Susanne Lajoie | McGill University (original) (raw)

Papers by Susanne Lajoie

Advances in analytics for learning and teaching, 2023

Social sciences & humanities open, 2023

Purdue University Press eBooks, Aug 15, 2020

Journal of computers in education, Jul 19, 2023

Self-regulation of learning (SRL) involves students controlling of their own cognition, behaviour... more Self-regulation of learning (SRL) involves students controlling of their own cognition, behaviour, emotions and motivation through metacognitive awareness of their conditions and situations (Winne & Hadwin, 1998). While this self-reflective process is argued to be adaptive, students may also exhibit maladaptive responses through poorly orchestrated SRL skills. What is considered maladaptive, however, may be relative to one’s goal; the same behaviour can be adaptive in one’s current context but maladaptive for the long term development. This structured poster symposium brings together researchers using a variety of methods and approaches to conceptualize and capture adaptive regulatory patterns and processes. A range of research will be presented, including: using an online diary tool to reveal patterns of adaptive and maladaptive regulation around self-reported challenges and strategies, how self-handicapping affects SRL processes while completing a webquest, using multimodal data to distinguish adaptive and maladaptive regulation, and the interplay between mental health and SRL. Issues and future directions related to examining students’ adaptive and maladaptive regulation will be discussed in this session.

Computer Supported Collaborative Learning, 2013

Journal of Computer Assisted Learning, May 11, 2023

BackgroundComputer‐based scaffolding has been intensively used to facilitate students' self‐r... more BackgroundComputer‐based scaffolding has been intensively used to facilitate students' self‐regulated learning (SRL). However, most previous studies investigated how computer‐based scaffoldings affected the cognitive aspect of SRL, such as knowledge gains and understanding levels. In contrast, more evidence is needed to examine the effects of scaffolding on the metacognitive dimension and efficiency outcome of SRL.ObjectivesThis study aims to examine the role of computer‐based scaffolding in students' metacognitive monitoring and problem‐solving efficiency.MethodsSeventy‐two medical students completed two clinical reasoning tasks in BioWorld, an intelligent tutoring system (ITS) designed for promoting medical students' diagnostic expertise. During solving the tasks, students were asked to report their confidence judgements about proposed diagnoses. Computer trace data were used to identify task completion time (CT) and students' use of three scaffolding types, that is, conceptual, strategic, and metacognitive. Then we calculated students' metacognitive monitoring accuracy (i.e., calibration) and problem‐solving efficiency.Results and ConclusionsOne‐sample t‐test demonstrated that students inaccurately monitored their learning processes and were overconfident in both tasks. Linear mixed‐effects models (LMMs) indicated that the intensive use of metacognitive scaffolding positively predicted students' metacognitive monitoring accuracy. Moreover, strategic scaffolding was negatively related to problem‐solving efficiency, whereas metacognitive scaffolding positively influenced problem‐solving efficiency.TakeawaysThis study shows the importance of metacognitive scaffolding in improving the accuracy of metacognitive monitoring and problem‐solving efficiency. Findings from this study provide new insights for instructors and ITS developers to optimise the design of scaffoldings.

Computers in Human Behavior, Jul 1, 2023

Interdisciplinary Journal of Problem-based Learning, Dec 26, 2021

Interactive Learning Environments, Mar 5, 2021

This study applies a time-driven approach to model self-regulated learning (SRL) on the basis of ... more This study applies a time-driven approach to model self-regulated learning (SRL) on the basis of elapsed time metrics in the context of open-ended learning environments (OELEs), specifically, netwo...

Educational Technology & Society, Apr 1, 2016

Introduction Learner modeling is a critical component in the process of adapting instruction to t... more Introduction Learner modeling is a critical component in the process of adapting instruction to the specific needs of different learners with computerized systems. Adaptive instructional systems may be defined as a systematic process which consists of four steps: (1) capturing information about the learner; (2) analyzing learner interactions through a model of learner characteristics in relation to the domain; (3) selecting the appropriate instructional content and resources; and (4) delivering the content to the learner (Shute & Zapata-Rivera, 2012). The analytical function of the learner model can be further classified in terms of processes conducted at both the macro and micro levels (VanLehn, 2006). At the macro-level, a representation of the path towards competency within the domain is updated for each task with the aim of selecting the next task that is the most appropriate for the learner. At the micro-level, instructional materials such as hints and feedback are delivered to the learner on the basis of a representation that is repeatedly updated over the duration of task performance. Intelligent tutoring systems have been shown to improve upon typical classroom instruction in several disciplines, including mathematics (e.g., Cognitive Tutors; Koedinger & Corbett, 2006), computer science (e.g., Constraint-Based Tutors; Mitrovic, 2003), microbiology (e.g., Narrative-Based Tutors; Rowe, Shores, Mott, & Lester, 2011), as well as physics and computer literacy (e.g., Dialogue-Based Tutors; Graesser, VanLehn, Rose, Jordan, & Harter, 2001). The challenge in modeling learning in the context of solving ill-structured problems is that there are multiple paths towards attaining the correct solution (Lajoie, 2003, 2009). These paths should be documented in order to represent common misconceptions or impasses in moving along the trajectory towards competency. In the medical domain, experts have been shown to diagnose patient diseases in different ways, while at the same time, justifying plausible hypotheses on the basis of common evidence items, including patient symptoms and lab test information (Gauthier & Lajoie, 2014; Lajoie, Gauthier, & Lu, 2009). An intelligent tutoring system such as BioWorld can represent these evidence items in terms of a novice-expert overlay system, which compares novice solution paths to those of the experts in order to individualize feedback. A similar modeling technique has been applied in other systems, such as SlideTutor in diagnosing dermatopathology (Feyzi-Behnagh, Azevedo, Legowski, Reitmeyer, Tseytlin, & Crowley, 2014) as well as the MedU virtual patient cases (Berman, Fall, Chessman, Dell, Lang, Leong, Nixon, & Smith, 2011). On the basis of the user interactions that are recorded by BioWorld, the system will highlight areas of similarities and differences with the expert solution path, allowing novices to self-reflect on their own approach to resolving the problem (Lajoie & Poitras, 2014). Several studies investigating the novice-expert overlay model have been carried out with the aim of capturing linguistic features from written case summaries to tailor feedback content (Poitras, Doleck, & Lajoie, 2014; Lajoie, Poitras, Doleck, & Jarrell, 2015) as well as the impact of goal-orientations and affective reactions towards attention given to feedback (Lajoie, Naismith, Poitras, Hong, Panesso-Cruz, Ranelluci, & Wiseman, 2013). In this study, we examine the use of subgroup discovery for the induction of rules that characterize the relationship between impasses in problem-solving and lab-tests ordered in BioWorld. Whereas our previous research characterized how experts converged in their paths to solving a problem, the present study captures how novices diverged from an expert solution path. In doing so, we claim that subgroup discovery algorithms are particularly well suited towards describing the multiple paths that characterize problem-solving in ill-structured domains. …