HELPING STUDENTS MASTER HIGHLY COMPLEX DISCOURSE IN INSTANCES VOID OF CONCEPTS: MAKING “FOSSIL ANATOMY” LEARNABLE

There are moments in science education which offer no problems to solve, no notions to discover and zero concepts to think deeply about: medical students memorizing the 200-plus bones or those in university-level Systematic Paleontology courses who must describe fossils using conventionalized descriptor-discourse: “the glabella extends to the anterior border, has distinct furrows and the facial suture is opisthoparian.” Combining words precisely into lexically dense descriptors that evoke images of fossils is tantamount to being literate in Systematics, as are descriptors of systems in Engineering, descriptors of symptoms in Medicine etc. Void of concepts “descriptors” represent a pure form of “language and literacy” in STEM education. Like learners elsewhere, such discourse challenged our Systematics students, motivating us to design discourse-targeted learning materials to help them master fossil-descriptors. What started as a small-scale pilot of these materials evolved into a 4-year longitudinal between-group empirical comparison. Here, we discuss the learning theories guiding the development of descriptor-focusing tasks and present a set of tasks to illustrate how instructional design prompts the sustained use of descriptor language as students collaboratively completed tasks. More than 200 descriptors written during exams were analyzed using two criteria: lexical density (LD), the percentage of lexical words within a descriptor and structures described (SD), the percentage of anatomical structures mentioned with respect to those taught. To normalize descriptor quality across different fossil groups, LD and SD were factored into a descriptor index (DI), which increases with increasing quality of descriptors, i.e. more structures described within a more lexically dense text. Descriptors produced by Natural Science (NS) students who had learnt using the descriptor-focusing materials (n=127; academic years 2012-2015) were superior in quality than those written by Geological Sciences (GS) students following traditional teaching methods (n=102; 2013-2015). The values LD, SD and DI were statistically significantly higher in the descriptors of NS students (p<=0.05), and relatively constant over time (LD ranged from 0.47-0.67; SSD from 0.4-0.82; DI from 0.30-0.54) compared to descriptors of GS students (LD ranged from 0.29-0.39; SD from 0.40- 0.50; DI from 0.12-0.16). The average LD of fossil descriptors in textbooks ranges from 0.7-0.8, indicating that the instructional tasks enabled SN students to successfully master the dense and difficult language of descriptor discourse. We suggest how various task types can be adapted to other instances in education that are void of concepts, where the only objective is “learning discourse”. Identifying simple variables that make it possible for non-linguists and also students to objectively evaluate the quality of descriptors or texts that students have produced will bring “language and literacy” to the fore of STEM education.