Academic journal article Educational Technology & Society

Modeling and Intervening across Time in Scientific Inquiry Exploratory Learning Environment

Academic journal article Educational Technology & Society

Modeling and Intervening across Time in Scientific Inquiry Exploratory Learning Environment

Article excerpt

Introduction

Researchers in the field of science education have been employing scientific inquiry as an instructional strategy to maximize learner's engagement and experiences during the learning processes (Frederiksen & White, 1998; Hulshof & de Jong, 2006; Linn, 2000; Pryor & Soloway, 1997; Reiser et al., 2001; Shimoda, While & Frederiksen, 2002; Veermans & van Joolingen, 2004). The importance of this instructional strategy is proven by the increasing number of computer-assisted learning environments developed recently such as the Belvedere (Frederiksen & White, 1998), BGuiILE (Pryor & Soloway, 1997), KIE (Linn, 2000), SCI-WISE (Reiser et al., 2001), SimQuest (Veermans & van Joolingen, 2004), Rashi (Dragon et al., 2006), and SmithTown (Shute & Glaser, 1990). These learning environments do not directly deliver scientific facts to learners. Instead, learners are required to actively involve in scientific inquiry processes such as evidence gathering, constructing and testing hypotheses, manipulating variables, and the like. As an attempt to maximize scientific inquiry learning experience, exploratory learning approach is often the preference when it comes to development of computer-based learning environments (Chang et al., 2003; de Jong & Van Joolingen, 1998; de Jong, 2006). The exploratory learning approach provides learners with freedom to interact with the learning environment by means of test-and-retest their idea.

To date, although there has been an attempt to integrate intelligence into scientific inquiry learning environment by employing a learner model (Meyer et al., 1999), solutions to the following challenges remain unclear. Firstly, how should a learner model be integrated into a scientific inquiry learning environment that consists of more than a single GUI and rooted in a particular instructional model? Secondly, having gathered a learner's exploratory behaviours, how should a system effectively assess the mastery levels of scientific inquiry skills which evolve across time? Thirdly, how should a system generate tailored pedagogical interventions in a timely manner to cater for temporally variable scientific inquiry skills? These challenges are not trivial as the system has to deal with a high level of uncertainty inherent in inferring a learner's mastery level of scientific inquiry skills from exploratory behaviours (Schum, 1994; de Jong, 2006).

To handle the uncertainty in an efficient modeling of interaction behaviours, Bayesian Network (BN) (Pearl, 1998) has been employed as one of the solutions (Jameson, 1995). However, a BN does not provide decision making under uncertainty. As a complementary solution, a Decision Network (DN) (Howard & Matheson, 1981; Russell & Norvig, 1995), which is an extension of BN with decision and utility nodes, is proposed by researchers in the field of Artificial Intelligence in Education (e.g., Murray et al., 2004; Conati, 2002, Pek & Poh, 2005). When time is a crucial factor in learner modeling, a static DN is then extended to a Dynamic Decision Network (DDN). By employing a DDN, not only the system is able to model the variables that evolve across time, but at the same time provides tailored feedback in a timely manner. A study by Murray et al. (2004) has shown that the decision-theoretic approach outperforms Fixed-Policy Tutor in selecting the optimal tutorial action. Here, we review instances of decision-theoretic Intelligent Tutoring Systems (ITSs) such as DT Tutor (Murray et al., 2004), CAPIT (Mayo & Mitrovic, 2001), Prime Climb (Conati, 2002), and iTutor (Pek & Poh, 2005), and subsequently highlight the major differences between their work and ours.

DT Tutor is an ITS designed for two domains, namely, the Calculus-related rate problems and Elementary reading. DT Tutor takes into consideration of not only the learner's goals, focus of attention, and affective state, but also its objectives to provide optimal pedagogical decision. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.