When we need to explore behaviour change in our projects we opt for longitudinal studies, building on traditions in psychology and sociology. In a recent project, we were asked to explore how people reach their health and fitness goals, as well as the mechanisms that enable or hinder the formation of new habits and routines. To do this, we set up a 10-month international study with over 80 participants, who were at various stages of progress in reaching a health or fitness goal.<\/p>\n
This timespan allowed us to take into account changes in personal circumstances, as well as seasonal variables that can affect routines significantly. Conducting several rounds of interviews and surveys, and asking similar questions each time, enabled us to get rich insights into participants\u2019 health and fitness intentions. These covered both short- and long-term goals, getting to the essence of what truly motivated them, or prevented them, to achieve those goals.<\/p>\n
The research used a blended method of both qualitative and quantitative research. The former was done through multiple one-hour interviews (once every three months) and the latter through regular surveys with multiple choice questions.<\/p>\n
Whereas in a singular study one snips insights from a particular time in people\u2019s lives, rather like cutting flowers for a bouquet of insights, longitudinal research is more like witnessing flowers growing over time, observing what supports (or hinders) them from reaching blossom.<\/p>\n
Working with massive quantities of qualitative data<\/b><\/span><\/p>\n There are many ways of synthesising and analysing data in design research. Considering the longitudinal, blended nature of this study, we went through several loops of iterations and developed two complex matrices to keep track of the cumulating data and observations. These matrices served different purposes but were also complimentary.<\/span><\/p>\n The first matrix was set up at the beginning of the study with the main research questions and interview rounds. We gradually filled in the qualitative data we collected through interviews (quotes and stories) as we went along. This way we were able to provide details and concrete examples of quantitative findings which were popping up throughout the study. For example, if researchers spotted a pattern from surveys, they were able to easily find evidence for that pattern using the matrix variables, such as participant segment, age, country, stage of change, technology in use, etc.<\/span><\/p>\n The second matrix was set up after we completed the interviews, and we used it to incorporate elements of trajectory analysis. To keep this matrix straightforward and uncluttered, we quantified the analysed qualitative data with simple drop-downs and coded variables and did not include quotes from the participants (which were in the first matrix). This way, we were able to easily track main changes over time and spot multiple patterns.<\/span><\/p>\n Longitudinal studies and complementary data visualisation<\/b><\/span><\/p>\n With such an enormous quantity of data, it can be a challenge to concisely communicate insights at a glance \u2014 particularly with a blend of qualitative and quantitative data. Even the second matrix, with its clean overview of numbers and figures, didn\u2019t give an instantly graspable picture of the insights. It also lacked the colour of the qualitative data, which reflected the richness of the participants\u2019 personal experiences over the ten months. And yet the mass of text recording that richness wasn\u2019t easy to absorb.<\/span><\/p>\n To visualise the data and illustrate our insights, we synthesised our findings in user journeys and videos that present the key changes over time and add real faces and voices to the matrices. These longitudinal user journeys plotted participants\u2019 key health and fitness behaviour changes, as well as the motivators and barriers to reaching their goals. Distilling mountains of data, this overview communicated a 10-months story at a glance. This enabled us to identify new patterns of behaviour and common triggers of change.<\/span><\/p>\n