This study aimed to establish an accurate baseline for Facebook usage and to investigate its learning potential via biometric perspective. This was explored by asking college students to browse their own Facebook page and peruse the content while wearing a high-fidelity EEG headset to record brainwave activity they experienced. Their actions were then coded using the empirically based Interactive-Constructive-Active-Passive (ICAP) engagement model to identify elements that could potentially be leveraged for learning. Based on these results, Facebook does not seem to make an ideal learning platform, due to higher frustration levels during higher cognitive tasks compared to passively consuming personally relevant non-instructional content. However, it may have more positive impacts on learning as long as initial frustration level is transformed smoothly to engagement and motivation for some learning opportunities and activities.

Many teens struggle with school writing and particularly in identifying personally relevant topics, which motivate writing improvement. However, a wealth of potential topics is available through their social media data. We explore the design of Sparkfolio, a prewriting support tool that aims to help students successfully prepare meaningful writing topics from their own social media content. This web-delivered tool was evaluated with 46 teen users in the context of their high school English class. Findings demonstrate that the use of social media can enhance the quality of written outcomes, as Sparkfolio users had significantly greater gains than a control group.

Virtual worlds enable users' interactions through avatars. Avatars embody individual characteristics from their owners and exhibit those characteristics outward to the community. Motivated by the role of avatars in interpersonal communication, we integrated a generic real-time multimodal affect recognition hub as an input within an online virtual world to make an avatar mirror its owner's affect. Affect vectors (determined by pleasure, arousal, and dominance coordinates) in a continuous affective space are applied to characterize the user's affective state in real time.

A step-based tutoring system for linear circuit analysis is being developed with the capabilities to automatically generate circuit problems with specified characteristics, including randomly generated topologies and element values. The system further generates fully-worked, error-free solutions using the methods typically taught in such classes, and accepts a rich variety of student input such as equations, matrix equations, numerical and multiple-choice answers, re-drawn circuit diagrams, and sketches of waveforms. A randomized, controlled study was conducted using paid student volunteers to compare the effectiveness of two of our tutorials in comparison to working conventional textbook-based problems. The average learning gain was only 3/100 points for the textbook users, but 29/100 points, about 10 times higher, for the tutorial users. The effect size on the post-test scores was 1.21 pooled standard deviations (Cohen d-value) and was statistically significant. A motivational survey administered to these students yielded a 0.53 point higher rating for the software than for the textbook (on a 1-5 scale). The system is being used in Spring 2013 by over 340 students in EEE 202 at Arizona State and two community colleges. About 99% of these students rated the system as “very helpful” or “somewhat helpful”.

The ability of a learning system to infer a student's affects has become highly relevant to be able to adjust its pedagogical strategies. Several methods have been used to infer affects. One of the most recognized for its reliability is face-based affect recognition. Another emerging one involves the use of brain-computer interfaces. In this paper we compare those strategies and explore if, to a great extent, it is possible to infer the values of one source from the other source.

The purpose of this investigation was to develop the STEM Active Listening Skills Assessment (STEM-ALSA), a conceptually grounded instrument designed to measure four components of active listening, a key element of communication in an academic setting. The STEM-ALSA is comprised of three unique scales that measure a person's knowledge (12 items), ability to apply (25 items), and self-efficacy (5 items) with respect to active listening. Two pilot studies were conducted with N = 99 upper level undergraduate students enrolled in STEM disciplines to develop and evaluate the instrument. Results of an exploratory factor analysis identified both a unidimensional factor structure for each of the three scales and total scores with adequate internal consistency reliability estimates. The STEM-ALSA provides a mechanism for measuring active listening skills among students in STEM.

The human-element is crucial for designing and implementing interactive intelligent systems, and therefore on instructional design. This tutorial provides a description and hands-on demonstration for detection of affective states and a description of devices, methodologies and tools necessary for automatic detection of affective states. Automatic detection of affective states requires that the computer sense information that is complex and diverse, it can range from brain-waves signals, and biofeedback readings to face-based and gesture emotion recognition to posture and pressure sensing. Obtaining, processing and understanding that information, to create systems that improve learning, requires the use of several sensing devices (and their perceiving algorithms) and the application of software tools.

The computer's ability to recognize human emotional states given physiological signals is gaining in popularity to create empathetic systems such as learning environments, health care systems and videogames. Despite that, there are few frameworks, libraries, architectures, or software tools, which allow systems developers to easily integrate emotion recognition into their software projects. The work reported here offers a first step to fill this gap in the lack of frameworks and models, addressing: (a) the modeling of an agent-driven component-based architecture for multimodal emotion recognition, called ABE, and (b) the use of ABE to implement a multimodal emotion recognition framework to support third-party systems becoming empathetic systems.