Blair Lehman

Guru: a computer tutor that models expert human tutors

We present Guru, an intelligent tutoring system for high school biology that has conversations with students, gestures and points to virtual instructional materials, and presents exercises for extended practice. Gurus instructional strategies are modeled after expert tutors and focus on brief interactive lectures followed by rounds of scaffolding as well as summarizing, concept mapping, and Cloze tasks. This paper describes the Guru session and presents learning outcomes from an in-school study comparing Guru, human tutoring, and classroom instruction. Results indicated significant learning gains for students in the Guru and human tutoring conditions compared to classroom controls.

A Motivationally Supportive Affect-Sensitive AutoTutor

This chapter describes a fully automated affect-sensitive Intelligent Tutoring System (ITS) called the Affective AutoTutor. AutoTutor is an ITS that helps students learn topics in Newtonian physics, computer literacy, and critical thinking via natural language dialogues that simulate the dialogue patterns observed in human–human tutoring. AutoTutor uses state-of-the-art natural language understanding mechanisms to model learners’ cognitive states and plan its dialogue moves in a manner that is sensitive to these states. While the original AutoTutor is sensitive to learners’ cognitive states, the affect-sensitive tutor is responsive to their affective states as well. This Affective tutor automatically detects learners’ boredom, confusion, and frustration by monitoring conversational cues, gross body language, and facial features. The sensed affective states guide the tutor’s responses in a manner that helps learners regulate their negative emotions. The tutor also synthesizes affect via the verbal content of its responses and the facial expressions and speech of an embodied pedagogical agent. An experiment comparing the affect-sensitive and nonaffective tutors indicated that the affective tutor improved learning for low-domain knowledge learners, particularly at deeper levels of comprehension.

Inducing and tracking confusion with contradictions during critical thinking and scientific reasoning

Cognitive disequilibrium and its affiliated affective state of confusion have been found to be beneficial to learning due to the effortful cognitive activities that accompany their experience. Although confusion naturally occurs during learning, it can be induced and scaffolded to increase learning opportunities. We addressed the possibility of induction in a study where learners engaged in trialogues on critical thinking and scientific reasoning topics with animated tutor and student agents. Confusion was induced by staging disagreements and contradictions between the animated agents, and the (human) learners were invited to provide their opinions. Self-reports of confusion and learner responses to embedded forced-choice questions indicated that the contradictions were successful at inducing confusion in the minds of the learners. The contradictions also resulted in enhanced learning gains under certain conditions.

What Makes Learning Fun? Exploring the Influence of Choice and Difficulty on Mind Wandering and Engagement during Learning

Maintaining learner engagement is critical for all types of learning technologies. This study investigated how choice over a learning topic and the difficulty of the materials influenced mind wandering, engagement, and learning during a computerized learning task. 59 participants were randomly assigned to a text difficulty and choice condition (i.e., self-selected or experimenter-selected topic) and measures of mind wandering and engagement were collected during learning. Participants who studied the difficult version of the texts reported significantly higher rates of mind wandering (d = .41) and lower arousal both during (d = .52) and after the learning session (d = .48). Mind wandering and arousal were not affected by choice. However, participants who were assigned to study the topic they selected reported significantly more positive valence during (d = .57) but not after learning. These participants also scored substantially higher on a subsequent knowledge test (d = 1.27). These results suggest that choice and text difficulty differentially impact mind wandering, engagement, and learning and provide important considerations for the design of ITSs and serious games with a reading component.

Interventions to regulate confusion during learning

Experiences of confusion have been found to correlate with learning, particularly for learning at deeper levels of comprehension. Previously, we have induced confusion within learning environments that teach critical scientific reasoning. Confusion was successfully induced with the presentation of contradictory information and false feedback. Next, we would like to regulate experiences of confusion to increase learning. In the current paper, we propose a series of experiments that investigate potential interventions to help regulate confusion during learning. Specifically, these experiments will address the impact of feedback specificity and emotional support.

How do they do it? investigating dialogue moves within dialogue modes in expert human tutoring

Expert human tutors are widely considered to be the gold standard for increasing student learning. While not every student has access to an expert tutor, it is possible to model intelligent tutoring systems after expert tutors. In an effort to achieve this goal, we have analyzed a corpus of 50 hours of one-to-one expert human tutoring sessions. This corpus was coded for speech acts (dialogue moves) and larger pedagogical strategies (dialogue modes). Using mixed-effects modeling, we found that expert tutors differentially used dialogue moves depending on the dialogue mode. Specifically, tutor posed questions, explanations, and motivational statements were predictive of different dialogue modes (e.g., Lecture, Scaffolding).

To Resolve or not to Resolve? that is the Big Question About Confusion

Positive relationships between confusion and learning have been found for the last decade. Most theoretical foundations for confusion hypothesize that it is not the mere occurrence of confusion, but rather the successful resolution that benefits learning. Empirical research has provided some support for this hypothesis, but investigations of the confusion resolution process are still sparse. The present work is a preliminary investigation of the confusion resolution process within two learning environments that experimentally induce confusion (false feedback, contradictory information). Findings showed that learners did benefit from confusion resolution compared to when confusion was unresolved, but it was not merely from increased effort. The nature of the confusion induction method also influenced the positive impact of confusion resolution on learning. Implications for intelligent tutoring systems are discussed.

Automatic evaluation of learner self-explanations and erroneous responses for dialogue-based ITSs

Self-explanations (SE) are an effective method to promote learning because they can help students identify gaps and inconsistencies in their knowledge and revise their faulty mental models. Given this potential, it is beneficial for intelligent tutoring systems (ITS) to promote SEs and adaptively respond based on SE quality. We developed and evaluated classification models using combinations of SE content (e.g., inverse weighted word-overlap) and contextual cues (e.g., SE response time, topic being discussed). SEs were coded based on correctness and presence of different types of errors. We achieved some success at classifying SE quality using SE content and context. For correct vs. incorrect discrimination, context-based features were more effective, whereas content-based features were more effective when classifying different types of errors. Implications for automatic assessment of learner SEs by ITSs are discussed.

Who Benefits from Confusion Induction during Learning? An Individual Differences Cluster Analysis

Recent research has indicated that learning environments that intentionally induce confusion to promote deep inquiry can be beneficial for learning if students engage in confusion resolution processes and if relevant scaffolds are provided. However, it is unlikely that these environments will benefit all students, so it is necessary to identify the student profiles that most benefit from confusion induction. We investigated how individual differences (e.g., prior knowledge, interest, attributional complexity) impacted confusion and learning outcomes in an environment that induced confusion via false system feedback (e.g., negative feedback after a correct response). A k-means cluster analysis revealed four clusters that varied on cognitive ability and cognitive drive. We found that students in the high cognitive ability + high cognitive drive cluster reported more confusion after receiving false feedback compared to the other clusters. These students also performed better on tasks requiring knowledge transfer, but only when they were meaningfully confused.

Impact of Agent Role on Confusion Induction and Learning

The presentation of contradictory information to trigger deeper processing and increase learning has been investigated in a variety of ways (e.g., conversational agents, worked examples). However, the impact of information source (e.g., expertise, gender) and the relationship between the contradicting sources (e.g., status level) has not been investigated to the same degree. We previously reported that confusion can successfully be induced and learning increased when contradictory information was presented by two conversational agents (tutor, peer student). In the present experiment we investigated contradictions posed by two peer student agents. Self-reports of confusion and learner responses to embedded forced-choice questions revealed that the contradictions still successfully induced confusion. There were, however, differences in the nature of confusion induction based on the inter-agent relationship (i.e., student-student vs. tutor-student). Learners performed better on transfer tasks when presented with contradictions compared to a no-contradiction control, but only when they were successfully confused.