W05: Harnessing the Power of Longitudinal Qualitative Data

In this interactive workshop, the facilitators introduced methods and strategies for longitudinal qualitative research (LQR). They introduced LQR as a very powerful way to describe and explain change over time. Qualitative data from open-ended survey questions, in-depth interviews, focus groups, and even participant observation can help researchers understand the processes and factors that influence behaviors and decision-making. By collecting and analyzing qualitative data longitudinally, program evaluations and research studies are able to capture the complex factors and conditions that lead to a particular outcome, such as a career decision.

The most basic definition of LQR is as a method that explores change over time, explained Robin Remich, Research Associate at the Northwestern University Feinberg School of Medicine. The goals of this method are to describe and explore the how and why of observed change while considering contextual elements that might influence individual experiences. One major benefit to longitudinal qualitative approaches is the ability to capture a person’s thinking and decision making in vivo, giving researchers a chance to understand more deeply the contexts, feelings, and thoughts that factor into people’s decisions or actions.

When using longitudinal methods, Remich continued, an important question is how much time is adequate to address an issue and how to choose the intervals for data collection. These factors may change over the course of the study, she explained, but thinking about them beforehand has major benefits, particularly when trying to fund a project.

Remich’s colleague Christine Wood discussed the different levels of change referred to by social scientists. The top level is macro-sociological change, which happens within a national context and applies to broad swaths of society. The middle or meso-level encompasses spheres between the individual and broader society and includes studies of organizations and communities. The third level is micro-sociological change, which happens with individuals or small groups such as schools or neighborhoods. “One thing to keep in mind when you’re designing a longitudinal qualitative study is how these different levels may interact with one another,” she said, giving as an example the influence of societal trends on individual students.

Another important factor is sample size and the longevity of your research team, Remich said. When collecting data through face-to-face interviews, an important factor is whether the same interviewers will be with the study long term, because keeping subjects in a study is often easier when they have a good rapport with researchers. She also suggested piloting questions and getting feedback before data collection begins.

Remich, Wood, and their colleague Remi Jones presented examples from their ongoing study, begun in 2008, which examines how the career intentions of PhD students in the biomedical sciences change over time. The study uses one-on-one interviews of about an hour and a half conducted annually with more than 200 participants. Students were enrolled in the study as juniors, post-baccalaureate, and first year PhD students.

Timelines and flowcharts are tremendously useful in organizing longitudinal data, Remich explained. These allow researchers to pull out snapshots of students at one point in time and also look at patterns within the group over time. Major events that provide context can be noted in the timeline data, such as the NIH funding sequestration and natural disasters, because these might have influenced students’ experiences and actions.

Another tool that the researchers demonstrated is a coding rubric, which they used to organize and identify subgroups in their large data set. Remich and her colleagues used a scale from -2 to +2 to assess career intention at each interview for three kinds of academic careers (research-focused, teaching-focused, and research/teaching), industry, and other careers. The coding rubric served as a more quantitative way to analyze the data and to identify sub-groups for more in-depth qualitative analysis. After using the rubric, the researchers displayed their data using Excel and Adobe Illustrator to map career decisions over time. “This was a process for us to see possible patterns that we have with these 200 people that would allow us to ask interesting questions,” Remich said.

At this point, Remich introduced the workshop participants to the work of “Saldaña (2003) who proposes seven descriptive questions for interrogating longitudinal data effectively. These questions guide researchers to examine their data for instances of things such as what emerges, accumulates, and stays consistent over time, and when an epiphany occurs. The facilitators concluded the workshop by reflecting on the patience and multiple reading of data involved in longitudinal qualitative analysis and distributed an LQR methods resources list.

Don't forget to join us in San Antonio for the 2017 UI Conference!

Our 9th Conference on Understanding Interventions that Broaden Participation in Science Careers will be held at the Sheraton Gunter Hotel, in downtown San Antonio.