, criminal justice and welfare costs etc.