Interactivity of Question Prompts and Feedback on Secondary Students' Science Knowledge Acquisition and Cognitive Load

Introduction Problem solving is an essential 21st century skill (Trilling & Fadel, 2009), and is continuously incorporated as an integral part of school curricula (Barron, 2000; Barrows & Tamblyn, 1980; Qin, Johnson, & Johnson, 1995). As web-based learning becomes the mainstream in many educational settings, it is increasingly important to adopt research-based guidelines to design effective web-based instruction for problem solving. Two research foci have become prominent: scaffolding and cognitive load. Scaffolding research focuses on designing tools and strategies that provide learners with optimal support as they work on learning tasks. A large body of research investigated the scaffolding of open-ended or ill-structured problem solving (e.g., Chen, 2010; Davis, 2000). Closely related to scaffolding is the provision of feedback to students' performance as a way to support learning (e.g., Bangert-Drowns, Kulik, Kulik, & Morgan, 1991; Corbalan, Kester, & van Merrienboer, 2009). Cognitive load theories approach the design of web-based learning from a different angle. Focusing on making optimal use of humans' limited working memory, cognitive load theorists have identified design principles and guidelines that minimize extraneous cognitive load while focusing learners' cognitive resources on tasks directly related to learning (Sweller, 2010; Sweller, Ayres & Kalyuga, 2011). While the two lines of research relate to each other, they do not intersect much in the research literature. Presumably, scaffolding which is intended to support problem solving should help to streamline learners' cognitive processes and facilitate schema construction. Yet existing research appears to cast some doubts on such a presumption (e.g., Chen, Wu, & Jen, 2013; Ge, Chen, & Davis, 2005; Hwang, Kuo, Chen, & Ho, 2014). Therefore, this study set out to implement scaffolding and feedback in a web-based learning environment to support students' problem solving in science, for the purpose of examining how scaffolding and feedback impact learners' cognitive load as well as knowledge acquisition. Theoretical framework Scaffolding problem solving in science Solving problems is an essential practice in the disciplines of science. Polya (1957) proposed an influential model to characterize a four-step process in problem solving: understanding the problem, planning a solution, executing the plan, and checking the result. In the context of solving a science problem, the process involves identifying relevant information in the problem, determining known and unknown concepts, selecting rules or principles applicable to the problem, applying rules or principles, and ensuring that a satisfactory solution is reached (Jonassen, 2000; Simon, 1978). In other words, learners are in on the active construction, manipulation, and testing of mental models of the problem (Jonassen, 2011). While problem-solving processes are known to researchers and intuitive to skillful problem solvers, students are often not strategic in their problem-solving approaches. Instead of taking time to comprehend a problem and build a conceptual model of it, learners often jump quickly to solutions (Bransford, Brown, & Cocking, 1999). As such, researchers have employed various strategies to scaffold students through the problem-solving process (Arnau, Arevalillo-Herraez, Puig, & Gonzalez-Calero, 2013; Fund, 2007; Palinscar, 1986; Rosenshine & Meister, 1992; Rosenshine, Meister, & Chapman, 1996; van Merrienboer, Kirschner, & Kester, 2003). Question prompting is a frequently used approach to scaffolding learners' problem solving (e.g., Chen, 2010; Ge, Chen, & Davis, 2005; Saye & Brush, 2002). By presenting questions to students, question prompts focus students' attention on relevant aspects of problem solving and guide them through the process. Numerous studies have found the effectiveness of question prompts in promoting problem solving, knowledge acquisition, and metacognition (Davis, 2000; Rae, Schellens, Wever, & Vanderhoven, 2012; Zydney, 2010). …