Crossing the threshold : Visualization design and conceptual understanding of evolution

Abstract: The theory of evolution is considered the unifying theory of biology. An accurate understanding of evolution is vital both for the understanding of diverse topics in biology, but also for societal issues such as antibiotic resistance or biodiversity. In contrast, decades of research in science education have revealed that students have difficulties to accurately understand evolutionary processes such as mutation and natural selection. The majority of this research relies on a conceptual framework of so-called key concepts (variation, selection, inheritance), derived from scholarly descriptions of natural selection. Recent research suggests that non-domain specific concepts such as randomness, probability, spatial and temporal scales, so called threshold concepts, are important for evolution understanding in addition to the key concepts. Thus, many important elements of evolutionary theory are counter-intuitive or lie outside direct perception. Hence, representations such as visualizations, models and simulations are considered to be important for teaching and learning evolution. While the importance of visualizations is generally acknowledged for science education, less is known about how visual design can facilitate students understanding of threshold concepts, such as random mutations or spatial scales.This thesis uses the Model of Educational Reconstruction (MER) as the guiding framework for exploring the significance of threshold concepts by analysing the conceptual content of students’ explanations and extant visualizations of natural selection. MER combines scientific content with teaching and learning perspectives for the analysis and design of learning environments. Content analysis of visualizations available online showed that most fail to fully represent the basic principles of natural selection (variation, selection and inheritance). Moreover, the representational potential of visualizations was seldom used to represent threshold concepts such as randomness in origin of variation. Visualizations were also biased to animals as the context of evolution. Similarly, upper-secondary and tertiary students’ explanations of natural selection were seldom complete in terms of the basic principles and threshold concepts such as randomness were often lacking. Especially significant was the almost complete lack of randomness in upper-secondary students’ explanations. In addition, threshold concepts were context-sensitive across the items used (bacteria, cheetah and salamander), for example spatial scale and randomness was significantly more common in responses to the bacteria item compared to the cheetah and salamander items. Considering the results from these studies, three interactive visualizations were developed (evolution of antibiotic resistance and fur colouration in mice). The visualization design was conducted iteratively following a Design-Based Research approach and evaluated in classroom settings in secondary and upper-secondary Swedish schools. The results showed that visualizations targeting randomness and genetic level events such as mutations can guide students towards a more scientific conception of natural selection. However, there were differences across the visualizations and student samples. In addition, while students often inferred randomness from the visuals, the results showed that integration of randomness into explanations of natural selection may be challenging. Hence, future research should explore the role of guidance and reflection for students understanding of randomness.  The thesis also discusses the role of students’ intuitive conceptions in relation to the use of interactive visualizations and how these preconceptions interact with the presented message. By using the theory of frame semantics, framing effects and conceptual integration, students’ issues of achieving an accurate understanding of evolution are discussed in relation to the theory of conceptual change. Implications for teaching and learning natural selection as well as visualization design for learning are also discussed.

  This dissertation MIGHT be available in PDF-format. Check this page to see if it is available for download.