Search

Pedagogical agents are on-screen characters that help facilitate learning in a virtual or mixed-reality setting. The research on pedagogical agents has shifted focus since their creation, from understanding their underlying principles, identifying their potential roles and usage, and recently finetuning their design for applied practice. This paper explores the history of pedagogical agents and summarizes the current state of knowledge. Once the current state of pedagogical agents is established, barriers and opportunities are discussed to suggest how researchers can best improve pedagogical agents and their implementation.

Virtual humans are on-screen characters that are often embedded in learning technologies to deliver educational content. Little research has investigated how virtual humans can be used to correct common misconceptions. In this study, we explored how different types of narrative structures, refutation text and expository text, influence perceptions of trust, credibility, and learning outcomes. In addition, we conducted exploratory analyses examining how different measures of trust and credibility are related to each other and how these measures may mediate learning outcomes. Results showed that the type of narrative used did not influence any measure. However, the trust and credibility measures, while related to one another, were measurably distinct. In addition, only perceptions of message trust were significantly related to learning. Perceptions of message trust did not mediate learning outcomes, but were significantly predictive of learning at nearly the same effect as prior knowledge.

As the areas of education and workforce training tend more toward online training, they have emphasized the rapid setup and expansion of online environments. This scaling-up and scaling-out of training programs can be difficult, but there are examples of success available for us to learn from. To capture the lessons learned, we conducted semi-structured interviews across academic, public, private, and non-profit organizations. Using deductive thematic analysis, three major groups of determinants were identified: technological infrastructure, human infrastructure, and governance. A framework consisting of three actionable recommendations was created to facilitate online learning organizational scaling-up including learner (learning experiences and user-friendly design), technical (technical competencies and the use of integrated data), and management (technical competencies for both learners and instructors as well as the use of integrated data) factors.

Virtual humans are virtual characters within multimedia learning environments designed to aid the learning process. While there is a large variety of research examining how to design the physical appearance of the character or the teaching strategies it should embody, there is comparatively little work around the design of the narrative the virtual human uses to communicate with the learner. In this study, we examine the use of four different types of text structures to structure the character's narrative in an instructional video: expository, refutation text, deep reasoning questions, and refutation text and deep reasoning questions; and their effects on learning, perceptions, attitudes, and emotions about genetically modified foods. Our results largely indicated no significant differences between the different communication strategies. We hypothesize that the length, dosage, and pacing of the intervention could explain why we did not see benefits from the different narrative structures that have been found in other contexts.

ObjectivesTo evaluate the impact of bowel resection at the time of interval cytoreductive surgery on survival.MethodsWe identified patients with advanced ovarian cancer who underwent neoadjuvant chemotherapy and interval cytoreductive surgery between 2008 and 2018 from a single-institution tumor registry. Kaplan-Meier survival analysis and Cox proportional hazards models were performed comparing patients who underwent bowel resection to those who did not.ResultsOf 158 patients, 43 (27%) underwent bowel resection. Rates of optimal (95%) and sub-optimal (5%) resection did not differ with bowel resection. Patients that required bowel resection had worse three-year survival (43% vs. 63%), even after adjusting for confounding variables of age, stage, number of neoadjuvant cycles, R0 resection, and ASA score (HR 2.27, p

Virtual reality (VR) has a high potential to facilitate education. However, the design of many VR learning applications was criticized for lacking the guidance of explicit and appropriate learning theories. To advance the use of VR in effective instruction, this study proposed a model that extended the cognitive-affective theory of learning with media (CATLM) into a VR learning context and evaluated this model using a structural equation modeling (SEM) approach. Undergraduate students (n = 77) learned about the solar system in a VR environment over three sessions. Overall, the results supported the core principles and assumptions of CATLM in a VR context (CATLM-VR). In addition, the CATLM-VR model illustrated how immersive VR may impact learning. Specifically, immersion had an overall positive impact on user experience and motivation. However, the impact of immersion on cognitive load was uncertain, and that uncertainty made the final learning outcomes less predictable. Enhancing students' motivation and cognitive engagement may more directly increase learning achievement than increasing the level of immersion and may be more universally applicable in VR instruction.

Detection and responding to a player's affect are important for serious games. A method for this purpose was tested within Chem-o-crypt, a game that teaches chemical equation balancing. The game automatically detects boredom, flow, and frustration using the Affdex SDK from Affectiva. The sensed affective state is then used to adapt the game play in an attempt to engage the player in the game. A randomized controlled experiment incorporating a Dynamic Bayesian Network that compared results from groups with the affect-sensitive states vs those without revealed that measuring affect and adapting the game improved learning for low domain-knowledge participants.

This study investigated changes in learners' motivation, engagement, performance, and spatial reasoning over time and across different levels of virtual reality (VR) immersion. Undergraduate participants explored a virtual solar system via a moderately immersive or highly immersive VR platform over three sessions. In a third condition, participants initially learned with moderate immersion and transitioned to higher immersion after the second session. Following research on novelty effects, we explored whether subjective experiences and performance would decline over time (e.g., decreasing motivation or performance) as participants became familiar with the virtual environment and tools. However, we hypothesized that transitional immersion (i.e., switching from moderate to higher immersion) might lead to a renewed sense of novelty. Results suggested that both moderate and higher levels of immersion were motivating, engaging, and supportive of learning. In contrast to predictions based on novelty effects, these outcomes did not decline overall as learners gained familiarity with the systems. However, transitional immersion emerged as a promising and testable pedagogical approach for future VR education. All participants also showed gains in spatial reasoning.

Technology has enabled various alternative educational platforms, such as online courses. Compared to human instructors in traditional educational environments, alternative platforms often show a limited capacity to evaluate the learning progress of students and implement intervention strategies based on the evaluation. Here, we tested participants' hand motions, recorded using a computer mouse and a touchscreen, to determine if the data can predict learners' struggles. Hand motions of participants concurrently performing arithmetic and motor tasks were examined to investigate how the hand motions varied depending on the difficulty of the arithmetic task. The results indicated that working memory load affected both the temporal and spatial features of hand motions, which were predictive of participants' level of working memory load. These findings demonstrate that the assessment of struggles in learning may be achieved at relatively high accuracy with input devices commonly used by learners to access online education systems.

Experimental research around virtual humans acting as pedagogical agents has often taken place in learner-paced learning environments. However, virtual humans are increasingly embedded in educational materials such as instructional videos, where the pacing of the environment can be fundamentally different than a stand-alone learner-controlled software package. This study examined the influence of three types of pacing with varying levels of learner control when learning from an instructional video with an embedded virtual human. The results of our three-group randomized study indicate that increased learner control led to the strongest learning outcomes, although moderate learner control was the most instructionally efficient. The results suggest that some aspects of learner control can be beneficial when learning from instructional videos with embedded virtual humans.

ObjectiveIn this study we aimed to investigate the content and process of contraceptive counseling surrounding combined hormonal contraceptive (CHC) methods (combined oral contraceptives, the ring, and the patch).Study designWe performed a mixed methods analysis of data collected as part of the Patient-Provider Communication about Contraception study, in which reproductive age women and their providers were recruited at several San Francisco Bay Area clinics from 2009-2012. Participants completed pre- and post-visit surveys, and had their visits audio recorded and transcribed. We performed descriptive and bivariate analyses of the entire cohort to examine associations between demographic characteristics and pre-existing method preferences with method selection and counseling content, and coded transcripts of a subset of the sample for salient themes related to content and process of counseling about combined hormonal contraceptive methods using a directed content analysis approach.ResultsThe overall sample included 342 women, with 152 women (44%) having a preference for a specific CHC prior to their visit, 127 women (37%) had a preference for a non-CHC method, and 63 (18%) having no existing method preference. Of the women who reported preferring a CHC in their pre-visit survey, the majority (72%) chose that method. We found that women were inconsistently counseled about the range of CHC methods. For example, women who had no pre-visit method preference (52%) or who preferred the ring (54%) or the patch (73%) were more likely to receive comprehensive counseling about the three CHC methods than were women who preferred combined oral contraceptives (35%) or non-CHC methods (33%). Providers mentioned the patch the least often, and in qualitative analysis indicated discomfort with prescribing this method. Side effects and benefits of methods, as well as strategies to enhance successful use of the chosen method, were inconsistently discussed. In only 73% of visits in which a woman chose a CHC did the provider assess the patient's ability to use the chosen method correctly, and in 66% of all visits in which women chose a CHC method, providers discussed what to do if she was dissatisfied with the method.ConclusionsCounseling about combined hormonal contraceptive methods often does not include information about all available methods, or comprehensive information about side effects, benefits, or logistics of use. As this counseling can impact patient's satisfaction with and continuation of their chosen method of contraception, future work should focus on designing interventions to improve providers' ability to meet patients' needs.ImplicationsShort acting hormonal contraception is widely used, but counseling for these methods often neglects key features. Comprehensive counseling about all methods and their individual features can improve contraceptive selection and use.

The current study examined youths' and their parents' perceptions concerning participation in an investigation of spontaneous and induced pain during recovery from laparoscopic appendectomy. Youth (age range 5-17 years) and their parents independently completed surveys about their study participation. On a 0 (very negative) -to-10 (very positive) scale, both parents 9.4(1.3) [mean(SD)] and youth 7.9(2.4) rated their experience as positive. Among youth, experience ratings did not differ by pain severity and survey responses did not differ by age. Most youth (83%) reported they would tell another youth to participate. Ethical issues regarding instigation of pain in youth for research purposes are examined.

Objective pain assessment is required for appropriate pain management in the clinical setting. However, clinical gold standard pain assessment is based on subjective methods. Automated pain detection from physiological data may provide important objective information to better standardize pain assessment. Specifically, electrodermal activity (EDA) can identify features of stress and anxiety induced by varying pain levels. However, notable variability in EDA measurement exists and research to date has demonstrated sensitivity but lack of specificity in pain assessment. In this paper, we use timescale decomposition (TSD) to extract salient features from EDA signals to identify an accurate and automated EDA pain detection algorithm to sensitively and specifically distinguish pain from no-pain conditions.

Accurately determining pain levels in children is difficult, even for trained professionals and parents. Facial activity provides sensitive and specific information about pain, and computer vision algorithms have been developed to automatically detect Facial Action Units (AUs) defined by the Facial Action Coding System (FACS). Our prior work utilized information from computer vision, i.e., automatically detected facial AUs, to develop classifiers to distinguish between pain and no-pain conditions. However, application of pain/no-pain classifiers based on automated AU codings across different environmental domains results in diminished performance. In contrast, classifiers based on manually coded AUs demonstrate reduced environmentally-based variability in performance. In this paper, we train a machine learning model to recognize pain using AUs coded by a computer vision system embedded in a software package called iMotions. We also study the relationship between iMotions (automatically) and human (manually) coded AUs. We find that AUs coded automatically are different from those coded by a human trained in the FACS system, and that the human coder is less sensitive to environmental changes. To improve classification performance in the current work, we applied transfer learning by training another machine learning model to map automated AU codings to a subspace of manual AU codings to enable more robust pain recognition performance when only automatically coded AUs are available for the test data. With this transfer learning method, we improved the Area Under the ROC Curve (AUC) on independent data from new participants in our target domain from 0.67 to 0.72.

Biases influence the decisions people make in everyday life, even if they are unaware of it. The current study investigates the extent bias behavior transfers into social interactions in virtual worlds by investigating the effect of aversive racism on helping behaviors and learning within a virtual world for medical triage training. In a 2 × 2 × 2 mixed design, two between subjects variables, participant skin tone (light, dark) and avatar skin tone (light, dark), and one within subjects variable, agent skin tone (light, dark), were manipulated. Effects on helping behaviors were observed on three measures: time to initiate help, errors made while helping virtual patients, and learning. Participants, regardless of their skin tone or their avatar's skin tone, took more time to initiate help and made more errors while triaging dark-skinned agents in comparison to light-skinned agents. The bias against virtual patients with a darker skin tone also served as a mediating factor for learning with lower prior knowledge increasing the errors made for dark skinned virtual patients, which had more of a negative impact on learning than the errors made on light skin virtual patients. This study showed that participants applied general biases against dark-skinned agents regardless of participant's ethnicity or avatar's skin-tone. It indicates the importance of considering biases when designing training systems.

Technology advances quickly in today's society. This is particularly true in regard to instructional multimedia. One increasingly important aspect of instructional multimedia design is determining the type of voice that will provide the narration; however, research in the area is dated and limited in scope. Using a randomized pretest-posttest design, we examined the efficacy of learning from an instructional animation where narration was provided by an older text-to-speech engine, a modern text-to-speech engine, or a recorded human voice. In most respects, those who learned from the modern text-to-speech engine were not statistically different in regard to their perceptions, learning outcomes, or cognitive efficiency measures compared with those who learned from the recorded human voice. Our results imply that software technologies may have reached a point where they can credibly and effectively deliver the narration for multimedia learning environments.