This Fostering Interdisciplinary Research in Education (FIRE) project is a partnership between Columbia University and Temple University that seeks to understand the perceptual and cognitive processes by which undergraduate students and scientists transform earth science data into explanations and prediction.
The principal investigator of this FIRE award, a geoscientist, will be working with a mentor in spatial cognitive science to continue to develop her expertise in this area and eye tracking in particular. The researcher proposes two studies using eye-tracking and think aloud protocols. The first study will investigate students? perceptions and reasoning as they interpret a data set (from topography/bathymetry) that is known to be subject to misinterpretation. Study two is an intervention study, testing whether providing hypothesis templates will help students interpret a three-dimensional data volume in the area of ocean salinity. The research will be evaluated by experts in spatial cognition, epistemology of data, visualization, and earth science.
In the geosciences, many students have a difficult time learning the skills that are necessary to interpret maps and large amounts of spatial data. Such skills are central to what geoscientists do and how knowledge is rendered in this discipline. This study could lead to improvements in how students learn the skills necessary to visualize geoscience data and make claims based on data interpretation. The project clearly advances discovery and promote teaching and learning in the geosciences. The research may also have implications in the learning in other fields of science, technology, engineering, and mathematics that place high demands on a student's visualization abilities.
Being able to make meaning from data is central to science, and it is becoming more important for non-scientists as society becomes more data infused. How people take a data visualization in the form of a graph or map and go from the dots or squiggles or blotches of color to some inference about structures or processes in the real world is poorly understood. This project was funded by a special program at NSF that funds collaborations between scientists in two different disciplines, for the purpose of building capacity in research in science education. Our collaboration was between a geoscientist (KK) and a cognitive scientist (TS). The project had 3 main components: (1) a study comparing how experts and novices interpret data visualizations of landforms on continents and on the seafloor (2) a study of how college students interpret data visualizations of oceanographic data, and (3) an interdisciplinary graduate seminar. Both study 1 and study 2 used similar techniques. On a computer monitor, participants viewed maps that showed data using different colors. For each map, the experimenter asked participants to describe what they were seeing and asked questions probing what inferences they were making about the data. The verbal responses and gestures were recorded by video. An eye-tracking device recorded where their gaze was focused as they examined the data. Eye-tracking data can provide insights about what the participants think is important in the image, and about what strategies they are using to coordinate information from different parts of an image. Study 1 compared the behaviors and interpretations of expert geoscientists versus novices with little to no Earth Science education. As expected, the experts were much more skillful at describing and explaining what they were seeing. However, to our surprise, the experts and novices invested their exploration time similarly, spending approximately the same amount of time, small or large, on the various features in the data. To a first approximation, they are looking at the same things, but extracting different meaning. From an education perspective, this means that there is not a simple silver bullet for improving data literacy by redirecting students’ attention, as had been found for some previously researched learning tasks. Study 2 looked at college students who were enrolled in an environmental science course where they had had some exposure to concepts and data related to the data they were viewing during the experiment. We found that providing the students with an array of candidate interpretations helped them organize their exploration of complex 3D data more productively. This is a promising strategy for use in curriculum materials, not just in geosciences, but across the disciplines. The interdisciplinary seminar was designed to build capacity for researching how people think and learn with data. Participants were Temple University graduate students and postdocs in cognitive science. We used a series of historical case studies of important scientific discoveries that were enabled by new data to understand the spatial reasoning employed by the scientists. In sum, the funding supported development of interdisciplinary research skills allowing the PIs to build their research programs that span disciplinary science practice, cognitive science, and education.
