The purpose of this STEP Type 2 project is to improve undergraduate retention in the biological sciences through the reformulation of instruction in the core course sequences. Prior research using a national sample indicates 90% of undergraduates who leave STEM majors cite poor instruction as a primary cause. Further, 74% of students successfully graduating from their STEM programs identify poor instruction as a major obstacle (Seymour and Hewitt, 1997). Using a double-blind design, the proposed study is testing the hypothesis that the lack of explicit instructions in scientific problem solving is a major factor in low STEM retention. Typically, experts in a field (e.g., professors in biological sciences) have automated their procedural knowledge during years of training and professional practice (Anderson, 2005; Bereiter and Scardamalia, 1993; Feldon, in press). As such, when they provide instruction through lectures or mentoring conversations, they frequently omit many of the steps necessary to complete a problem without necessarily realizing that they are doing so (Chao and Salvendy, 1994; Feldon, 2006). However, cognitive task analysis (CTA) is emerging as a viable means to elicit accurate and complete information from experts to serve as the basis for highly effective instruction (Feldon, in press). The Intellectual Merit of the project is its application of robust findings in cognitive science research to STEM instruction. For several decades, cognitive psychologists have called for greater consideration of skill automaticity in the design of complex instruction (e.g., Schneider, 1985; Clark and Estes, 1996; Rogers, Maurer, Salas, and Fisk, 1997). Those training systems that have explicitly accommodated the tacit nature of their subject matter experts' knowledge have proven to be significantly more effective than those that have not (e.g., Merrill, 2002; Schaafstal, Schraagen, and van Berlo, 2000; Velmahos et al., 2004). However, university based STEM instruction has not reflected any consideration of these advances in instructional design (van Merrienboer, 1997). The resulting gaps in instructional content induce a much higher level of cognitive load in learners that typically results in less effective learning and significant drops in motivation for challenging material (Britt, 2005; Kirschner, Sweller, and Clark, 2006; Paas, Tuovinen, van Merrienboer, and Darabi, 2005). Because this study encompasses the instructional design process and implementation for core laboratory courses in genetics, molecular, and cell biology, it is well-positioned to (1) document the knowledge gaps typically included in skill-based STEM instruction, (2) systematically eliminate those gaps in the treatment group, and (3) longitudinally track the impact of each instructional condition over the course of students' undergraduate coursework in biology. It is expected that students in the CTA treatment condition will be more likely to perform better in lab-based coursework and remain in the biological sciences major to a greater degree than their counterparts in the control condition. The Broad Impacts of this project exist at two levels. First, it is determining the extent to which explicit, comprehensive problem solving instruction contributes to STEM retention. Previous descriptive studies provide strong evidence of a correlation, but few experimental studies have examined the direct causal relationship (Seymour, 2001). Second, this project is validating a general model of instructional design for STEM disciplines that can be easily transferred and adapted across fields and institutions. Instructional design methods that utilize cognitive task analysis methods to identify and generate curriculum have had a major impact on the effectiveness and efficiency of complex skills training in many non-academic arenas (Clark et al., in press). This study can validate and leverage its strengths to enhance the preparation of future scientists and provide stronger scientific reasoning skills for college graduates who enter the workforce in science-related fields.
